Energy consumption simulation and prediction method, system, device and medium based on factory operating state

Abstract

The present disclosure relates to a method, system, device and medium for energy consumption simulation and prediction based on factory operation state, and relates to the technical field of energy consumption simulation and prediction. The method for energy consumption simulation based on factory operation state comprises: training a preset energy consumption simulation evaluation generator by using historical use time, historical power consumption data and / or historical gas consumption data corresponding to at least one collection time in a set time interval of each set energy consumption device in a plurality of set energy consumption devices under a factory operation state and historical energy consumption evaluation corresponding to the historical use time, the historical power consumption data and / or the historical gas consumption data, to obtain the preset energy consumption simulation evaluation generator; and generating a corresponding energy consumption simulation evaluation by using current use time, current power consumption data and / or current gas consumption data corresponding to each set energy consumption device in the plurality of set energy consumption devices under the factory operation state based on the preset energy consumption simulation evaluation generator. The present disclosure can realize energy consumption simulation and prediction based on factory operation state.

Description

Technical Field

[0001] This disclosure relates to the field of energy consumption simulation and prediction technology, and in particular to a method, system, device and medium for energy consumption simulation and prediction based on factory operating status. Background Technology

[0002] From the perspective of energy utilization in national economic activities, unit energy consumption is a key indicator reflecting energy consumption levels and energy conservation efforts. The ratio of total primary energy supply to GDP is an energy efficiency indicator. Therefore, efficiency indicators illustrate the extent of energy utilization in a country's economic activities and reflect changes in economic structure and energy efficiency.

[0003] For factories, unit energy consumption is a key indicator reflecting energy consumption levels and energy conservation efforts. However, the production value of energy-consuming equipment within a factory is generally fixed during its operation. But as usage time increases, the equipment ages, leading to increased energy consumption to maintain production levels. Furthermore, aging equipment generates more heat during operation, requiring cooling systems to cool it or the surrounding environment, further increasing energy consumption. Currently, energy consumption assessments do not consider the impact of equipment usage time. Moreover, factory operation requires the collaboration of multiple energy-consuming devices, making it difficult to establish a linear relationship between the usage time of these devices and the overall energy consumption of the factory. Therefore, it is necessary to consider the impact of equipment usage time on energy consumption assessment to achieve energy consumption simulation and prediction during factory operation. Summary of the Invention

[0004] This disclosure proposes a technical solution for energy consumption simulation and prediction based on factory operating conditions, including a method, system, equipment, and corresponding media.

[0005] According to one aspect of this disclosure, an energy consumption simulation method based on factory operating status is provided, comprising:

[0006] By utilizing the historical usage time, historical electricity consumption data, and / or historical gas consumption data of each of the multiple energy-consuming devices under factory operation conditions at at least one collection moment within a set time interval, along with their corresponding historical energy consumption evaluations, a preset energy consumption simulation evaluation generation network is trained to obtain the preset energy consumption simulation evaluation generator. Based on the preset energy consumption simulation evaluation generator, a corresponding energy consumption simulation evaluation is generated using the current usage time, current electricity consumption data, and / or current gas consumption data of each of the multiple energy-consuming devices under factory operation conditions.

[0007] Preferably, the step of training a preset energy consumption simulation evaluation generation network using historical usage time, historical electricity consumption data, and / or historical gas consumption data corresponding to at least one collection moment within a set time interval for each of multiple set energy consumption devices under factory operation status, along with their corresponding historical energy consumption evaluations, to obtain the preset energy consumption simulation evaluation generator includes: configuring the historical usage time, historical electricity consumption data, and / or historical gas consumption data as classification feature vectors; configuring the energy consumption evaluations corresponding to the historical usage time, historical electricity consumption data, and / or historical gas consumption data as classification training labels corresponding to the classification feature vectors; training the preset energy consumption simulation evaluation generation network using the classification feature vectors to obtain corresponding training process energy consumption simulation evaluations; calculating the loss values ​​corresponding to the training process energy consumption simulation evaluations and the classification training labels, and then training the preset energy consumption simulation evaluation generation network corresponding to the preset energy consumption simulation evaluation generator.

[0008] Preferably, the method of calculating the loss value corresponding to the energy consumption simulation evaluation of the training process and the label, and then training the preset energy consumption simulation evaluation generation network corresponding to the preset energy consumption simulation evaluation generator, includes: if the loss value corresponding to the energy consumption simulation evaluation of the training process and the label is less than or equal to the preset loss value within a set number of training iterations, then the training of the preset energy consumption simulation evaluation generation network corresponding to the preset energy consumption simulation evaluation generator is completed.

[0009] Preferably, before training the preset energy consumption simulation evaluation generation network, a connection to a host computer database is established through an industrial bus. The industrial bus is used to obtain the historical usage time, historical electricity consumption data and / or historical gas consumption data and their corresponding historical energy consumption evaluations corresponding to at least one collection moment within a set time interval for each of the multiple set energy consumption devices under the factory operating state.

[0010] Preferably, the step of training the preset energy consumption simulation evaluation generation network using the classification feature vector to obtain the corresponding training process energy consumption simulation evaluation includes: acquiring random noise under the conditions of the historical electricity consumption data and / or historical gas consumption data; configuring the historical usage time, the historical electricity consumption data and / or historical gas consumption data, and the random noise as classification feature vectors; configuring the energy consumption evaluations corresponding to the historical usage time, the historical electricity consumption data and / or historical gas consumption data, and the random noise as classification training labels corresponding to the classification feature vectors; and training the preset energy consumption simulation evaluation generation network using the classification feature vectors to obtain the corresponding training process energy consumption simulation evaluation.

[0011] Preferably, before training the preset energy consumption simulation evaluation generation network using the classification feature vector, the classification feature vector is standardized to obtain a standardized classification feature vector; the preset energy consumption simulation evaluation generation network is then trained using the standardized classification feature vector.

[0012] Preferably, the step of calculating the loss value corresponding to the energy consumption simulation evaluation and the classification training label during the training process includes: based on a preset energy consumption simulation evaluation network, using the historical usage time, the historical electricity consumption data and / or gas consumption data, the energy consumption simulation evaluation, and the energy consumption evaluation, determining whether the energy consumption simulation evaluation is true or false; calculating the first loss of the energy consumption simulation evaluation corresponding to random noise under the conditions of the historical electricity consumption data and / or historical gas consumption data, and the second loss corresponding to determining whether the energy consumption simulation evaluation is true or false, and training the preset energy consumption simulation evaluation generation network.

[0013] Preferably, the preset energy consumption simulation evaluation network is configured as one or more of support vector regression, support vector machine, K-nearest neighbor algorithm, random forest, artificial neural network, deep neural network, etc.

[0014] Preferably, the energy consumption simulation evaluation and / or the energy consumption evaluation is configured as one of a high energy consumption level, a medium energy consumption level, a low energy consumption level, and a normal energy consumption level.

[0015] Preferably, each energy-consuming device is equipped with a corresponding power consumption detection sensor and / or gas consumption sensor; using the corresponding power consumption detection sensor and / or gas consumption detection sensor for each energy-consuming device, historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are collected; and using an industrial bus connected to the power consumption detection sensor and / or gas consumption detection sensor for each energy-consuming device, the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are sent to a host computer database, and the host computer database stores the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data.

[0016] Preferably, a timer is used to time the usage time of each set energy-consuming device from turning on to turning off, thereby obtaining the usage time corresponding to each set energy-consuming device; the usage time corresponding to each set energy-consuming device is sent to a host computer database through an industrial bus connected to the timer, and the host computer database is used to store the usage time corresponding to each set energy-consuming device.

[0017] Preferably, the preset energy consumption simulation evaluation generation network is configured as one or more of support vector regression, support vector machine, K-nearest neighbor algorithm, random forest, artificial neural network, deep neural network, etc.

[0018] The energy consumption simulation method based on factory operating status further includes: training a preset energy consumption simulation evaluation generation network using the historical usage time, historical electricity consumption data, and / or historical gas consumption data of each of the multiple set energy consumption devices in the factory operating status at a first moment within a set time interval, and the historical energy consumption evaluation corresponding to a second moment after the first moment, to obtain the preset energy consumption simulation evaluation generator; and predicting the energy consumption evaluation corresponding to a fourth moment after the third moment using the current usage time, current electricity consumption data, and / or current gas consumption data of each of the multiple set energy consumption devices in the factory operating status at a third moment.

[0019] According to one aspect of this disclosure, an energy consumption prediction method based on factory operating status is provided, comprising:

[0020] By utilizing the historical usage time, historical electricity consumption data, and / or historical gas consumption data of each of the multiple energy-consuming devices under factory operation conditions at a first moment within a set time interval, and the historical energy consumption evaluation corresponding to the second moment after the first moment, a preset energy consumption simulation evaluation generation network is trained to obtain the preset energy consumption simulation evaluation generator; based on the preset energy consumption simulation evaluation generator, the energy consumption evaluation corresponding to the fourth moment after the third moment is predicted by using the current usage time, current electricity consumption data, and / or current gas consumption data of each of the multiple energy-consuming devices under factory operation conditions at a third moment.

[0021] Preferably, the step of training a preset energy consumption simulation evaluation generation network using the historical usage time, historical electricity consumption data, and / or historical gas consumption data of each of the multiple set energy consumption devices under factory operation conditions at a first moment within a set time interval, and the historical energy consumption evaluation at a second moment after the first moment, to obtain the preset energy consumption simulation evaluation generator, includes: configuring the historical usage time, historical electricity consumption data, and / or historical gas consumption data corresponding to the first moment as a prediction feature vector; configuring the historical energy consumption evaluation at the second moment after the first moment as a prediction training label corresponding to the prediction feature vector; training the preset energy consumption simulation evaluation generation network using the prediction feature vector to obtain a corresponding training process energy consumption simulation evaluation; calculating the loss value corresponding to the training process energy consumption simulation evaluation and the prediction training label, and then training the preset energy consumption simulation evaluation generation network corresponding to the preset energy consumption simulation evaluation generator.

[0022] Preferably, the method of calculating the loss value corresponding to the energy consumption simulation evaluation of the training process and the label, and then training the preset energy consumption simulation evaluation generation network corresponding to the preset energy consumption simulation evaluation generator, includes: if the loss value corresponding to the energy consumption simulation evaluation of the training process and the label is less than or equal to the preset loss value within a set number of training iterations, then the training of the preset energy consumption simulation evaluation generation network corresponding to the preset energy consumption simulation evaluation generator is completed.

[0023] Preferably, before training the preset energy consumption simulation evaluation generation network, a connection to a host computer database is established through an industrial bus. The industrial bus is used to obtain from the host computer database the historical usage time, historical electricity consumption data, and / or historical gas consumption data of each of the multiple set energy consumption devices in the factory operating state at the first moment within a set time interval, as well as the historical energy consumption evaluation corresponding to the second moment after the first moment.

[0024] Preferably, training the preset energy consumption simulation evaluation generation network using the predicted feature vector to obtain the corresponding training process energy consumption simulation evaluation includes: acquiring random noise under the conditions of the historical electricity consumption data and / or historical gas consumption data corresponding to the first time moment; configuring the historical usage time, the historical electricity consumption data and / or historical gas consumption data, and the random noise corresponding to the first time moment as the predicted feature vector; configuring the energy consumption evaluation corresponding to the second time moment as the predicted training label corresponding to the predicted feature vector; and training the preset energy consumption simulation evaluation generation network using the predicted feature vector to obtain the corresponding training process energy consumption simulation evaluation.

[0025] Preferably, before training the preset energy consumption simulation evaluation generation network using the predicted feature vector, the method includes: standardizing the predicted feature vector to obtain a standardized predicted feature vector; and training the preset energy consumption simulation evaluation generation network using the standardized predicted feature vector.

[0026] Preferably, the step of calculating the loss value corresponding to the energy consumption simulation evaluation of the training process and the predicted training label includes: based on a preset energy consumption simulation evaluation network, using the historical usage time, the historical electricity consumption data and / or gas consumption data, the energy consumption simulation evaluation and the energy consumption evaluation, determining whether the energy consumption simulation evaluation is true or false; calculating the first loss of the energy consumption simulation evaluation corresponding to random noise under the conditions of the historical electricity consumption data and / or historical gas consumption data and the second loss corresponding to determining whether the energy consumption simulation evaluation is true or false, and training the preset energy consumption simulation evaluation generation network.

[0027] Preferably, the energy consumption simulation method based on factory operating status is characterized in that the energy consumption simulation evaluation and / or the energy consumption evaluation is configured as one of high energy consumption level, medium energy consumption level, low energy consumption level and normal energy consumption level.

[0028] Preferably, each energy-consuming device is equipped with a corresponding power consumption detection sensor and / or gas consumption sensor; using the corresponding power consumption detection sensor and / or gas consumption detection sensor for each energy-consuming device, historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are collected; and using an industrial bus connected to the power consumption detection sensor and / or gas consumption detection sensor for each energy-consuming device, the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are sent to a host computer database, and the host computer database stores the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data.

[0029] Preferably, a timer is used to time the usage time of each set energy-consuming device from turning on to turning off, thereby obtaining the usage time corresponding to each set energy-consuming device; the usage time corresponding to each set energy-consuming device is sent to a host computer database through an industrial bus connected to the timer, and the host computer database is used to store the usage time corresponding to each set energy-consuming device.

[0030] Preferably, the preset energy consumption simulation evaluation generation network is configured as one or more of support vector regression, support vector machine, K-nearest neighbor algorithm, random forest, artificial neural network, deep neural network, etc.

[0031] Preferably, the energy consumption prediction method based on factory operating status further includes: using the historical usage time, historical electricity consumption data and / or historical gas consumption data corresponding to at least one collection moment in a set time interval for each of the multiple set energy consumption devices under factory operating status, and their corresponding historical energy consumption evaluations, to train a preset energy consumption simulation evaluation generation network to obtain the preset energy consumption simulation evaluation generator.

[0032] Based on the preset energy consumption simulation evaluation generator, the corresponding energy consumption simulation evaluation is generated by using the current usage time, current electricity consumption data and / or current gas consumption data of each of the multiple set energy consumption devices in the factory operation state.

[0033] According to one aspect of this disclosure, an energy consumption simulation system based on factory operating status is provided, comprising: a first training unit, configured to train a preset energy consumption simulation evaluation generation network by utilizing historical usage time, historical electricity consumption data, and / or historical gas consumption data corresponding to at least one acquisition time within a set time interval for each of a plurality of set energy consumption devices under factory operating status, and their corresponding historical energy consumption evaluations, to obtain the preset energy consumption simulation evaluation generator; and a simulation evaluation unit, configured to generate a corresponding energy consumption simulation evaluation based on the preset energy consumption simulation evaluation generator, using the current usage time, current electricity consumption data, and / or current gas consumption data corresponding to each of the plurality of set energy consumption devices under factory operating status.

[0034] According to one aspect of this disclosure, an energy consumption simulation system based on factory operating status is provided, comprising: an electronic device, the electronic device being configured with a processor and a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the aforementioned energy consumption simulation method based on factory operating status.

[0035] According to one aspect of this disclosure, an energy consumption simulation system based on factory operating status is provided, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to perform the aforementioned energy consumption simulation method based on factory operating status.

[0036] According to one aspect of this disclosure, an energy consumption simulation system based on factory operating status is provided, comprising: a computer program product, the computer program product being provided with a computer program / instruction, which, when executed by a processor, implements the energy consumption simulation method based on factory operating status as described above.

[0037] According to one aspect of this disclosure, an energy consumption simulation device based on factory operating status is provided, comprising: a first training unit, configured to train a preset energy consumption simulation evaluation generation network by utilizing historical usage time, historical electricity consumption data and / or historical gas consumption data corresponding to at least one collection moment within a set time interval for each of a plurality of set energy consumption devices under factory operating status, and their corresponding historical energy consumption evaluations, to obtain the preset energy consumption simulation evaluation generator; and a simulation evaluation unit, configured to generate a corresponding energy consumption simulation evaluation based on the preset energy consumption simulation evaluation generator, using the current usage time, current electricity consumption data and / or current gas consumption data corresponding to each of the plurality of set energy consumption devices under factory operating status.

[0038] According to one aspect of this disclosure, an energy consumption simulation device based on factory operating status is provided, comprising: an electronic device, the electronic device being configured with a processor and a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the aforementioned energy consumption simulation method based on factory operating status.

[0039] According to one aspect of this disclosure, an energy consumption simulation device based on factory operating status is provided, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to perform the aforementioned energy consumption simulation method based on factory operating status.

[0040] According to one aspect of this disclosure, an energy consumption simulation device based on factory operating status is provided, comprising: a computer program product, the computer program product being provided with a computer program / instruction, which, when executed by a processor, implements the energy consumption simulation method based on factory operating status as described above.

[0041] According to one aspect of this disclosure, an energy consumption prediction system based on factory operating status is provided, comprising: a second training unit, configured to train a preset energy consumption simulation evaluation generation network using historical usage time, historical electricity consumption data, and / or historical gas consumption data of each of a plurality of set energy consumption devices in the factory operating status at a first moment within a set time interval, and historical energy consumption evaluation at a second moment after the first moment, to obtain the preset energy consumption simulation evaluation generator; and an energy consumption evaluation prediction unit, configured to predict the energy consumption evaluation at a fourth moment after the third moment based on the preset energy consumption simulation evaluation generator, using current usage time, current electricity consumption data, and / or current gas consumption data of each of the plurality of set energy consumption devices in the factory operating status at a third moment.

[0042] According to one aspect of this disclosure, an energy consumption prediction system based on factory operating status is provided, comprising: an electronic device, the electronic device being configured with a processor and a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the aforementioned energy consumption prediction method based on factory operating status.

[0043] According to one aspect of this disclosure, an energy consumption prediction system based on factory operating status is provided, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the above-described energy consumption prediction method based on factory operating status.

[0044] According to one aspect of this disclosure, an energy consumption prediction system based on factory operating status is provided, comprising: a computer program product, the computer program product being provided with a computer program / instruction, which, when executed by a processor, implements the energy consumption prediction method based on factory operating status as described above.

[0045] According to one aspect of this disclosure, an energy consumption prediction device based on factory operating status is provided, comprising: a second training unit, configured to train a preset energy consumption simulation evaluation generation network using historical usage time, historical electricity consumption data and / or historical gas consumption data of each of a plurality of set energy consumption devices in the factory operating status at a first moment within a set time interval, and historical energy consumption evaluation corresponding to a second moment after the first moment, to obtain the preset energy consumption simulation evaluation generator; and an energy consumption evaluation prediction unit, configured to predict the energy consumption evaluation corresponding to a fourth moment after the third moment based on the preset energy consumption simulation evaluation generator and using the current usage time, current electricity consumption data and / or current gas consumption data of each of the plurality of set energy consumption devices in the factory operating status at a third moment.

[0046] According to one aspect of this disclosure, an energy consumption prediction device based on factory operating status is provided, comprising: an electronic device, the electronic device being configured with a processor and a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the aforementioned energy consumption prediction method based on factory operating status.

[0047] According to one aspect of this disclosure, an energy consumption prediction device based on factory operating status is provided, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the above-described energy consumption prediction method based on factory operating status.

[0048] According to one aspect of this disclosure, an energy consumption prediction device based on factory operating status is provided, comprising: a computer program product, the computer program product being provided with a computer program / instruction, which, when executed by a processor, implements the energy consumption prediction method based on factory operating status as described above.

[0049] According to one aspect of this disclosure, a computer-readable storage medium is provided that stores computer program instructions thereon, which, when executed by a processor, implement the above-described energy consumption simulation method based on factory operating conditions.

[0050] According to one aspect of this disclosure, a computer-readable storage medium is provided that stores computer program instructions thereon, which, when executed by a processor, implement the above-described energy consumption prediction method based on factory operating status.

[0051] According to one aspect of this disclosure, a computer-readable storage medium is provided that stores computer program instructions thereon, which, when executed by a processor, implement the above-described energy consumption simulation method based on factory operating status and the above-described energy consumption prediction method based on factory operating status.

[0052] In the embodiments of this disclosure, the proposed technical solutions for energy consumption simulation and prediction methods, systems, equipment and media based on factory operating status are intended to solve at least one technical problem in the current energy consumption simulation and prediction of factory operating status that does not consider the impact of the usage time of energy-consuming equipment on energy consumption evaluation.

[0053] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure.

[0054] Other features and aspects of this disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0055] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the specification, serve to illustrate the technical solutions of this disclosure.

[0056] Figure 1 A flowchart illustrating an energy consumption simulation method based on factory operating status according to an embodiment of the present disclosure is shown;

[0057] Figure 2 A schematic diagram of an energy consumption prediction method based on factory operating status according to an embodiment of the present disclosure is shown;

[0058] Figure 3 A schematic diagram of an energy consumption simulation system based on factory operating status according to an embodiment of the present disclosure is shown;

[0059] Figure 4 A schematic diagram of an energy consumption prediction system based on factory operating status according to an embodiment of the present disclosure is shown. Detailed Implementation

[0060] Various exemplary embodiments, features, and aspects of this disclosure will now be described in detail with reference to the accompanying drawings. The same reference numerals in the drawings denote elements that have the same or similar functions. Although various aspects of the embodiments are shown in the drawings, they are not necessarily drawn to scale unless specifically indicated otherwise.

[0061] The term “exemplary” as used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments.

[0062] In this document, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Furthermore, the term "at least one" in this document means any combination of at least two of any one or more elements. For example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C.

[0063] Furthermore, to better illustrate this disclosure, numerous specific details are set forth in the following detailed description. Those skilled in the art will understand that this disclosure can be practiced without certain specific details. In some instances, methods, means, components, and circuits well known to those skilled in the art have not been described in detail in order to highlight the main points of this disclosure.

[0064] It is understood that the various method embodiments mentioned above in this disclosure can be combined with each other to form combined embodiments without violating the principle and logic. Due to space limitations, this disclosure will not elaborate further.

[0065] In addition, this disclosure also provides an energy consumption simulation device and an energy consumption simulation and prediction device or system based on factory operating status, electronic equipment, computer-readable storage medium, and program products, all of which can be used to implement any of the energy consumption simulation and prediction methods based on factory operating status provided in this disclosure. The corresponding technical solutions and descriptions are described in the corresponding descriptions in the method section and will not be repeated here.

[0066] Figure 1 A flowchart illustrating an energy consumption simulation method based on factory operating conditions according to an embodiment of this disclosure is shown. Figure 1 As shown, the energy consumption simulation method based on factory operating status includes: Step S101: Using historical usage time, historical electricity consumption data, and / or historical gas consumption data corresponding to at least one collection moment within a set time interval for each of multiple set energy consumption devices under factory operating status, and their corresponding historical energy consumption evaluations, a preset energy consumption simulation evaluation generation network is trained to obtain the preset energy consumption simulation evaluation generator; Step S102: Based on the preset energy consumption simulation evaluation generator, a corresponding energy consumption simulation evaluation is generated using the current usage time, current electricity consumption data, and / or current gas consumption data corresponding to each of the multiple set energy consumption devices under factory operating status. This addresses the technical problem that current energy consumption prediction methods for factory operating status do not consider the impact of energy consumption device usage time on energy consumption evaluation.

[0067] Step S101: Using the historical usage time, historical electricity consumption data and / or historical gas consumption data corresponding to at least one collection moment within a set time interval for each of the multiple set energy consumption devices under factory operation status, and their corresponding historical energy consumption evaluations, train the preset energy consumption simulation evaluation generation network to obtain the preset energy consumption simulation evaluation generator.

[0068] In the embodiments disclosed herein and other possible embodiments, those skilled in the art can configure multiple energy-consuming devices under factory operating conditions according to actual needs. For example, production line equipment and its corresponding drive motors, lighting equipment, air conditioning equipment, etc., one or more. For example, factory circulating water equipment and its corresponding drive motors, lighting equipment, air conditioning equipment, etc., one or more. For example, steel rolling production line equipment and its corresponding drive motors, lighting equipment, air conditioning equipment, etc., one or more. For example, sewing machines, ironing machines, etc., in a garment factory and their corresponding drive motors, lighting equipment, air conditioning equipment, etc., one or more.

[0069] In the embodiments of this disclosure and other possible embodiments, those skilled in the art can configure the set time interval according to actual needs. For example, the set time interval can be configured as a week, 30 days, 6 months, or other time intervals.

[0070] In the embodiments of this disclosure and other possible embodiments, the set time interval may include one or more of the following: peak electricity and / or gas consumption periods, off-peak electricity and / or gas consumption periods, and electricity and / or gas consumption periods during holidays.

[0071] In the embodiments of this disclosure, the step of training a preset energy consumption simulation evaluation generation network using historical usage time, historical electricity consumption data, and / or historical gas consumption data corresponding to at least one collection moment within a set time interval for each of multiple set energy consumption devices under factory operation status, along with their corresponding historical energy consumption evaluations, to obtain the preset energy consumption simulation evaluation generator, includes: configuring the historical usage time, the historical electricity consumption data, and / or historical gas consumption data as classification feature vectors; configuring the energy consumption evaluations corresponding to the historical usage time, historical electricity consumption data, and / or historical gas consumption data as classification training labels corresponding to the classification feature vectors; training the preset energy consumption simulation evaluation generation network using the classification feature vectors to obtain corresponding training process energy consumption simulation evaluations; calculating the loss values ​​corresponding to the training process energy consumption simulation evaluations and the classification training labels, and then training the preset energy consumption simulation evaluation generation network corresponding to the preset energy consumption simulation evaluation generator.

[0072] In the embodiments of this disclosure, the method of calculating the loss value corresponding to the energy consumption simulation evaluation of the training process and the label, and then training the preset energy consumption simulation evaluation generation network corresponding to the preset energy consumption simulation evaluation generator, includes: if the loss value corresponding to the energy consumption simulation evaluation of the training process and the label is less than or equal to the preset loss value within a set number of training iterations, then the training of the preset energy consumption simulation evaluation generation network corresponding to the preset energy consumption simulation evaluation generator is completed.

[0073] In the embodiments of this disclosure, before training the preset energy consumption simulation evaluation generation network, a connection to a host computer database is established through an industrial bus. The industrial bus is used to obtain the historical usage time, historical electricity consumption data and / or historical gas consumption data and their corresponding historical energy consumption evaluations corresponding to at least one collection moment within a set time interval for each of the multiple set energy consumption devices under factory operation status.

[0074] In the embodiments of this disclosure, the energy consumption simulation evaluation and / or the energy consumption evaluation is configured as one of a high energy consumption level, a medium energy consumption level, a low energy consumption level, and a normal energy consumption level. Meanwhile, those skilled in the art can configure the energy consumption simulation evaluation and / or the energy consumption evaluation according to actual needs. For example, the energy consumption simulation evaluation and / or the energy consumption evaluation can be configured as energy efficiency, with the energy efficiency levels divided into energy efficiency levels A to G; wherein energy efficiency level A represents the highest energy efficiency level; conversely, energy efficiency level G represents the lowest energy efficiency level.

[0075] In the embodiments of this disclosure and other possible embodiments, determining the energy consumption evaluation includes: acquiring electricity consumption data and / or gas consumption data of multiple set energy-consuming devices and their corresponding production values; and determining the energy consumption evaluation based on the electricity consumption data and / or gas consumption data of the multiple set energy-consuming devices and the production values. Specifically, determining the energy consumption evaluation based on the electricity consumption data and / or gas consumption data of the multiple set energy-consuming devices and the production values ​​includes: acquiring multiple set ratios; calculating the ratios corresponding to the production values ​​and the electricity consumption data and / or gas consumption data of the multiple set energy-consuming devices; comparing the ratios and the multiple set ratios to determine the energy consumption evaluation, and then configuring the energy consumption evaluation as one of a high energy consumption level, a medium energy consumption level, a low energy consumption level, and a normal energy consumption level.

[0076] In the embodiments of this disclosure and other possible embodiments, the step of comparing the ratio and the plurality of set ratios to determine the energy consumption evaluation, and then configuring the energy consumption evaluation as one of high energy consumption level, medium energy consumption level, low energy consumption level, and normal energy consumption level, includes: obtaining a first set ratio corresponding to the plurality of set ratios, a second set ratio less than the first set ratio, and a third set ratio less than the second set ratio; if the ratio is greater than or equal to the first set ratio, then the energy consumption evaluation is configured as high energy consumption level; if the ratio is less than the first set ratio but greater than or equal to the second set ratio, then the energy consumption evaluation is configured as medium energy consumption level; if the ratio is less than the second set ratio but greater than or equal to the third set ratio, then the energy consumption evaluation is configured as low energy consumption level; if the ratio is less than the third set ratio, then the energy consumption evaluation is configured as normal energy consumption level.

[0077] In the embodiments of this disclosure, each energy-consuming device is equipped with a corresponding power consumption detection sensor and / or gas consumption sensor; using the corresponding power consumption detection sensor and / or gas consumption detection sensor for each energy-consuming device, historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are collected; and using an industrial bus connected to the power consumption detection sensor and / or gas consumption detection sensor for each energy-consuming device, the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are sent to a host computer database, and the host computer database stores the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data.

[0078] In the embodiments of this disclosure, a timer is used to time the usage time of each set energy-consuming device from turning on to turning off, thereby obtaining the usage time corresponding to each set energy-consuming device; the usage time corresponding to each set energy-consuming device is sent to a host computer database through an industrial bus connected to the timer, and the host computer database is used to store the usage time corresponding to each set energy-consuming device.

[0079] In the embodiments of this disclosure, the preset energy consumption simulation evaluation generation network is configured as one or more of the following: Support Vector Regression (SVR), Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Decision Tree (XGBoost), Random Forest (RF), Artificial Neural Network (ANN), and Deep Neural Network (DNN).

[0080] In embodiments of this disclosure and other possible embodiments, the input and output layers of the deep neural network (DNN) respectively include a first number of neurons and a second number of neurons, and each of the plurality of intermediate hidden layers between the input and output layers includes a third number of neurons. To overcome the vanishing gradient problem and accelerate learning, in addition to the Sigmoid and Tanh activation functions, the ReLU activation function is used to approximate the nonlinear function. Specifically, the ReLU activation function is placed after each neuron in the second and fourth layers of the intermediate hidden layers, and the Sigmoid and Tanh activation functions are placed after each neuron in the first and third layers of the intermediate hidden layers, respectively.

[0081] In embodiments of this disclosure, training the preset energy consumption simulation evaluation generation network using the classification feature vector to obtain the corresponding training process energy consumption simulation evaluation includes: acquiring random noise under the conditions of historical electricity consumption data and / or historical gas consumption data; configuring the historical usage time, the historical electricity consumption data and / or historical gas consumption data, and the random noise as classification feature vectors; configuring the energy consumption evaluations corresponding to the historical usage time, the historical electricity consumption data and / or historical gas consumption data, and the random noise as classification training labels corresponding to the classification feature vectors; and training the preset energy consumption simulation evaluation generation network using the classification feature vectors to obtain the corresponding training process energy consumption simulation evaluation.

[0082] In the embodiments of this disclosure, before training the preset energy consumption simulation evaluation generation network using the classification feature vector, the classification feature vector is standardized to obtain a standardized classification feature vector; the preset energy consumption simulation evaluation generation network is then trained using the standardized classification feature vector.

[0083] In the embodiments of this disclosure, calculating the loss value corresponding to the energy consumption simulation evaluation of the training process and the classification training label includes: based on a preset energy consumption simulation evaluation network, using the historical usage time, the historical electricity consumption data and / or gas consumption data, the energy consumption simulation evaluation and the energy consumption evaluation, determining whether the energy consumption simulation evaluation is true or false; calculating the first loss of the energy consumption simulation evaluation corresponding to random noise under the conditions of the historical electricity consumption data and / or historical gas consumption data and the second loss corresponding to determining whether the energy consumption simulation evaluation is true or false, and training the preset energy consumption simulation evaluation generation network.

[0084] In embodiments of this disclosure and other possible embodiments, the calculation of the first loss of the energy consumption simulation evaluation corresponding to random noise under the conditions of historical usage time, historical electricity consumption data and / or gas consumption data, and the second loss corresponding to determining the truth or falsehood of the energy consumption simulation evaluation, includes: constructing a first set loss function corresponding to the first loss and a second set loss function corresponding to the second loss; calculating the first loss of the energy consumption simulation evaluation corresponding to random noise under the conditions of historical usage time, historical electricity consumption data and / or gas consumption data based on the first set loss function; and calculating the second loss corresponding to determining the truth or falsehood of the energy consumption simulation evaluation corresponding to random noise under the conditions of historical electricity consumption data and / or gas consumption data based on the second set loss function.

[0085] In embodiments of this disclosure and other possible embodiments, constructing the first set loss function corresponding to the first loss includes: calculating the difference between a set value and an energy consumption simulation evaluation corresponding to random noise under the conditions of historical usage time, historical electricity consumption data and / or gas consumption data; taking the logarithm of the difference to obtain a logarithmic value of the difference; calculating a first expected value of the logarithmic value of the difference, and constructing the first set loss function corresponding to the first loss; wherein the set value is configured to be 1.

[0086] In the embodiments of this disclosure and other possible embodiments, constructing the first set loss function corresponding to the first loss includes: taking the logarithm of the discrimination probability value corresponding to the random noise under the conditions of the historical usage time, the historical electricity consumption data and / or gas consumption data to determine whether the energy consumption simulation evaluation is true or false, to obtain a discrimination logarithm value; calculating the second expected value of the logarithm value, and constructing the second set loss function corresponding to the second loss. The step of calculating the first loss of the energy consumption simulation evaluation corresponding to random noise under the conditions of the historical usage time, the historical electricity consumption data and / or gas consumption data, and the second loss corresponding to the determination of the true or false nature of the energy consumption simulation evaluation, and training the preset energy consumption simulation evaluation generation network to obtain the preset energy consumption simulation evaluation generator, includes: constructing the objective functions corresponding to the preset energy consumption simulation evaluation generation network and the preset energy consumption simulation evaluation identification network based on the first set loss function corresponding to the first loss and the second set loss function corresponding to the second loss in the objective function; maximizing the first expected value of the first set loss function corresponding to the first loss in the objective function and minimizing the second expected value of the second set loss function corresponding to the second loss in the objective function, and training the preset energy consumption simulation evaluation generation network to obtain the preset energy consumption simulation evaluation generator.

[0087] In the embodiments of this disclosure, the preset energy consumption simulation evaluation network is configured as one or more of support vector regression, support vector machine, K-nearest neighbor algorithm, random forest, artificial neural network, deep neural network, etc.

[0088] In the embodiments of this disclosure and other possible embodiments, the acquisition period or acquisition frequency corresponding to the power consumption detection sensor and / or gas consumption sensor corresponding to each set energy consumption device is obtained; based on the acquisition period or acquisition frequency corresponding to the power consumption detection sensor and / or gas consumption sensor corresponding to each set energy consumption device, power consumption data and / or gas consumption data are collected for each set energy consumption device, and the collected power consumption data and / or gas consumption data corresponding to each set energy consumption device and the acquisition time are sent to the host computer database through the industrial bus.

[0089] In embodiments of this disclosure and other possible embodiments, the energy consumption simulation method based on factory operating status includes: training a preset energy consumption simulation evaluation generation network using the historical usage time, historical electricity consumption data, and / or historical gas consumption data of each of the multiple set energy consumption devices in the factory operating status at a first moment within a set time interval, and the historical energy consumption evaluation corresponding to a second moment after the first moment, to obtain the preset energy consumption simulation evaluation generator; and predicting the energy consumption evaluation corresponding to a fourth moment after the third moment using the current usage time, current electricity consumption data, and / or current gas consumption data of each of the multiple set energy consumption devices in the factory operating status at a third moment, based on the preset energy consumption simulation evaluation generator; and predicting the energy consumption evaluation corresponding to a fourth moment after the third moment using the current usage time, current electricity consumption data, and / or current gas consumption data of each of the multiple set energy consumption devices in the factory operating status at a third moment.

[0090] Step S102: Based on the preset energy consumption simulation evaluation generator, generate a corresponding energy consumption simulation evaluation by using the current usage time, current electricity consumption data and / or current gas consumption data of each of the multiple set energy consumption devices in the factory operation state.

[0091] Figure 2 A schematic diagram of an energy consumption prediction method based on factory operating status according to an embodiment of the present disclosure is shown. Figure 2As shown in the embodiments of this disclosure, an energy consumption prediction method based on factory operating status is also provided, including: Step S201: Using the historical usage time, historical electricity consumption data, and / or historical gas consumption data of each of the multiple set energy consumption devices in the factory operating status at a first moment within a set time interval, and the historical energy consumption evaluation corresponding to a second moment after the first moment, a preset energy consumption simulation evaluation generation network is trained to obtain the preset energy consumption simulation evaluation generator; Step S202: Based on the preset energy consumption simulation evaluation generator, using the current usage time, current electricity consumption data, and / or current gas consumption data of each of the multiple set energy consumption devices in the factory operating status at a third moment, the energy consumption evaluation corresponding to a fourth moment after the third moment is predicted. This solves the technical problem that current energy consumption prediction methods for factory operating status do not consider the impact of the usage time of energy consumption devices on energy consumption evaluation.

[0092] In the embodiments of this disclosure, the step of training a preset energy consumption simulation evaluation generation network using the historical usage time, historical electricity consumption data, and / or historical gas consumption data of each of the multiple set energy consumption devices in the factory operation state at a first moment within a set time interval, and the historical energy consumption evaluation at a second moment after the first moment, to obtain the preset energy consumption simulation evaluation generator, includes: configuring the historical usage time, historical electricity consumption data, and / or historical gas consumption data corresponding to the first moment as a prediction feature vector; configuring the historical energy consumption evaluation at the second moment after the first moment as a prediction training label corresponding to the prediction feature vector; training the preset energy consumption simulation evaluation generation network using the prediction feature vector to obtain a corresponding training process energy consumption simulation evaluation; calculating the loss value corresponding to the training process energy consumption simulation evaluation and the prediction training label, and then training the preset energy consumption simulation evaluation generation network corresponding to the preset energy consumption simulation evaluation generator.

[0093] In the embodiments of this disclosure and other possible embodiments, the second time point can be configured as 72 hours after the first time point; similarly, the third time point can be configured as 72 hours after the fourth time point. Furthermore, based on the preset energy consumption simulation evaluation generator, the energy consumption evaluation corresponding to the fourth time point 72 hours after the third time point is predicted using the current usage time, current electricity consumption data, and / or current gas consumption data of each of the multiple set energy consumption devices under the factory operating state at the third time point.

[0094] In the embodiments of this disclosure, the method of calculating the loss value corresponding to the energy consumption simulation evaluation of the training process and the label, and then training the preset energy consumption simulation evaluation generation network corresponding to the preset energy consumption simulation evaluation generator, includes: if the loss value corresponding to the energy consumption simulation evaluation of the training process and the label is less than or equal to the preset loss value within a set number of training iterations, then the training of the preset energy consumption simulation evaluation generation network corresponding to the preset energy consumption simulation evaluation generator is completed.

[0095] In the embodiments of this disclosure, before training the preset energy consumption simulation evaluation generation network, a connection to a host computer database is established through an industrial bus. The industrial bus is used to obtain from the host computer database the historical usage time, historical electricity consumption data, and / or historical gas consumption data of each of the multiple set energy consumption devices in the factory operating state at the first moment within a set time interval, as well as the historical energy consumption evaluation corresponding to the second moment after the first moment.

[0096] In embodiments of this disclosure, training the preset energy consumption simulation evaluation generation network using the predicted feature vector to obtain the corresponding training process energy consumption simulation evaluation includes: acquiring random noise under the conditions of the historical electricity consumption data and / or historical gas consumption data corresponding to the first time moment; configuring the historical usage time, the historical electricity consumption data and / or historical gas consumption data, and the random noise corresponding to the first time moment as the predicted feature vector; configuring the energy consumption evaluation corresponding to the second time moment as the predicted training label corresponding to the predicted feature vector; and training the preset energy consumption simulation evaluation generation network using the predicted feature vector to obtain the corresponding training process energy consumption simulation evaluation.

[0097] In embodiments of this disclosure, before training the preset energy consumption simulation evaluation generation network using the predicted feature vector, the method includes: standardizing the predicted feature vector to obtain a standardized predicted feature vector; and training the preset energy consumption simulation evaluation generation network using the standardized predicted feature vector.

[0098] In embodiments of this disclosure, calculating the loss value corresponding to the energy consumption simulation evaluation during the training process and the predicted training label includes: based on a preset energy consumption simulation evaluation network, using the historical usage time, the historical electricity consumption data and / or gas consumption data, the energy consumption simulation evaluation, and the energy consumption evaluation, determining whether the energy consumption simulation evaluation is true or false; calculating a first loss of the energy consumption simulation evaluation corresponding to random noise under the conditions of the historical electricity consumption data and / or historical gas consumption data, and a second loss corresponding to determining whether the energy consumption simulation evaluation is true or false, and training the preset energy consumption simulation evaluation generation network.

[0099] In embodiments of this disclosure, the energy consumption simulation evaluation and / or the energy consumption evaluation is configured as one of a high energy consumption level, a medium energy consumption level, a low energy consumption level, and a normal energy consumption level. Similarly, the preset energy consumption simulation evaluation network is configured as one or more of support vector regression, support vector machine, K-nearest neighbor algorithm, random forest, artificial neural network, deep neural network, etc.

[0100] Similarly, in the embodiments of this disclosure and other possible embodiments, the calculation of the first loss of the energy consumption simulation evaluation corresponding to random noise under the conditions of historical usage time, historical electricity consumption data and / or gas consumption data, and the second loss corresponding to the determination of the true or false nature of the energy consumption simulation evaluation, includes: constructing a first set loss function corresponding to the first loss and a second set loss function corresponding to the second loss; calculating the first loss of the energy consumption simulation evaluation corresponding to random noise under the conditions of historical usage time, historical electricity consumption data and / or gas consumption data based on the first set loss function; and calculating the second loss corresponding to the determination of the true or false nature of the energy consumption simulation evaluation corresponding to random noise under the conditions of historical electricity consumption data and / or gas consumption data based on the second set loss function.

[0101] Similarly, in the embodiments of this disclosure and other possible embodiments, constructing the first set loss function corresponding to the first loss includes: calculating the difference between a set value and an energy consumption simulation evaluation corresponding to random noise under the conditions of historical usage time, historical electricity consumption data and / or gas consumption data; taking the logarithm of the difference to obtain a logarithmic value of the difference; calculating a first expected value of the logarithmic value of the difference, and constructing the first set loss function corresponding to the first loss; wherein the set value is configured to be 1.

[0102] Similarly, in the embodiments of this disclosure and other possible embodiments, constructing the first set loss function corresponding to the first loss includes: taking the logarithm of the discrimination probability value corresponding to the random noise under the conditions of the historical usage time, the historical electricity consumption data and / or gas consumption data to determine whether the energy consumption simulation evaluation is true or false, to obtain a discrimination logarithm value; calculating the second expected value of the logarithm value, and constructing the second set loss function corresponding to the second loss. The step of calculating the first loss of the energy consumption simulation evaluation corresponding to random noise under the conditions of the historical usage time, the historical electricity consumption data and / or gas consumption data, and the second loss corresponding to the determination of the true or false nature of the energy consumption simulation evaluation, and training the preset energy consumption simulation evaluation generation network to obtain the preset energy consumption simulation evaluation generator, includes: constructing the objective functions corresponding to the preset energy consumption simulation evaluation generation network and the preset energy consumption simulation evaluation identification network based on the first set loss function corresponding to the first loss and the second set loss function corresponding to the second loss in the objective function; maximizing the first expected value of the first set loss function corresponding to the first loss in the objective function and minimizing the second expected value of the second set loss function corresponding to the second loss in the objective function, and training the preset energy consumption simulation evaluation generation network to obtain the preset energy consumption simulation evaluation generator.

[0103] In the embodiments of this disclosure and other possible embodiments, in order to solve the technical problem that traditional machine learning networks rely on a large amount of usage time, electricity consumption and / or gas consumption data, and when the data on usage time, electricity consumption and / or gas consumption is limited and the distribution of energy consumption evaluations corresponding to usage time, electricity consumption and / or gas consumption is uneven, it is difficult to obtain the optimal energy consumption simulation evaluation and to accurately predict the energy consumption evaluation, a conditional generative adversarial network based on a preset energy consumption simulation evaluation generation network and a preset energy consumption simulation evaluation identification network is proposed.

[0104] In embodiments of this disclosure and other possible embodiments, the step of training a preset energy consumption simulation evaluation generation network using the historical usage time, the historical electricity consumption data and / or gas consumption data, random noise under the conditions of the historical electricity consumption data and / or gas consumption data, the energy consumption evaluation corresponding to the historical usage time, the historical electricity consumption data and / or gas consumption data, and a preset energy consumption simulation evaluation network to obtain a preset energy consumption simulation evaluation generator includes: based on the preset energy consumption simulation evaluation generation network, using the historical usage time, the historical electricity consumption data and / or gas consumption data, and random noise under the conditions of the historical electricity consumption data and / or gas consumption data, the energy consumption evaluation corresponding to the historical usage time, the historical electricity consumption data and / or gas consumption data, and a preset energy consumption simulation evaluation network, to obtain a preset energy consumption simulation evaluation generator. Random noise under the conditions of electricity consumption data and / or gas consumption data is used to generate an energy consumption simulation evaluation; based on the preset energy consumption simulation evaluation network, the true or false nature of the energy consumption simulation evaluation is determined by using the historical usage time, the historical electricity consumption data and / or gas consumption data, the energy consumption simulation evaluation, and the energy consumption evaluation; the first loss of the energy consumption simulation evaluation corresponding to the random noise under the conditions of the historical usage time, the historical electricity consumption data and / or gas consumption data, and the second loss corresponding to the determination of the true or false nature of the energy consumption simulation evaluation are calculated, and the preset energy consumption simulation evaluation generation network is trained to obtain the preset energy consumption simulation evaluation generator.

[0105] In embodiments of this disclosure and other possible embodiments, the calculation of the first loss of the energy consumption simulation evaluation corresponding to random noise under the conditions of historical usage time, historical electricity consumption data and / or gas consumption data, and the second loss corresponding to the determination of the true or false nature of the energy consumption simulation evaluation, includes: constructing a first set loss function E[log(1-D(G(z|y)))] corresponding to the first loss and a second set loss function E[logD(x|y)] corresponding to the second loss; calculating the first loss of the energy consumption simulation evaluation corresponding to random noise under the conditions of historical usage time, historical electricity consumption data and / or gas consumption data based on the first set loss function E[log(1-D(G(z|y)))]; and calculating the second loss corresponding to the determination of the true or false nature of the energy consumption simulation evaluation based on the second set loss function E[logD(x|y)]. Wherein, G represents the preset energy consumption simulation evaluation generation network, D represents the preset energy consumption simulation evaluation identification network, z represents random noise, D(G(z|y)) is the judgment probability corresponding to the generated energy consumption simulation evaluation; E represents expectation; the known condition y corresponds to the historical usage time, historical electricity consumption data and / or gas consumption data, and x represents the actual energy consumption evaluation corresponding to the energy consumption simulation evaluation.

[0106] In embodiments of this disclosure and other possible embodiments, constructing the first set loss function corresponding to the first loss includes: calculating the difference 1-D(G(z|y)) between a set value and the energy consumption simulation evaluation corresponding to random noise under the conditions of historical usage time, historical electricity consumption data and / or gas consumption data; taking the logarithm of the difference to obtain the logarithmic difference value log(1-D(G(z|y))); calculating the first expected value E[log(1-D(G(z|y)))] of the logarithmic difference value log(1-D(G(z|y))); and constructing the first set loss function corresponding to the first loss; wherein the set value is configured to be 1.

[0107] In the embodiments of this disclosure and other possible embodiments, constructing the first set loss function corresponding to the first loss includes: taking the logarithm of the discrimination probability value D(x|y) corresponding to the random noise under the conditions of the historical usage time, the historical electricity consumption data and / or gas consumption data to determine whether the energy consumption simulation evaluation is true or false, to obtain the discrimination logarithm value logD(x|y); calculating the second expected value E[logD(x|y)] of the logarithm value, and constructing the second set loss function corresponding to the second loss.

[0108] In embodiments of this disclosure and other possible embodiments, the step of calculating the first loss of the energy consumption simulation evaluation corresponding to random noise under the conditions of historical usage time, historical electricity consumption data and / or gas consumption data, and the second loss corresponding to the determination of the true or false nature of the energy consumption simulation evaluation, and training a preset energy consumption simulation evaluation generation network to obtain a preset energy consumption simulation evaluation generator, includes: constructing an objective function E[logD(x|y)]+E[log(1-D(G(z|y)))] corresponding to the preset energy consumption simulation evaluation generation network and the preset energy consumption simulation evaluation identification network based on the first set loss function corresponding to the first loss and the second set loss function corresponding to the second loss in the objective function; maximizing the first expected value of the first set loss function corresponding to the first loss in the objective function and minimizing the second expected value of the second set loss function corresponding to the second loss in the objective function, that is:

[0109] m G inm D The preset energy consumption simulation evaluation generator is obtained by training the preset energy consumption simulation evaluation generation network with the formula axE[logD(x|y)]+E[log(1-D(G(z|y)))].

[0110] In the embodiments of this disclosure and other possible embodiments, known prior conditions corresponding to random noise under the conditions of historical usage time, historical electricity consumption data and / or gas consumption data are added to control the generation process of the conditional generative adversarial network based on the preset energy consumption simulation evaluation generation network and the preset energy consumption simulation evaluation identification network.

[0111] In embodiments of this disclosure and other possible embodiments, the preset energy consumption simulation evaluation generation network includes: a first energy consumption evaluation shallow feature extraction module, an encoder connected to the shallow feature extraction module, and a decoder connected to the shallow feature extraction module and the encoder respectively; wherein, the shallow feature extraction module is used to extract the first energy consumption evaluation shallow features of the historical usage time, the historical electricity consumption data and / or gas consumption data, and random noise under the conditions of the historical usage time, the historical electricity consumption data and / or gas consumption data; the encoder is used to extract the first energy consumption evaluation deep features and the second energy consumption evaluation deep features corresponding to the first energy consumption evaluation shallow features; the decoder is used to generate an energy consumption simulation evaluation based on the first energy consumption evaluation shallow features and their corresponding first energy consumption evaluation deep features and the second energy consumption evaluation deep features.

[0112] In embodiments of this disclosure and other possible embodiments, the shallow feature extraction module includes: a first predefined convolutional neural unit corresponding to a first energy consumption evaluation shallow feature for extracting the historical usage time, the historical electricity consumption data and / or gas consumption data, and random noise under the conditions of the historical electricity consumption data and / or gas consumption data. The first predefined convolutional neural unit includes at least one one-dimensional convolutional layer or multiple cascaded one-dimensional convolutional layers.

[0113] In embodiments of this disclosure and other possible embodiments, the encoder includes: a first dense feature extraction module and a second dense feature extraction module connected to the first dense feature extraction module; wherein, the first dense feature extraction module includes: a second set convolutional neural unit, a third set convolutional neural unit connected to the second set convolutional neural unit, a first set batch normalization unit connected to the third set convolutional neural unit, and a first set activation function connected to the first set batch normalization unit; wherein, the input configuration of the second set convolutional neural unit is the first energy consumption evaluation shallow feature, and the output configuration of the first set activation function is the first energy consumption evaluation deep feature; the second dense feature extraction module includes: a fourth set convolutional neural unit, a fifth set convolutional neural unit connected to the fourth set convolutional neural unit, and a third set convolutional neural unit connected to the fifth set batch normalization unit. The system defines a second set batch normalization unit connected to a set convolutional neural unit and a second set activation function connected to the second set batch normalization unit; a first skip connection unit is configured between the input of the fourth set convolutional neural unit, the input of the second set convolutional neural unit, and the output of the third set convolutional neural unit; the first skip connection unit is used to concatenate the first energy consumption evaluation shallow features and the processing features output by the third set convolutional neural unit to obtain a first concatenated feature; the first concatenated feature is configured as the input of the fourth set convolutional neural unit, and the output of the second set activation function is configured as the second energy consumption evaluation deep features; wherein, the second set convolutional neural unit, the third set convolutional neural unit, the fourth set convolutional neural unit, and the fifth set convolutional neural unit include at least one one-dimensional convolutional layer or multiple cascaded one-dimensional convolutional layers.

[0114] In embodiments of this disclosure and other possible embodiments, a second skip connection unit is configured between the input of the second set convolutional neural unit and the input of the third set convolutional neural unit; wherein, the second skip connection unit is used to splice the first energy consumption evaluation shallow feature and the processing feature output by the second set convolutional neural unit to obtain a second spliced ​​feature; the second spliced ​​feature is configured as the input of the third set convolutional neural unit.

[0115] In embodiments of this disclosure and other possible embodiments, a third skip connection unit is configured between the input of the fourth set convolutional neural unit and the input of the fifth set convolutional neural unit; wherein, the third skip connection unit is used to splice the splicing feature with the processing feature output by the fourth set convolutional neural unit to obtain a third splicing feature; the third splicing feature is configured as the input of the fifth set convolutional neural unit.

[0116] In embodiments of this disclosure and other possible embodiments, the decoder includes: a sixth predefined convolutional neural unit, a feature self-attention module connected to the sixth predefined convolutional neural unit, and a linear layer connected to the feature self-attention module; wherein, the sixth predefined convolutional neural unit is used to perform convolution processing on the first energy consumption evaluation shallow features to obtain second energy consumption evaluation shallow features; the feature self-attention module is used to perform self-attention processing on the second energy consumption evaluation shallow features, the first energy consumption evaluation deep features, and the second energy consumption evaluation deep features to obtain self-attention features; the linear layer is used to perform linear processing on the self-attention features to generate the energy consumption simulation evaluation; wherein, the sixth predefined convolutional neural unit includes at least one one-dimensional convolutional layer or multiple cascaded one-dimensional convolutional layers.

[0117] In embodiments of this disclosure and other possible embodiments, the feature self-attention module includes: a channel attention module and a spatial attention module connected to the channel attention module; wherein, the channel attention module is used to perform channel attention processing on the second energy consumption evaluation shallow features, the first energy consumption evaluation deep features and the second energy consumption evaluation deep features to obtain channel attention features; the spatial attention module is used to perform spatial attention processing on the spatial attention features to obtain self-attention features.

[0118] In the embodiments of this disclosure and other possible embodiments, before the channel attention module performs channel attention processing on the second energy consumption evaluation shallow features, the first energy consumption evaluation deep features, and the second energy consumption evaluation deep features, the second energy consumption evaluation shallow features, the first energy consumption evaluation deep features, and the second energy consumption evaluation deep features are spliced ​​together to obtain spliced ​​features to be processed; the channel attention module performs channel attention processing on the spliced ​​features to be processed to obtain channel attention features.

[0119] In embodiments of this disclosure and other possible embodiments, the channel attention module includes: a first pooling unit and a second pooling unit different from the first pooling unit, a shared multilayer perceptron connected to the first pooling unit and the second pooling unit respectively, an adder connected to the shared multilayer perceptron, a third set activation function connected to the adder, and a first multiplier; wherein, the first pooling unit and the second pooling unit respectively perform pooling operations on the shallow features of the second energy consumption evaluation, the deep features of the first energy consumption evaluation, and the unprocessed splicing features corresponding to the deep features of the second energy consumption evaluation, to obtain the first pooling feature and the second pooling feature. The shared multilayer perceptron is used to perform weight processing on the first pooling feature and the second pooling feature respectively to obtain a first weighted feature and a second weighted feature; the adder is used to perform addition processing on the first weighted feature and the second weighted feature to obtain an additive feature; the third set activation function is used to perform nonlinear processing on the additive feature to obtain a channel attention weight; the first multiplier is used to perform multiplication operation processing on the channel attention weight and the spliced ​​feature to be processed to obtain a channel attention feature; wherein, the first pooling unit is configured as a max pooling layer and the second pooling unit is configured as a mean pooling layer.

[0120] In embodiments of this disclosure and other possible embodiments, the spatial attention module includes: a third pooling unit, a fourth pooling unit connected to the third pooling unit, a convolutional layer connected to the fourth pooling unit, a fourth set activation function connected to the convolutional layer, and a second multiplier; wherein, the third pooling unit and the fourth pooling unit are sequentially used to perform pooling processing on the channel attention features to obtain third pooling features; the convolutional layer is used to perform convolution processing on the third pooling features to obtain convolutional features; the fourth set activation function is used to perform nonlinear processing on the convolutional features to obtain spatial attention weights; the second multiplier is used to perform multiplication operations on the channel attention features and the spatial attention weights to obtain self-attention features; wherein, the third pooling unit is configured as a max pooling layer, and the fourth pooling unit is configured as a mean pooling layer.

[0121] In embodiments of this disclosure and other possible embodiments, the linear layer connected to the feature self-attention module includes: a first linear layer and a second linear layer connected to the first linear layer; wherein, the second linear layer is used to perform linear processing on the linear features after linear processing of the self-attention features by the first linear layer, and then perform further linear processing to generate the energy consumption simulation evaluation.

[0122] In the embodiments of this disclosure and other possible embodiments, the preset energy consumption simulation evaluation network includes: at least one convolutional correlation weight module or multiple cascaded convolutional correlation weight modules; wherein, the convolutional correlation weight module is used to train the preset energy consumption simulation evaluation generation network using the historical usage time, the historical electricity consumption data and / or gas consumption data, random noise under the conditions of the historical electricity consumption data and / or gas consumption data, and the energy consumption evaluation corresponding to the historical electricity consumption data and / or gas consumption data; wherein, the number of the multiple cascaded convolutional correlation weight modules can be configured to be 3.

[0123] In embodiments of this disclosure and other possible embodiments, each of the convolutional correlation weight modules includes: a seventh set convolutional neural unit, a feature self-attention unit connected to the seventh set convolutional neural unit, a third set batch normalization unit connected to the feature self-attention unit, and a fifth set activation function connected to the third set batch normalization unit; wherein, the seventh set convolutional neural unit is used to perform correlation weight features on the historical usage time, the historical electricity consumption data and / or gas consumption data, the energy consumption evaluation and reservoir quality pseudo-evaluation corresponding to the historical usage time and the historical electricity consumption data and / or gas consumption data, or the correlation weight features output by the previous level convolutional correlation weight module. Convolution processing; obtaining the convolutional features to be discriminated; the feature self-attention unit is used to perform self-attention processing on the convolutional features to be discriminated, to obtain the self-attention features to be discriminated; the third set batch normalization unit is used to perform batch normalization processing on the self-attention features to be discriminated, to obtain the batch normalized features to be discriminated; the fifth set activation function is used to perform nonlinear processing on the batch normalized features to be discriminated, to obtain the discrimination result corresponding to each convolutional correlation weight module; wherein, the seventh set convolutional neural unit includes at least one one-dimensional convolutional layer or multiple cascaded one-dimensional convolutional layers; wherein, the discrimination result is configured as true or false corresponding to the energy consumption simulation evaluation and / or the energy consumption evaluation.

[0124] In the embodiments of this disclosure and other possible embodiments, a skip connection unit is provided between the inputs of each of the convolutional correlation weight modules; wherein, the skip connection unit is used to concatenate the discrimination result corresponding to the output of the previous level convolutional correlation weight module with the input features of the previous level convolutional correlation weight module to obtain the correlation weight concatenation features of the next level convolutional correlation weight module.

[0125] In the embodiments of this disclosure and other possible embodiments, feature extraction is better achieved from the historical usage time, the historical electricity consumption data and / or gas consumption data, and random noise under the conditions of the historical electricity consumption data and / or gas consumption data. Specifically, a shallow energy consumption evaluation feature is first obtained using a first set convolutional neural unit. This shallow energy consumption evaluation feature is then input into a dense feature extraction module to enhance the feature learning ability on limited data, resulting in a deep energy consumption evaluation feature, thus encoding the input feature. Next, the shallow and deep energy consumption evaluation features are fused using a hybrid attention module. The fused feature is then used to generate and predict energy consumption through two linear network layers, thus decoding the input feature. Furthermore, a three-layer set convolutional neural unit and a feature self-attention mechanism are used to learn features from the input data. The original data and the extracted features are then concatenated, and finally, two linear layers are used to determine whether the input data comes from a preset energy consumption simulation evaluation generator or real data. Finally, through continuous adversarial interaction between the preset energy consumption simulation evaluation generator and the preset energy consumption simulation evaluation identification network model, the preset energy consumption simulation evaluation generator gains better data generation capabilities, thus enabling the evaluation of unknown energy consumption quality. Specifically, the preset energy consumption simulation evaluation identification network classifies real energy consumption evaluations as true and identifies generated energy consumption simulation evaluations that are opposite to real energy consumption evaluations as false.

[0126] In the embodiments of this disclosure and other possible embodiments, during the training process, the preset energy consumption simulation evaluation generator and the preset energy consumption simulation evaluation identification network need to be optimized and trained simultaneously. They promote and improve each other through adversarial game theory. Specifically, the goal of the preset energy consumption simulation evaluation generator is to continuously optimize the network parameters and improve the quality of the generated energy consumption simulation evaluations, making it impossible for the preset energy consumption simulation evaluation identification network to distinguish between true and false evaluations. In other words, it aims to maximize the probability that the preset energy consumption simulation evaluation identification network classifies the generated energy consumption simulation evaluations as true. Meanwhile, the preset energy consumption simulation evaluation identification network needs to continuously minimize the classification error rate of the input data, that is, maximize the difference between its probability of classifying true energy consumption evaluations and its probability of classifying generated data.

[0127] In the embodiments of this disclosure and other possible embodiments, by jointly optimizing the objective function of the preset energy consumption simulation evaluation generator and the preset energy consumption simulation evaluation identification network, the training objective is to find a Nash equilibrium point of the preset energy consumption simulation evaluation generator and the preset energy consumption simulation evaluation identification network, so that the energy consumption simulation evaluation generated by the preset energy consumption simulation evaluation generator is sufficiently realistic, and the preset energy consumption simulation evaluation identification network cannot distinguish between the real energy consumption evaluation and the generated energy consumption simulation evaluation.

[0128] In the embodiments of this disclosure and other possible embodiments, during the training process, the preset energy consumption simulation evaluation generator and the preset energy consumption simulation evaluation identification network are updated alternately. First, the preset energy consumption simulation evaluation identification network is trained using real data and generated data, and its parameters are updated to better distinguish between real energy consumption evaluations and generated energy consumption simulation evaluations. Second, the parameters of the preset energy consumption simulation evaluation identification network are fixed, while the parameters of the preset energy consumption simulation evaluation generator are updated, enabling the generator to produce more realistic energy consumption evaluations and increasing the probability that the preset energy consumption simulation evaluation identification network classifies the generated data as real. Finally, through continuous alternating updates and optimizations, the error rate of the preset energy consumption simulation evaluation identification network is increased, the generator can estimate the distribution of energy consumption simulation evaluations, and the generated energy consumption simulation evaluations become more realistic.

[0129] In the embodiments of this disclosure and other possible embodiments, conditional generative adversarial networks (GANs) can effectively utilize given conditional constraint information, transforming the probabilities of generative adversarial networks into conditional probabilities. During the implementation using conditional GANs, the historical usage time, historical electricity consumption data, and / or gas consumption data corresponding to the conditions in the conditional probabilities are output as energy consumption evaluations.

[0130] In the embodiments of this disclosure and other possible embodiments, in order to better apply conditional generative adversarial networks directly to the energy consumption evaluation process, that is, under the condition that each of the multiple set energy consumption devices in a given factory operating state has historical usage time, historical electricity consumption data and / or historical gas consumption data at least at a set time interval within a set time interval, the model can give the corresponding energy consumption evaluation.

[0131] In embodiments of this disclosure and other possible embodiments, the preset energy consumption simulation evaluation generation network includes an encoder and a decoder to perform dense feature extraction. The dense feature extraction preset energy consumption simulation evaluation generator model takes potential noise and historical usage time, historical electricity consumption data, and / or gas consumption data as constraints as input. For the historical usage time, the historical electricity consumption data, and / or gas consumption data, and random noise under the conditions of the historical electricity consumption data and / or gas consumption data, a first unencoded shallow energy consumption evaluation feature is obtained using a single layer of predefined convolutional neural units (the first predefined convolutional neural unit). This first shallow energy consumption evaluation feature is then input into the designed dense feature extraction module to obtain encoded first and second deep energy consumption evaluation features, thus achieving the encoding of the input features. Then, the unencoded shallow features of the first energy consumption evaluation are extracted again through a set convolutional neural unit, and then fused with the deep features of the first and second energy consumption evaluations through an attention module. The fused features are then used to generate and predict energy consumption through two cascaded linear network layers, corresponding to the first linear layer and the second linear layer, to decode the input features and obtain the simulated energy consumption evaluation.

[0132] In the embodiments of this disclosure and other possible embodiments, the entire encoder mainly consists of two cascaded energy consumption evaluation dense feature modules. Each energy consumption evaluation dense feature module includes: a second and third set convolutional neural unit corresponding to two layers of set convolutional neural units, or a fourth and fifth set convolutional neural unit, a first set batch normalization unit or a second set batch normalization unit, and a first or second set activation function ReLU corresponding to a set activation function layer. To effectively achieve dense extraction between data and enhance feature extraction effect, the module establishes a second or third skip connection unit corresponding to a residual link between the two layers of set convolutional neural units. The input features and the output features of the second or fourth set convolutional neural unit corresponding to the first layer of set convolutional neural units are concatenated in the channel dimension and used together as the input of the dense module to the third or fifth set convolutional neural unit corresponding to the second layer of set convolutional neural units. After being processed by the third or fifth predefined convolutional neural unit corresponding to the predefined convolutional neural unit, the features sequentially pass through the first or second predefined batch normalization unit corresponding to the BN layer and the first or second predefined activation function corresponding to the ReLU layer. The first or second predefined batch normalization unit corresponding to the BN layer normalizes the feature distribution. Finally, the output of a single energy consumption evaluation dense feature module is obtained.

[0133] In the embodiments of this disclosure and other possible embodiments, to further achieve dense feature extraction, data transfer is also implemented in the same way in the two dense modules. The output features of the previous energy consumption evaluation dense feature module are used as the input of the next energy consumption evaluation dense feature module by feature concatenation (a second skip connection unit is configured between the input of the second set convolutional neural unit and the input of the third set convolutional neural unit; wherein, the second skip connection unit is used to concatenate the first energy consumption evaluation shallow features with the processing features output by the second set convolutional neural unit to obtain a second concatenated feature; the second concatenated feature is configured as the input of the third set convolutional neural unit). This cascaded feature transfer allows subsequent modules to directly utilize the feature information extracted by the previous modules, significantly improving the feature utilization rate, and also enabling the fusion of multi-level features, allowing the network to simultaneously utilize shallow local features and deep global features. Finally, the outputs of the two energy consumption evaluation dense feature modules are concatenated together to provide a portion of the input for the subsequent decoder (a first skip connection unit is configured between the input of the fourth set convolutional neural unit, the input of the second set convolutional neural unit, and the output of the third set convolutional neural unit; the first skip connection unit is used to concatenate the first energy consumption evaluation shallow features and the processing features output by the third set convolutional neural unit to obtain a first concatenated feature; the first concatenated feature is configured as the input of the fourth set convolutional neural unit).

[0134] In the embodiments of this disclosure and other possible embodiments, for the decoder, the first shallow energy consumption evaluation features extracted by the first set convolutional neural unit corresponding to the set convolutional neural unit are input into the encoder and the decoder. The first shallow energy consumption evaluation features are then processed again by the sixth set convolutional neural unit corresponding to a set convolutional neural unit to extract features, resulting in the second shallow energy consumption evaluation features. Subsequently, the extracted second shallow energy consumption evaluation features, along with the first and second deep energy consumption evaluation features from the encoder, are fused together by the convolutional attention mechanism module to improve the model's ability to focus on features at different levels, achieving effective feature fusion. Finally, the fused features are sequentially input into the first linear layer and the second linear layer corresponding to the two linear network layers to generate and predict the preset energy consumption simulation evaluation, thus obtaining the energy consumption simulation evaluation.

[0135] In the embodiments of this disclosure and other possible embodiments, the convolutional attention mechanism module performs attention mechanism operations in the channel and spatial dimensions, making the network pay more attention to important channel features and spatial location features, suppressing unwanted channel and spatial location features, and increasing the network's ability to capture input features.

[0136] In the embodiments of this disclosure and other possible embodiments, channel attention is a method that compresses the spatial dimension through pooling layers, focusing on the importance of each feature map channel, aiming to enhance the feature representation ability between different channels. First, the channel attention weights are calculated by using max pooling corresponding to the first pooling unit and average pooling corresponding to the second pooling unit to calculate the maximum feature corresponding to the first pooling feature and the average feature corresponding to the second pooling feature for each channel. Second, the obtained feature vectors are passed through a shared multilayer perceptron to learn the first weight feature and the second weight feature corresponding to the weight parameters of each channel. Finally, the channel attention weights are obtained by normalization using a third activation function σ (such as Sigmoid), and multiplied by the concatenated feature obtained by concatenating the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature, and the second energy consumption evaluation deep feature to achieve weight adjustment for different channels.

[0137] In the embodiments of this disclosure and other possible embodiments, spatial attention is an important method for enhancing the network's representation ability in the spatial dimension. It primarily identifies important features at each spatial location, compensating for the spatial limitations of channel attention. First, the channel attention features are compressed using average pooling corresponding to the fourth pooling unit and max pooling corresponding to the third pooling unit, resulting in third and fourth pooling features. Then, the third and fourth pooling features obtained from average and max pooling are concatenated to obtain the third pooling feature; alternatively, the third and fourth pooling units are sequentially used to pool the channel attention features to obtain the third pooling feature. Finally, the channel information is fused using 7*7 convolutional kernels corresponding to the convolutional layers to obtain convolutional features.

[0138] In the embodiments of this disclosure and other possible embodiments, the attention module takes as input the concatenated features obtained by stitching together the first and second deep energy consumption evaluation features from the encoder and the second shallow energy consumption evaluation features extracted from the decoder (i.e., the concatenated features to be processed obtained by stitching together the second shallow energy consumption evaluation features, the first deep energy consumption evaluation features, and the second deep energy consumption evaluation features). First, channel attention is used to calculate the weights of different channels, obtaining channel attention weights, which are then multiplied with the input features to obtain channel attention features. Then, using these as input, the weights of different spatial locations are calculated to obtain spatial attention weights, which are then multiplied with the concatenated features to be processed, completing the final self-attention features. Using the attention module enables the effective fusion of the second shallow energy consumption evaluation features, the first deep energy consumption evaluation features, and the second deep energy consumption evaluation features, completing the extraction of the importance of features at different levels and the extraction of information at different spatial locations. This enhances the overall representational capability of the preset energy consumption simulation evaluation generation network, enabling the generation of high-quality energy consumption simulation evaluations, thereby achieving accurate energy consumption evaluation.

[0139] In embodiments of this disclosure and other possible embodiments, a preset energy consumption simulation evaluation network is included, comprising: one layer of predefined convolutional neural units and two linear connection layers configured within the convolutional correlation weight module, i.e., at least one linear layer is provided after at least one convolutional correlation weight module or multiple cascaded convolutional correlation weight modules. For example, two linear layers are provided after at least one convolutional correlation weight module or multiple cascaded convolutional correlation weight modules, respectively configured as a third linear layer and a fourth linear layer connected to the third linear layer. Within the convolutional correlation weight module, after the seventh predefined convolutional neural unit corresponding to each predefined convolutional neural unit, a feature self-attention module corresponding to the feature self-attention unit, a third predefined batch normalization unit, and a ReLU predefined activation function (the fifth predefined activation function) are sequentially added. The feature self-attention unit enhances the expressive power of important features by calculating the correlation weights between features; the third predefined batch normalization unit is used to stabilize the feature distribution and accelerate convergence; and the fifth predefined activation function introduces nonlinear transformation capability.

[0140] In the embodiments of this disclosure and other possible embodiments, the preset energy consumption simulation evaluation network adopts a dual-input design. The first input is configured as a first joint feature obtained by concatenating the real energy consumption evaluation with the energy consumption evaluation corresponding to the known conditions, namely the historical usage time, historical electricity consumption data and / or gas consumption data, and the energy consumption evaluation corresponding to the historical electricity consumption data and / or gas consumption data. The other input is a second joint feature obtained by concatenating the energy consumption simulation evaluation generated by the preset energy consumption simulation evaluation generator with the energy consumption simulation evaluation corresponding to the known conditions, namely the historical usage time, historical electricity consumption data and / or gas consumption data, and the energy consumption simulation evaluation corresponding to the historical electricity consumption data and / or gas consumption data. During training, after the first joint feature passes through the preset energy consumption simulation evaluation network, its discrimination probability value of being judged as a real sample is calculated using a binary classification cross-entropy loss function. At the same time, the second joint feature is also processed by the same preset energy consumption simulation evaluation network, and its discrimination probability value of being identified as a generated sample is calculated using a cross-entropy loss function. This adversarial training mechanism forces the preset energy consumption simulation evaluation network to continuously improve its discrimination ability, and also promotes the preset energy consumption simulation evaluation generation network to improve its generation quality.

[0141] In the embodiments of this disclosure and other possible embodiments, there is no strict distinction between the historical usage time, historical electricity consumption data and / or historical gas consumption data corresponding to at least one collection moment within a set time interval and their corresponding historical energy consumption evaluations, and the historical usage time, historical electricity consumption data and / or historical gas consumption data corresponding to the first moment within the set time interval and the historical energy consumption evaluations corresponding to the second moment after the first moment. However, the process of training the preset energy consumption simulation evaluation generation network to obtain the preset energy consumption simulation evaluation generator in the energy consumption simulation method and energy consumption prediction method based on factory operating status corresponding to the embodiments of this disclosure is the same.

[0142] In the embodiments of this disclosure and other possible embodiments, the energy consumption prediction method based on factory operating status further includes: training a preset energy consumption simulation evaluation generation network using historical usage time, historical electricity consumption data and / or historical gas consumption data corresponding to at least one collection moment within a set time interval for each of a plurality of set energy consumption devices under factory operating status, and their corresponding historical energy consumption evaluations, to obtain the preset energy consumption simulation evaluation generator; and generating a corresponding energy consumption simulation evaluation based on the preset energy consumption simulation evaluation generator using the current usage time, current electricity consumption data and / or current gas consumption data corresponding to each of the plurality of set energy consumption devices under factory operating status.

[0143] The entity executing the energy consumption simulation and prediction method based on factory operating status can be an energy consumption simulation and prediction device or system based on factory operating status. For example, the energy consumption simulation and prediction method based on factory operating status can be executed by a terminal device, server, or other processing device. The terminal device can be user equipment (UE), mobile device, user terminal, terminal, cellular phone, cordless phone, personal digital assistant (PDA), handheld device, computing device, in-vehicle device, wearable device, etc. In some possible implementations, the energy consumption simulation and prediction method based on factory operating status can be implemented by a processor calling computer-readable instructions stored in memory.

[0144] Those skilled in the art will understand that in the above-described method for simulating and predicting energy consumption based on factory operating status in specific implementations, the order in which each step is written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0145] Figure 3 A schematic diagram of an energy consumption simulation system based on factory operating conditions according to an embodiment of the present disclosure is shown. Figure 3 As shown, embodiments of this disclosure also provide an energy consumption simulation system based on factory operating status, comprising: a first training unit 101, used to train a preset energy consumption simulation evaluation generation network by utilizing historical usage time, historical electricity consumption data and / or historical gas consumption data corresponding to at least one collection moment within a set time interval for each of a plurality of set energy consumption devices under factory operating status, and their corresponding historical energy consumption evaluations, to obtain the preset energy consumption simulation evaluation generator; and a simulation evaluation unit 102, used to generate a corresponding energy consumption simulation evaluation based on the preset energy consumption simulation evaluation generator, utilizing the current usage time, current electricity consumption data and / or current gas consumption data corresponding to each of the plurality of set energy consumption devices under factory operating status.

[0146] Embodiments of this disclosure also provide an energy consumption simulation system based on factory operating status, including: an electronic device, the electronic device being configured with a processor and a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the above-described energy consumption simulation method based on factory operating status.

[0147] Embodiments of this disclosure also provide an energy consumption simulation system based on factory operating status, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to perform the above-described energy consumption simulation method based on factory operating status.

[0148] Embodiments of this disclosure also provide an energy consumption simulation system based on factory operating status, comprising: a computer program product, wherein the computer program product is provided with a computer program / instruction, which, when executed by a processor, implements the energy consumption simulation method based on factory operating status as described above.

[0149] The embodiments of this disclosure also provide an energy consumption simulation device based on factory operating status, comprising: a first training unit, used to train a preset energy consumption simulation evaluation generation network by using historical usage time, historical electricity consumption data and / or historical gas consumption data corresponding to at least one collection moment within a set time interval for each of a plurality of set energy consumption devices under factory operating status, and their corresponding historical energy consumption evaluations, to obtain the preset energy consumption simulation evaluation generator; and a simulation evaluation unit, used to generate a corresponding energy consumption simulation evaluation based on the preset energy consumption simulation evaluation generator, using the current usage time, current electricity consumption data and / or current gas consumption data corresponding to each of the plurality of set energy consumption devices under factory operating status.

[0150] Embodiments of this disclosure also provide an energy consumption simulation device based on factory operating status, comprising: an electronic device, the electronic device being configured with a processor and a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the above-described energy consumption simulation method based on factory operating status.

[0151] Embodiments of this disclosure also provide an energy consumption simulation device based on factory operating status, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to perform the above-described energy consumption simulation method based on factory operating status.

[0152] Embodiments of this disclosure also provide an energy consumption simulation device based on factory operating status, comprising: a computer program product, wherein the computer program product is provided with a computer program / instruction, which, when executed by a processor, implements the energy consumption simulation method based on factory operating status as described above.

[0153] Figure 4 A schematic diagram of an energy consumption prediction system based on factory operating status according to an embodiment of the present disclosure is shown. Figure 4As shown, embodiments of this disclosure also provide an energy consumption prediction system based on factory operating status, comprising: a second training unit 201, used to train a preset energy consumption simulation evaluation generation network by utilizing the historical usage time, historical electricity consumption data and / or historical gas consumption data of each of the multiple set energy consumption devices in the factory operating status at a first moment within a set time interval, and the historical energy consumption evaluation corresponding to a second moment after the first moment, to obtain the preset energy consumption simulation evaluation generator; and an energy consumption evaluation prediction unit 202, used to predict the energy consumption evaluation corresponding to a fourth moment after the third moment based on the preset energy consumption simulation evaluation generator and utilizing the current usage time, current electricity consumption data and / or current gas consumption data of each of the multiple set energy consumption devices in the factory operating status at a third moment.

[0154] Embodiments of this disclosure also provide an energy consumption prediction system based on factory operating status, including: an electronic device, the electronic device being configured with a processor and a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the above-described energy consumption prediction method based on factory operating status.

[0155] Embodiments of this disclosure also provide an energy consumption prediction system based on factory operating status, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the above-described energy consumption prediction method based on factory operating status.

[0156] Embodiments of this disclosure also provide an energy consumption prediction system based on factory operating status, comprising: a computer program product, wherein the computer program product is provided with a computer program / instruction, which, when executed by a processor, implements the energy consumption prediction method based on factory operating status as described above.

[0157] The embodiments of this disclosure also provide an energy consumption prediction device based on factory operating status, comprising: a second training unit, configured to train a preset energy consumption simulation evaluation generation network using historical usage time, historical electricity consumption data and / or historical gas consumption data of each of a plurality of set energy consumption devices in the factory operating status at a first moment within a set time interval, and historical energy consumption evaluation at a second moment after the first moment, to obtain the preset energy consumption simulation evaluation generator; and an energy consumption evaluation prediction unit, configured to predict the energy consumption evaluation at a fourth moment after the third moment based on the preset energy consumption simulation evaluation generator, using the current usage time, current electricity consumption data and / or current gas consumption data of each of the plurality of set energy consumption devices in the factory operating status at a third moment.

[0158] Embodiments of this disclosure also provide an energy consumption prediction device based on factory operating status, comprising: an electronic device, the electronic device being configured with a processor and a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the above-described energy consumption prediction method based on factory operating status.

[0159] Embodiments of this disclosure also provide an energy consumption prediction device based on factory operating status, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the above-described energy consumption prediction method based on factory operating status.

[0160] Embodiments of this disclosure also provide an energy consumption prediction device based on factory operating status, comprising: a computer program product, wherein the computer program product is provided with a computer program / instruction, which, when executed by a processor, implements the energy consumption prediction method based on factory operating status as described above.

[0161] Embodiments of this disclosure also provide a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, implement the above-described energy consumption simulation method based on factory operating conditions.

[0162] Embodiments of this disclosure also provide a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, implement the above-described energy consumption prediction method based on factory operating status.

[0163] Embodiments of this disclosure also provide a computer-readable storage medium storing computer program instructions thereon, which, when executed by a processor, implement the above-described energy consumption simulation method based on factory operating status and the above-described energy consumption prediction method based on factory operating status.

[0164] In some embodiments, the functions or modules of the apparatus provided in this disclosure can be used to execute the energy consumption simulation and prediction method based on factory operating status described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0165] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0166] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0167] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0168] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0169] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or technical improvements to the embodiments in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for energy consumption simulation based on plant operating state, characterized by, The method comprises the following steps: training a preset energy consumption simulation evaluation generator by using historical use time, historical power consumption data and / or historical gas consumption data corresponding to at least one collection time of each of a plurality of set energy consumption devices in a factory running state within a set time interval and historical energy consumption evaluation thereof; wherein the preset energy consumption simulation evaluation generation network comprises a first energy consumption evaluation shallow feature extraction module, an encoder connected to the first energy consumption evaluation shallow feature extraction module, and a decoder connected to the first energy consumption evaluation shallow feature extraction module and the encoder; the first set of convolutional neural units corresponding to the shallow feature extraction module at least includes one one-dimensional convolutional layer or a plurality of cascaded one-dimensional convolutional layers, which are used to extract first energy consumption evaluation shallow features of the historical use time, the historical power consumption data and / or the gas consumption data and random noise under the condition of the historical use time, the historical power consumption data and / or the gas consumption data; the encoder is used to extract first energy consumption evaluation deep features and second energy consumption evaluation deep features corresponding to the first energy consumption evaluation shallow features; the decoder is used to generate an energy consumption simulation evaluation based on the first energy consumption evaluation shallow features, the first energy consumption evaluation deep features and the second energy consumption evaluation deep features corresponding thereto; the decoder comprises a sixth set of convolutional neural units, a feature self-attention module connected to the sixth set of convolutional neural units, and a linear layer connected to the feature self-attention module; wherein the one one-dimensional convolutional layer or the plurality of cascaded one-dimensional convolutional layers of the sixth set of convolutional neural units are used to perform convolution processing on the first energy consumption evaluation shallow features to obtain second energy consumption evaluation shallow features; the feature self-attention module is used to perform self-attention processing on the second energy consumption evaluation shallow features, the first energy consumption evaluation deep features and the second energy consumption evaluation deep features to obtain self-attention features; the linear layer is used to perform linear processing on the self-attention features to generate the energy consumption simulation evaluation; the energy consumption simulation evaluation is configured as one of high energy consumption level, medium energy consumption level, low energy consumption level and normal energy consumption level; based on the preset energy consumption simulation evaluation generator, generating a corresponding energy consumption simulation evaluation by using current use time, current power consumption data and / or current gas consumption data corresponding to each of a plurality of set energy consumption devices in the factory running state.

2. The energy consumption simulation method based on a factory operation state according to claim 1, characterized by, Before training the preset energy consumption simulation evaluation generation network by using historical use time, historical power consumption data and / or historical gas consumption data corresponding to at least one collection time of each of a plurality of set energy consumption devices in a factory running state within a set time interval and historical energy consumption evaluation thereof, comprising: establishing a connection between the industrial bus and the host computer database; The industrial bus is used to obtain historical use time, historical power consumption data and / or historical gas consumption data corresponding to at least one collection time in a set time interval of each set energy consumption device in the plurality of set energy consumption devices under the factory operation state and historical energy consumption evaluation corresponding thereto from the host computer database.

3. The method of claim 1 or 2, wherein the method is characterized by, The historical use time, the historical power consumption data and / or the historical gas consumption data corresponding to at least one collection time in a set time interval of each set energy consumption device in the plurality of set energy consumption devices under the factory operation state and the historical energy consumption evaluation corresponding thereto are used to train a preset energy consumption simulation evaluation generation network to obtain the preset energy consumption simulation evaluation generator, including: Obtaining random noise under the condition of the historical power consumption data and / or the historical gas consumption data; The historical use time, the historical power consumption data and / or the historical gas consumption data, and the random noise are configured as a classification feature vector; The energy consumption evaluation corresponding to the historical use time, the historical power consumption data and / or the historical gas consumption data, and the random noise is configured as a classification training label corresponding to the classification feature vector; The classification feature vector is used to train the preset energy consumption simulation evaluation generation network to obtain a training process energy consumption simulation evaluation.

4. The energy consumption simulation method based on a factory operation state according to claim 3, characterized by, Before the classification feature vector is used to train the preset energy consumption simulation evaluation generation network, including: The classification feature vector is standardized to obtain a standardized classification feature vector; The standardized classification feature vector is used to train the preset energy consumption simulation evaluation generation network.

5. The energy consumption simulation method based on a factory operation state according to claim 3, characterized by, The historical use time, the historical power consumption data and / or the historical gas consumption data corresponding to at least one collection time in a set time interval of each set energy consumption device in the plurality of set energy consumption devices under the factory operation state and the historical energy consumption evaluation corresponding thereto are used to train a preset energy consumption simulation evaluation generation network to obtain the preset energy consumption simulation evaluation generator, further including: Based on a preset energy consumption simulation evaluation discrimination network, the historical use time, the historical power consumption data and / or the gas consumption data, the energy consumption simulation evaluation and the energy consumption evaluation are used to discriminate whether the energy consumption simulation evaluation is true or false; The first loss of the energy consumption simulation evaluation corresponding to the random noise under the condition of the historical power consumption data and / or historical gas consumption data and the second loss of discriminating the true or false correspondence of the energy consumption simulation evaluation are calculated, and a preset energy consumption simulation evaluation generation network is trained, including: a first setting loss function corresponding to the first loss is constructed and a second setting loss function corresponding to the second loss is constructed ; the first loss of the energy consumption simulation evaluation corresponding to the random noise under the condition of the historical use time, the historical power consumption data and / or gas consumption data is calculated based on the first setting loss function , and the second loss of discriminating the true or false correspondence of the energy consumption simulation evaluation corresponding to the random noise under the condition of the historical use time, the historical power consumption data and / or gas consumption data is calculated based on the second setting loss function , and the preset energy consumption simulation evaluation generation network is trained; wherein, G represents a preset energy consumption simulation evaluation generation network, D represents a preset energy consumption simulation evaluation discrimination network, z represents a random noise, is a judgment probability corresponding to the generated energy consumption simulation evaluation; E represents expectation; the known prior condition y corresponds to the historical use time, the historical power consumption data and / or gas consumption data, x represents a real energy consumption evaluation corresponding to the energy consumption simulation evaluation.

6. The energy consumption simulation method based on a factory operation state according to claim 4, characterized by, The historical use time, the historical power consumption data and / or the historical gas consumption data corresponding to at least one collection time in a set time interval of each set energy consumption device in the plurality of set energy consumption devices under the factory operation state and the historical energy consumption evaluation corresponding thereto are used to train a preset energy consumption simulation evaluation generation network to obtain the preset energy consumption simulation evaluation generator, further including: Based on a preset energy consumption simulation evaluation discrimination network, the historical use time, the historical power consumption data and / or the gas consumption data, the energy consumption simulation evaluation and the energy consumption evaluation are used to discriminate whether the energy consumption simulation evaluation is true or false; The first loss of the energy consumption simulation evaluation corresponding to the random noise under the condition of the historical power consumption data and / or historical gas consumption data and the second loss of discriminating the true or false correspondence of the energy consumption simulation evaluation are calculated, and a preset energy consumption simulation evaluation generation network is trained, including: a first setting loss function corresponding to the first loss is constructed and a second setting loss function corresponding to the second loss is constructed; based on the first setting loss function , the first loss of the energy consumption simulation evaluation corresponding to the random noise under the condition of the historical use time, the historical power consumption data and / or gas consumption data is calculated; based on the second setting loss function , the second loss of discriminating the true or false correspondence of the energy consumption simulation evaluation corresponding to the random noise under the condition of the historical use time, the historical power consumption data and / or gas consumption data is calculated, and the preset energy consumption simulation evaluation generation network is trained; wherein, G represents a preset energy consumption simulation evaluation generation network, D represents a preset energy consumption simulation evaluation discrimination network, z represents a random noise, is a judgment probability corresponding to the generated energy consumption simulation evaluation; E represents expectation; the known prior condition y corresponds to the historical use time, the historical power consumption data and / or gas consumption data, x represents a real energy consumption evaluation corresponding to the energy consumption simulation evaluation.

7. The method of claim 1, 2, 4-6, wherein, The encoder comprises a first dense feature extraction module and a second dense feature extraction module connected with the first dense feature extraction module; the first dense feature extraction module comprises a second set of convolutional neural units, a third set of convolutional neural units connected with the second set of convolutional neural units, a first set of batch normalization units connected with the third set of convolutional neural units, and a first set of activation functions connected with the first set of batch normalization units; the input of the second set of convolutional neural units is configured as the first energy consumption evaluation shallow layer feature, and the output of the first set of activation functions is configured as the first energy consumption evaluation deep layer feature; the second dense feature extraction module comprises a fourth set of convolutional neural units, a fifth set of convolutional neural units connected with the fourth set of convolutional neural units, a second set of batch normalization units connected with the fifth set of convolutional neural units, and a second set of activation functions connected with the second set of batch normalization units; the input of the fourth set of convolutional neural units is configured with a first jump connection unit between the input of the second set of convolutional neural units and the output of the third set of convolutional neural units; the first jump connection unit is used for splicing the first energy consumption evaluation shallow layer feature and the processing feature output by the third set of convolutional neural units to obtain a first spliced feature; the first spliced feature is configured as the input of the fourth set of convolutional neural units, and the output of the second set of activation functions is configured as the second energy consumption evaluation deep layer feature; wherein the second set of convolutional neural units, the third set of convolutional neural units, the fourth set of convolutional neural units, and the fifth set of convolutional neural units at least comprise one one-dimensional convolution layer or multiple cascaded one-dimensional convolution layers; the input of the second set of convolutional neural units is configured with a second jump connection unit between the input of the third set of convolutional neural units; wherein the second jump connection unit is used for splicing the first energy consumption evaluation shallow layer feature and the processing feature output by the second set of convolutional neural units to obtain a second spliced feature; the second spliced feature is configured as the input of the third set of convolutional neural units; the input of the fourth set of convolutional neural units is configured with a third jump connection unit between the input of the fifth set of convolutional neural units; wherein the third jump connection unit is used for splicing the spliced feature and the processing feature output by the fourth set of convolutional neural units to obtain a third spliced feature; the third spliced feature is configured as the input of the fifth set of convolutional neural units.

8. The energy consumption simulation method based on the factory running state according to claim 3, wherein The encoder comprises a first dense feature extraction module and a second dense feature extraction module connected with the first dense feature extraction module; the first dense feature extraction module comprises a second set of convolutional neural units, a third set of convolutional neural units connected with the second set of convolutional neural units, a first set of batch normalization units connected with the third set of convolutional neural units, and a first set of activation functions connected with the first set of batch normalization units; the input of the second set of convolutional neural units is configured as the first energy consumption evaluation shallow layer feature, and the output of the first set of activation functions is configured as the first energy consumption evaluation deep layer feature; the second dense feature extraction module comprises a fourth set of convolutional neural units, a fifth set of convolutional neural units connected with the fourth set of convolutional neural units, a second set of batch normalization units connected with the fifth set of convolutional neural units, and a second set of activation functions connected with the second set of batch normalization units; the input of the fourth set of convolutional neural units is configured with a first jump connection unit between the input of the second set of convolutional neural units and the output of the third set of convolutional neural units; the first jump connection unit is used for splicing the first energy consumption evaluation shallow layer feature and the processing feature output by the third set of convolutional neural units to obtain a first spliced feature; the first spliced feature is configured as the input of the fourth set of convolutional neural units, and the output of the second set of activation functions is configured as the second energy consumption evaluation deep layer feature; wherein the second set of convolutional neural units, the third set of convolutional neural units, the fourth set of convolutional neural units, and the fifth set of convolutional neural units at least comprise one one-dimensional convolution layer or multiple cascaded one-dimensional convolution layers; the input of the second set of convolutional neural units is configured with a second jump connection unit between the input of the third set of convolutional neural units; wherein the second jump connection unit is used for splicing the first energy consumption evaluation shallow layer feature and the processing feature output by the second set of convolutional neural units to obtain a second spliced feature; the second spliced feature is configured as the input of the third set of convolutional neural units; the input of the fourth set of convolutional neural units is configured with a third jump connection unit between the input of the fifth set of convolutional neural units; wherein the third jump connection unit is used for splicing the spliced feature and the processing feature output by the fourth set of convolutional neural units to obtain a third spliced feature; the third spliced feature is configured as the input of the fifth set of convolutional neural units.

9. The method of claim 1, 2, 4-6, 8, wherein, The corresponding power consumption detection sensor and / or gas consumption sensor are arranged on the energy consumption setting device. ​ Corresponding power consumption detection sensors and / or gas consumption detection sensors are arranged for each set energy consumption device, and historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are collected; an industrial bus connected with the power consumption detection sensors and / or gas consumption detection sensors arranged for each set energy consumption device is used to send the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data to an upper computer database, and the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are stored in the upper computer database; And / or, The feature self-attention module comprises a channel attention module and a spatial attention module connected with the channel attention module; the channel attention module is configured to perform channel attention processing on the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature to obtain a channel attention feature; the spatial attention module is configured to perform spatial attention processing on a spatial attention feature to obtain a self-attention feature; before the channel attention module performs channel attention processing on the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature, the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature are spliced to obtain a to-be-processed spliced feature; and the channel attention module is configured to perform channel attention processing on the to-be-processed spliced feature to obtain a channel attention feature.

10. The energy consumption simulation method based on a factory operation state according to claim 3, characterized by, Comprise: Corresponding power consumption detection sensors and / or gas consumption sensors are arranged for each set energy consumption device; Corresponding power consumption detection sensors and / or gas consumption detection sensors are arranged for each set energy consumption device, and historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are collected; an industrial bus connected with the power consumption detection sensors and / or gas consumption detection sensors arranged for each set energy consumption device is used to send the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data to an upper computer database, and the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are stored in the upper computer database; And / or, The feature self-attention module comprises a channel attention module and a spatial attention module connected with the channel attention module; the channel attention module is configured to perform channel attention processing on the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature to obtain a channel attention feature; the spatial attention module is configured to perform spatial attention processing on a spatial attention feature to obtain a self-attention feature; before the channel attention module performs channel attention processing on the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature, the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature are spliced to obtain a to-be-processed spliced feature; and the channel attention module is configured to perform channel attention processing on the to-be-processed spliced feature to obtain a channel attention feature.

11. The energy consumption simulation method based on a factory operation state according to claim 7, wherein Comprise: corresponding power consumption detection sensors and / or gas consumption sensors are arranged for each set energy consumption device; corresponding historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are collected by using the power consumption detection sensors and / or gas consumption detection sensors arranged for each set energy consumption device; the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are transmitted to an upper computer database by using an industrial bus connected with the power consumption detection sensors and / or gas consumption detection sensors arranged for each set energy consumption device; and the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are stored by using the upper computer database; and / or The feature self-attention module comprises a channel attention module and a spatial attention module connected with the channel attention module; the channel attention module is configured to perform channel attention processing on the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature to obtain a channel attention feature; the spatial attention module is configured to perform spatial attention processing on a spatial attention feature to obtain a self-attention feature; before the channel attention module performs channel attention processing on the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature, the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature are spliced to obtain a to-be-processed spliced feature; and the channel attention module is configured to perform channel attention processing on the to-be-processed spliced feature to obtain a channel attention feature.

12. The method of claim 1, 2, 4-6, 8, 10, 11, wherein, Comprise: The usage time of each set energy consumption device from turning on to turning off is timed by using a timer to obtain the corresponding usage time of each set energy consumption device; The use time corresponding to each set energy consumption device is sent to an upper computer database through an industrial bus connected with the timer, and the use time corresponding to each set energy consumption device is stored by using the upper computer database; And / or, The channel attention module of the feature self-attention module comprises a first pooling unit and a second pooling unit different from the first pooling unit, a shared multi-layer perceptron connected with the first pooling unit and the second pooling unit respectively, an adder connected with the shared multi-layer perceptron, a third set activation function and a first multiplier connected with the adder; wherein the first pooling unit and the second pooling unit perform pooling operation on the to-be-processed spliced features corresponding to the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature respectively to obtain first pooled features and second pooled features; the shared multi-layer perceptron is used for performing weight processing on the first pooled features and the second pooled features respectively to obtain first weight features and second weight features; the adder is used for performing addition processing on the first weight features and the second weight features to obtain addition features; the third set activation function is used for performing nonlinear processing on the addition features to obtain channel attention weights; and the first multiplier is used for performing multiplication operation processing on the channel attention weights and the to-be-processed spliced features to obtain channel attention features.

13. The energy consumption simulation method based on a factory operation state according to claim 3, characterized by, Comprise: The use time corresponding to each set energy consumption device is sent to an upper computer database through an industrial bus connected with the timer, and the use time corresponding to each set energy consumption device is stored by using the upper computer database; The use time corresponding to each set energy consumption device is sent to an upper computer database through an industrial bus connected with the timer, and the use time corresponding to each set energy consumption device is stored by using the upper computer database; And / or, The channel attention module of the feature self-attention module comprises: a first pooling unit and a second pooling unit different from the first pooling unit, a shared multi-layer perceptron connected with the first pooling unit and the second pooling unit respectively, an adder connected with the shared multi-layer perceptron, a third preset activation function connected with the adder, and a first multiplier; wherein the first pooling unit and the second pooling unit perform a pooling operation on the to-be-processed spliced features corresponding to the second energy consumption evaluation shallow layer feature, the first energy consumption evaluation deep layer feature and the second energy consumption evaluation deep layer feature respectively to obtain first pooled features and second pooled features; the shared multi-layer perceptron is configured to perform weight processing on the first pooled features and the second pooled features respectively to obtain first weight features and second weight features; the adder is configured to perform addition processing on the first weight features and the second weight features to obtain addition features; the third preset activation function is configured to perform nonlinear processing on the addition features to obtain channel attention weights; and the first multiplier is configured to perform multiplication operation processing on the channel attention weights and the to-be-processed spliced features to obtain channel attention features.

14. The energy consumption simulation method based on a factory operation state according to claim 7, wherein Comprise: using a timer, timing the use time of each set energy consumption device from opening to closing respectively to obtain the use time corresponding to each set energy consumption device; sending the use time corresponding to each set energy consumption device to an upper computer database through an industrial bus connected with the timer, and storing the use time corresponding to each set energy consumption device by using the upper computer database; and / or, The channel attention module of the feature self-attention module comprises: a first pooling unit and a second pooling unit different from the first pooling unit, a shared multi-layer perceptron connected with the first pooling unit and the second pooling unit respectively, an adder connected with the shared multi-layer perceptron, a third preset activation function connected with the adder, and a first multiplier; wherein the first pooling unit and the second pooling unit perform a pooling operation on the to-be-processed spliced features corresponding to the second energy consumption evaluation shallow layer feature, the first energy consumption evaluation deep layer feature and the second energy consumption evaluation deep layer feature respectively to obtain first pooled features and second pooled features; the shared multi-layer perceptron is configured to perform weight processing on the first pooled features and the second pooled features respectively to obtain first weight features and second weight features; the adder is configured to perform addition processing on the first weight features and the second weight features to obtain addition features; the third preset activation function is configured to perform nonlinear processing on the addition features to obtain channel attention weights; and the first multiplier is configured to perform multiplication operation processing on the channel attention weights and the to-be-processed spliced features to obtain channel attention features.

15. The method of claim 9, wherein, Comprise: using a timer, timing the use time of each set energy consumption device from opening to closing respectively to obtain the use time corresponding to each set energy consumption device; The use time corresponding to each set energy consumption device is transmitted to an upper computer database through an industrial bus connected with the timer, and the use time corresponding to each set energy consumption device is stored by using the upper computer database; And / or, The channel attention module of the feature self-attention module comprises a first pooling unit and a second pooling unit different from the first pooling unit, a shared multi-layer perceptron connected with the first pooling unit and the second pooling unit respectively, an adder connected with the shared multi-layer perceptron, a third set activation function and a first multiplier connected with the adder; wherein the first pooling unit and the second pooling unit perform a pooling operation on the to-be-processed spliced features corresponding to the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature respectively to obtain first pooled features and second pooled features; the shared multi-layer perceptron is configured to perform weight processing on the first pooled features and the second pooled features respectively to obtain first weight features and second weight features; the adder is configured to perform addition processing on the first weight features and the second weight features to obtain addition features; the third set activation function is configured to perform nonlinear processing on the addition features to obtain channel attention weights; and the first multiplier is configured to perform multiplication operation processing on the channel attention weights and the to-be-processed spliced features to obtain channel attention features.

16. A method for energy consumption prediction based on plant operating state, characterized by, Comprise: The preset energy consumption simulation evaluation generator is obtained by training a preset energy consumption simulation evaluation generation network using historical use time, historical power consumption data and / or historical gas consumption data of each of a plurality of set energy consumption devices in a factory operating state at a first time corresponding to a set time interval and historical energy consumption evaluation at a second time after the first time; wherein the preset energy consumption simulation evaluation generation network comprises: a first energy consumption evaluation shallow feature extraction module, an encoder connected to the first energy consumption evaluation shallow feature extraction module, and a decoder connected to the first energy consumption evaluation shallow feature extraction module and the encoder, respectively; the first set of convolutional neural units corresponding to the shallow feature extraction module includes at least one one-dimensional convolutional layer or a plurality of cascaded one-dimensional convolutional layers, which are used to extract first energy consumption evaluation shallow features of the historical use time, the historical power consumption data and / or the historical gas consumption data, and random noise under the conditions of the historical use time, the historical power consumption data and / or the historical gas consumption data; the encoder is used to extract first energy consumption evaluation deep features and second energy consumption evaluation deep features corresponding to the first energy consumption evaluation shallow features; the decoder is used to generate an energy consumption simulation evaluation based on the first energy consumption evaluation shallow features, the first energy consumption evaluation deep features and the second energy consumption evaluation deep features corresponding thereto; the decoder comprises a sixth set of convolutional neural units, a feature self-attention module connected to the sixth set of convolutional neural units, and a linear layer connected to the feature self-attention module; wherein the one-dimensional convolutional layer or the plurality of cascaded one-dimensional convolutional layers corresponding to the sixth set of convolutional neural units are used to perform convolution processing on the first energy consumption evaluation shallow features to obtain second energy consumption evaluation shallow features; the feature self-attention module is used to perform self-attention processing on the second energy consumption evaluation shallow features, the first energy consumption evaluation deep features and the second energy consumption evaluation deep features to obtain self-attention features; the linear layer is used to perform linear processing on the self-attention features to generate the energy consumption simulation evaluation; the energy consumption simulation evaluation is configured as one of high energy consumption level, medium energy consumption level, low energy consumption level and normal energy consumption level; Based on the preset energy consumption simulation evaluation generator, the energy consumption evaluation corresponding to a fourth time after a third time is predicted using current use time, current power consumption data and / or current gas consumption data of each of a plurality of set energy consumption devices in the factory operating state at the third time.

17. The method of claim 16, wherein, The preset energy consumption simulation evaluation generator is obtained by training a preset energy consumption simulation evaluation generation network using historical use time, historical power consumption data and / or historical gas consumption data of each of a plurality of set energy consumption devices in a factory operating state at a first time corresponding to a set time interval and historical energy consumption evaluation at a second time after the first time; wherein the preset energy consumption simulation evaluation generation network comprises: a first energy consumption evaluation shallow feature extraction module, an encoder connected to the first energy consumption evaluation shallow feature extraction module, and a decoder connected to the first energy consumption evaluation shallow feature extraction module and the encoder, respectively; the first set of convolutional neural units corresponding to the shallow feature extraction module includes at least one one-dimensional convolutional layer or a plurality of cascaded one-dimensional convolutional layers, which are used to extract first energy consumption evaluation shallow features of the historical use time, the historical power consumption data and / or the historical gas consumption data, and random noise under the conditions of the historical use time, the historical power consumption data and / or the historical gas consumption data; the encoder is used to extract first energy consumption evaluation deep features and second energy consumption evaluation deep features corresponding to the first energy consumption evaluation shallow features; the decoder is used to generate an energy consumption simulation evaluation based on the first energy consumption evaluation shallow features, the first energy consumption evaluation deep features and the second energy consumption evaluation deep features corresponding thereto; the decoder comprises a sixth set of convolutional neural units, a feature self-attention module connected to the sixth set of convolutional neural units, and a linear layer connected to the feature self-attention module; wherein the one-dimensional convolutional layer or the plurality of cascaded one-dimensional convolutional layers corresponding to the sixth set of convolutional neural units are used to perform convolution processing on the first energy consumption evaluation shallow features to obtain second energy consumption evaluation shallow features; the feature self-attention module is used to perform self-attention processing on the second energy consumption evaluation shallow features, the first energy consumption evaluation deep features and the second energy consumption evaluation deep features to obtain self-attention features; the linear layer is used to perform linear processing on the self-attention features to generate the energy consumption simulation evaluation; the energy consumption simulation evaluation is configured as one of high energy consumption level, medium energy consumption level, low energy consumption level and normal energy consumption level; Based on the preset energy consumption simulation evaluation generator, the energy consumption evaluation corresponding to a fourth time after a third time is predicted using current use time, current power consumption data and / or current gas consumption data of each of a plurality of set energy consumption devices in the factory operating state at the third time. corresponding to the first moment, the historical use time, the historical power consumption data and / or historical gas consumption data, and the random noise are configured as a prediction feature vector; corresponding to the second moment is configured as a prediction training label corresponding to the prediction feature vector; The preset energy consumption simulation evaluation generation network is trained using the prediction feature vector, and a corresponding training process energy consumption evaluation is obtained.

18. The method of energy consumption prediction based on plant operating state according to any one of claims 16 or 17, characterized in that, Before the preset energy consumption simulation evaluation generation network is trained using the historical use time, the historical power consumption data and / or historical gas consumption data corresponding to the first moment of each of the plurality of set energy consumption devices in the utilization factory running state within the set time interval and the historical energy consumption evaluation corresponding to the second moment after the first moment, the method comprises: establishing a connection between an industrial bus and an upper computer database; The historical use time, the historical power consumption data and / or historical gas consumption data corresponding to the first moment of each of the plurality of set energy consumption devices in the utilization factory running state within the set time interval and the historical energy consumption evaluation corresponding to the second moment after the first moment are obtained from the upper computer database using the industrial bus.

19. The method of claim 17, wherein, Before the preset energy consumption simulation evaluation generation network is trained using the prediction feature vector, the method comprises: The prediction feature vector is standardized to obtain a standardized prediction feature vector, and the preset energy consumption simulation evaluation generation network is trained using the standardized prediction feature vector.

20. The method of claim 16, 17, or 19, wherein, The preset energy consumption simulation evaluation generation network is trained using the historical use time, the historical power consumption data and / or historical gas consumption data corresponding to the first moment of each of the plurality of set energy consumption devices in the utilization factory running state within the set time interval and the historical energy consumption evaluation corresponding to the second moment after the first moment, and the preset energy consumption simulation evaluation generator is obtained. Based on the preset energy consumption simulation evaluation discrimination network, the historical use time, the historical power consumption data and / or gas consumption data, the energy consumption simulation evaluation, and the energy consumption evaluation are used to discriminate whether the energy consumption simulation evaluation is true or false. The first loss of the energy consumption simulation evaluation corresponding to the random noise under the condition of the historical power consumption data and / or historical gas consumption data and the second loss of discriminating the true or false correspondence of the energy consumption simulation evaluation are calculated, and a preset energy consumption simulation evaluation generation network is trained, including: a first setting loss function corresponding to the first loss is constructed and a second setting loss function corresponding to the second loss ; the first loss of the energy consumption simulation evaluation corresponding to the random noise under the condition of the historical use time, the historical power consumption data and / or gas consumption data is calculated based on the first setting loss function , and the second loss of discriminating the true or false correspondence of the energy consumption simulation evaluation corresponding to the random noise under the condition of the historical use time, the historical power consumption data and / or gas consumption data is calculated based on the second setting loss function , and the preset energy consumption simulation evaluation generation network is trained; wherein, G represents a preset energy consumption simulation evaluation generation network, D represents a preset energy consumption simulation evaluation discrimination network, z represents a random noise, is a judgment probability corresponding to the generated energy consumption simulation evaluation; E represents expectation; the known prior condition y corresponds to the historical use time, the historical power consumption data and / or gas consumption data, x represents a real energy consumption evaluation corresponding to the energy consumption simulation evaluation.

21. The method of claim 18, wherein, The preset energy consumption simulation evaluation generation network is trained using the historical use time, the historical power consumption data and / or historical gas consumption data corresponding to the first moment of each of the plurality of set energy consumption devices in the utilization factory running state within the set time interval and the historical energy consumption evaluation corresponding to the second moment after the first moment, and the preset energy consumption simulation evaluation generator is obtained. Based on the preset energy consumption simulation evaluation discrimination network, the historical use time, the historical power consumption data and / or gas consumption data, the energy consumption simulation evaluation, and the energy consumption evaluation are used to discriminate whether the energy consumption simulation evaluation is true or false. The first loss of the energy consumption simulation evaluation corresponding to the random noise under the condition of the historical power consumption data and / or historical gas consumption data and the second loss of discriminating the true or false correspondence of the energy consumption simulation evaluation are calculated, and a preset energy consumption simulation evaluation generation network is trained, including: a first setting loss function corresponding to the first loss is constructed and a second setting loss function corresponding to the second loss ; the first loss of the energy consumption simulation evaluation corresponding to the random noise under the condition of the historical use time, the historical power consumption data and / or gas consumption data is calculated based on the first setting loss function , and the second loss of discriminating the true or false correspondence of the energy consumption simulation evaluation corresponding to the random noise under the condition of the historical use time, the historical power consumption data and / or gas consumption data is calculated based on the second setting loss function , and the preset energy consumption simulation evaluation generation network is trained; wherein, G represents a preset energy consumption simulation evaluation generation network, D represents a preset energy consumption simulation evaluation discrimination network, z represents a random noise, is a judgment probability corresponding to the generated energy consumption simulation evaluation; E represents expectation; the known prior condition y corresponds to the historical use time, the historical power consumption data and / or gas consumption data x represents a real energy consumption evaluation corresponding to the energy consumption simulation evaluation.

22. The method of energy consumption prediction based on plant operating state according to any one of claims 16, 17, 19, 21, characterized in that, The encoder comprises a first dense feature extraction module and a second dense feature extraction module connected with the first dense feature extraction module; the first dense feature extraction module comprises a second set of convolutional neural units, a third set of convolutional neural units connected with the second set of convolutional neural units, a first set of batch normalization units connected with the third set of convolutional neural units, and a first set of activation functions connected with the first set of batch normalization units; the input of the second set of convolutional neural units is configured as the first energy consumption evaluation shallow layer feature, and the output of the first set of activation functions is configured as the first energy consumption evaluation deep layer feature; the second dense feature extraction module comprises a fourth set of convolutional neural units, a fifth set of convolutional neural units connected with the fourth set of convolutional neural units, a second set of batch normalization units connected with the fifth set of convolutional neural units, and a second set of activation functions connected with the second set of batch normalization units; the input of the fourth set of convolutional neural units is configured with a first jump connection unit between the input of the second set of convolutional neural units and the output of the third set of convolutional neural units; the first jump connection unit is used for splicing the first energy consumption evaluation shallow layer feature and the processing feature output by the third set of convolutional neural units to obtain a first spliced feature; the first spliced feature is configured as the input of the fourth set of convolutional neural units, and the output of the second set of activation functions is configured as the second energy consumption evaluation deep layer feature; wherein the second set of convolutional neural units, the third set of convolutional neural units, the fourth set of convolutional neural units, and the fifth set of convolutional neural units at least comprise one one-dimensional convolution layer or multiple cascaded one-dimensional convolution layers; the input of the second set of convolutional neural units is configured with a second jump connection unit between the input of the third set of convolutional neural units; wherein the second jump connection unit is used for splicing the first energy consumption evaluation shallow layer feature and the processing feature output by the second set of convolutional neural units to obtain a second spliced feature; the second spliced feature is configured as the input of the third set of convolutional neural units; the input of the fourth set of convolutional neural units is configured with a third jump connection unit between the input of the fifth set of convolutional neural units; wherein the third jump connection unit is used for splicing the spliced feature and the processing feature output by the fourth set of convolutional neural units to obtain a third spliced feature; the third spliced feature is configured as the input of the fifth set of convolutional neural units.

23. The method of claim 18, wherein, The encoder comprises a first dense feature extraction module and a second dense feature extraction module connected with the first dense feature extraction module; the first dense feature extraction module comprises a second set of convolutional neural units, a third set of convolutional neural units connected with the second set of convolutional neural units, a first set of batch normalization units connected with the third set of convolutional neural units, and a first set of activation functions connected with the first set of batch normalization units; the input of the second set of convolutional neural units is configured as the first energy consumption evaluation shallow layer feature, and the output of the first set of activation functions is configured as the first energy consumption evaluation deep layer feature; the second dense feature extraction module comprises a fourth set of convolutional neural units, a fifth set of convolutional neural units connected with the fourth set of convolutional neural units, a second set of batch normalization units connected with the fifth set of convolutional neural units, and a second set of activation functions connected with the second set of batch normalization units; the input of the fourth set of convolutional neural units is configured with a first jump connection unit between the input of the second set of convolutional neural units and the output of the third set of convolutional neural units; the first jump connection unit is used for splicing the first energy consumption evaluation shallow layer feature and the processing feature output by the third set of convolutional neural units to obtain a first spliced feature; the first spliced feature is configured as the input of the fourth set of convolutional neural units, and the output of the second set of activation functions is configured as the second energy consumption evaluation deep layer feature; wherein the second set of convolutional neural units, the third set of convolutional neural units, the fourth set of convolutional neural units, and the fifth set of convolutional neural units at least comprise one one-dimensional convolution layer or multiple cascaded one-dimensional convolution layers; the input of the second set of convolutional neural units is configured with a second jump connection unit between the input of the third set of convolutional neural units; wherein the second jump connection unit is used for splicing the first energy consumption evaluation shallow layer feature and the processing feature output by the second set of convolutional neural units to obtain a second spliced feature; the second spliced feature is configured as the input of the third set of convolutional neural units; the input of the fourth set of convolutional neural units is configured with a third jump connection unit between the input of the fifth set of convolutional neural units; wherein the third jump connection unit is used for splicing the spliced feature and the processing feature output by the fourth set of convolutional neural units to obtain a third spliced feature; the third spliced feature is configured as the input of the fifth set of convolutional neural units.

24. The method of claim 20, wherein, The encoder comprises a first dense feature extraction module and a second dense feature extraction module connected with the first dense feature extraction module; the first dense feature extraction module comprises a second set of convolutional neural units, a third set of convolutional neural units connected with the second set of convolutional neural units, a first set of batch normalization units connected with the third set of convolutional neural units, and a first set of activation functions connected with the first set of batch normalization units; the input of the second set of convolutional neural units is configured as the first energy consumption evaluation shallow layer feature, and the output of the first set of activation functions is configured as the first energy consumption evaluation deep layer feature; the second dense feature extraction module comprises a fourth set of convolutional neural units, a fifth set of convolutional neural units connected with the fourth set of convolutional neural units, a second set of batch normalization units connected with the fifth set of convolutional neural units, and a second set of activation functions connected with the second set of batch normalization units; the input of the fourth set of convolutional neural units is configured with a first jump connection unit between the input of the second set of convolutional neural units and the output of the third set of convolutional neural units; the first jump connection unit is used for splicing the first energy consumption evaluation shallow layer feature and the processing feature output by the third set of convolutional neural units to obtain a first spliced feature; the first spliced feature is configured as the input of the fourth set of convolutional neural units, and the output of the second set of activation functions is configured as the second energy consumption evaluation deep layer feature; wherein the second set of convolutional neural units, the third set of convolutional neural units, the fourth set of convolutional neural units, and the fifth set of convolutional neural units at least comprise one one-dimensional convolution layer or multiple cascaded one-dimensional convolution layers; the input of the second set of convolutional neural units is configured with a second jump connection unit between the input of the third set of convolutional neural units; wherein the second jump connection unit is used for splicing the first energy consumption evaluation shallow layer feature and the processing feature output by the second set of convolutional neural units to obtain a second spliced feature; the second spliced feature is configured as the input of the third set of convolutional neural units; the input of the fourth set of convolutional neural units is configured with a third jump connection unit between the input of the fifth set of convolutional neural units; wherein the third jump connection unit is used for splicing the spliced feature and the processing feature output by the fourth set of convolutional neural units to obtain a third spliced feature; the third spliced feature is configured as the input of the fifth set of convolutional neural units.

25. The method of energy consumption prediction based on plant operating state according to any one of claims 16, 17, 19, 21, 23, 24, characterized in that, The corresponding power consumption detection sensor and / or gas consumption sensor are arranged on the energy consumption setting device. ​ Corresponding power consumption detection sensors and / or gas consumption detection sensors are arranged at each of the set energy consumption equipment settings, historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are collected, the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are transmitted to the upper computer database through the industrial bus connected with the power consumption detection sensors and / or gas consumption detection sensors arranged at each of the set energy consumption equipment settings, and the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are stored in the upper computer database; And / or, The feature self-attention module comprises a channel attention module and a spatial attention module connected with the channel attention module; the channel attention module is configured to perform channel attention processing on the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature to obtain a channel attention feature; the spatial attention module is configured to perform spatial attention processing on the spatial attention feature to obtain a self-attention feature; before the channel attention module performs channel attention processing on the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature, the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature are spliced to obtain a to-be-processed spliced feature; and the channel attention module is configured to perform channel attention processing on the to-be-processed spliced feature to obtain a channel attention feature.

26. The method of claim 18, wherein, Comprise: Corresponding power consumption detection sensors and / or gas consumption sensors are arranged at each of the set energy consumption equipment settings; Corresponding power consumption detection sensors and / or gas consumption detection sensors are arranged at each of the set energy consumption equipment settings, historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are collected, the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are transmitted to the upper computer database through the industrial bus connected with the power consumption detection sensors and / or gas consumption detection sensors arranged at each of the set energy consumption equipment settings, and the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are stored in the upper computer database; And / or, The feature self-attention module comprises a channel attention module and a spatial attention module connected with the channel attention module; the channel attention module is configured to perform channel attention processing on the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature to obtain a channel attention feature; the spatial attention module is configured to perform spatial attention processing on a spatial attention feature to obtain a self-attention feature; before the channel attention module performs channel attention processing on the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature, the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature are spliced to obtain a to-be-processed spliced feature; and the channel attention module is configured to perform channel attention processing on the to-be-processed spliced feature to obtain a channel attention feature.

27. The method of claim 20, wherein, Comprise: corresponding power consumption detection sensors and / or gas consumption sensors are arranged for each set energy consumption device; corresponding historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are collected by using the power consumption detection sensors and / or gas consumption detection sensors arranged for each set energy consumption device; the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are transmitted to an upper computer database by using an industrial bus connected with the power consumption detection sensors and / or gas consumption detection sensors arranged for each set energy consumption device; the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are stored by using the upper computer database; and / or The feature self-attention module comprises a channel attention module and a spatial attention module connected with the channel attention module; the channel attention module is configured to perform channel attention processing on the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature to obtain a channel attention feature; the spatial attention module is configured to perform spatial attention processing on a spatial attention feature to obtain a self-attention feature; before the channel attention module performs channel attention processing on the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature, the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature are spliced to obtain a to-be-processed spliced feature; and the channel attention module is configured to perform channel attention processing on the to-be-processed spliced feature to obtain a channel attention feature. 28.The method of claim 22, wherein, Comprise: corresponding power consumption detection sensors and / or gas consumption sensors are arranged for each set energy consumption device; Corresponding power consumption detection sensors and / or gas consumption detection sensors are arranged for each set energy consumption device, and historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are collected; the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are transmitted to an upper computer database through an industrial bus connected with the power consumption detection sensors and / or gas consumption detection sensors arranged for each set energy consumption device; and the collected historical power consumption data and / or historical gas consumption data, current power consumption data and / or current gas consumption data are stored in the upper computer database; And / or, The feature self-attention module comprises a channel attention module and a spatial attention module connected with the channel attention module; the channel attention module is configured to perform channel attention processing on the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature to obtain a channel attention feature; the spatial attention module is configured to perform spatial attention processing on the spatial attention feature to obtain a self-attention feature; before the channel attention module performs channel attention processing on the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature, the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature are spliced to obtain a to-be-processed spliced feature; and the channel attention module is configured to perform channel attention processing on the to-be-processed spliced feature to obtain a channel attention feature.

29. The method of predicting energy consumption based on plant operating conditions according to any one of claims 16, 17, 19, 21, 23, 24, 26-28, wherein, Including: A timer is used to time the use time of each set energy consumption device from opening to closing, and the use time of each set energy consumption device is obtained; The use time of each set energy consumption device is transmitted to an upper computer database through an industrial bus connected with the timer, and the use time of each set energy consumption device is stored in the upper computer database; And / or, The channel attention module of the feature self-attention module comprises: a first pooling unit and a second pooling unit different from the first pooling unit, a shared multi-layer perceptron connected with the first pooling unit and the second pooling unit respectively, an adder connected with the shared multi-layer perceptron, a third preset activation function connected with the adder, and a first multiplier; wherein the first pooling unit and the second pooling unit perform a pooling operation on the to-be-processed spliced features corresponding to the second energy consumption evaluation shallow layer feature, the first energy consumption evaluation deep layer feature and the second energy consumption evaluation deep layer feature respectively to obtain first pooled features and second pooled features; the shared multi-layer perceptron is configured to perform weight processing on the first pooled features and the second pooled features respectively to obtain first weight features and second weight features; the adder is configured to perform addition processing on the first weight features and the second weight features to obtain addition features; the third preset activation function is configured to perform nonlinear processing on the addition features to obtain channel attention weights; and the first multiplier is configured to perform multiplication operation processing on the channel attention weights and the to-be-processed spliced features to obtain channel attention features. 30.The method of claim 18, wherein, Comprise: using a timer, timing the use time of each set energy consumption device from opening to closing respectively to obtain the use time corresponding to each set energy consumption device; sending the use time corresponding to each set energy consumption device to an upper computer database through an industrial bus connected with the timer, and storing the use time corresponding to each set energy consumption device by using the upper computer database; and / or, The channel attention module of the feature self-attention module comprises: a first pooling unit and a second pooling unit different from the first pooling unit, a shared multi-layer perceptron connected with the first pooling unit and the second pooling unit respectively, an adder connected with the shared multi-layer perceptron, a third preset activation function connected with the adder, and a first multiplier; wherein the first pooling unit and the second pooling unit perform a pooling operation on the to-be-processed spliced features corresponding to the second energy consumption evaluation shallow layer feature, the first energy consumption evaluation deep layer feature and the second energy consumption evaluation deep layer feature respectively to obtain first pooled features and second pooled features; the shared multi-layer perceptron is configured to perform weight processing on the first pooled features and the second pooled features respectively to obtain first weight features and second weight features; the adder is configured to perform addition processing on the first weight features and the second weight features to obtain addition features; the third preset activation function is configured to perform nonlinear processing on the addition features to obtain channel attention weights; and the first multiplier is configured to perform multiplication operation processing on the channel attention weights and the to-be-processed spliced features to obtain channel attention features. 31.The method of claim 20, wherein, Comprise: using a timer, timing the use time of each set energy consumption device from opening to closing respectively to obtain the use time corresponding to each set energy consumption device; The use time corresponding to each set energy consumption device is sent to an upper computer database through an industrial bus connected with the timer, and the use time corresponding to each set energy consumption device is stored by using the upper computer database; And / or, The channel attention module of the feature self-attention module comprises a first pooling unit and a second pooling unit different from the first pooling unit, a shared multi-layer perceptron connected with the first pooling unit and the second pooling unit respectively, an adder connected with the shared multi-layer perceptron, a third set activation function and a first multiplier connected with the adder; wherein the first pooling unit and the second pooling unit perform pooling operation on the to-be-processed spliced features corresponding to the second energy consumption evaluation shallow feature, the first energy consumption evaluation deep feature and the second energy consumption evaluation deep feature respectively to obtain first pooled features and second pooled features; the shared multi-layer perceptron is used for performing weight processing on the first pooled features and the second pooled features respectively to obtain first weight features and second weight features; the adder is used for performing addition processing on the first weight features and the second weight features to obtain addition features; the third set activation function is used for performing nonlinear processing on the addition features to obtain channel attention weights; and the first multiplier is used for performing multiplication operation processing on the channel attention weights and the to-be-processed spliced features to obtain channel attention features.

32. The method of claim 22, wherein, Comprise: The use time corresponding to each set energy consumption device is sent to an upper computer database through an industrial bus connected with the timer, and the use time corresponding to each set energy consumption device is stored by using the upper computer database; The use time corresponding to each set energy consumption device is sent to an upper computer database through an industrial bus connected with the timer, and the use time corresponding to each set energy consumption device is stored by using the upper computer database; And / or, The channel attention module of the feature self-attention module comprises: a first pooling unit and a second pooling unit different from the first pooling unit, a shared multi-layer perceptron connected with the first pooling unit and the second pooling unit respectively, an adder connected with the shared multi-layer perceptron, a third preset activation function and a first multiplier connected with the adder; wherein the first pooling unit and the second pooling unit perform a pooling operation on the to-be-processed spliced features corresponding to the second energy consumption evaluation shallow layer feature, the first energy consumption evaluation deep layer feature and the second energy consumption evaluation deep layer feature respectively to obtain first pooled features and second pooled features; the shared multi-layer perceptron is configured to perform weight processing on the first pooled features and the second pooled features respectively to obtain first weight features and second weight features; the adder is configured to perform addition processing on the first weight features and the second weight features to obtain addition features; the third preset activation function is configured to perform nonlinear processing on the addition features to obtain channel attention weights; and the first multiplier is configured to perform multiplication operation processing on the channel attention weights and the to-be-processed spliced features to obtain channel attention features.

33. The method of claim 25, wherein, Comprising: using a timer, timing the use time of each preset energy consumption device from turning on to turning off to obtain the use time corresponding to each preset energy consumption device; sending the use time corresponding to each preset energy consumption device to an upper computer database through an industrial bus connected with the timer, and storing the use time corresponding to each preset energy consumption device by using the upper computer database; and / or, The channel attention module of the feature self-attention module comprises: a first pooling unit and a second pooling unit different from the first pooling unit, a shared multi-layer perceptron connected with the first pooling unit and the second pooling unit respectively, an adder connected with the shared multi-layer perceptron, a third preset activation function and a first multiplier connected with the adder; wherein the first pooling unit and the second pooling unit perform a pooling operation on the to-be-processed spliced features corresponding to the second energy consumption evaluation shallow layer feature, the first energy consumption evaluation deep layer feature and the second energy consumption evaluation deep layer feature respectively to obtain first pooled features and second pooled features; the shared multi-layer perceptron is configured to perform weight processing on the first pooled features and the second pooled features respectively to obtain first weight features and second weight features; the adder is configured to perform addition processing on the first weight features and the second weight features to obtain addition features; the third preset activation function is configured to perform nonlinear processing on the addition features to obtain channel attention weights; and the first multiplier is configured to perform multiplication operation processing on the channel attention weights and the to-be-processed spliced features to obtain channel attention features.

34. A plant operation state-based energy consumption simulation system, comprising: Comprising: The first training unit is configured to train a preset energy consumption simulation evaluation generator by using historical use time, historical power consumption data and / or historical gas consumption data corresponding to at least one acquisition time in a set time interval of each set energy consumption device in a plurality of set energy consumption devices under a factory operation state and historical energy consumption evaluation of the historical use time, the historical power consumption data and / or the historical gas consumption data, and to obtain the preset energy consumption simulation evaluation generator. The preset energy consumption simulation evaluation network includes a first energy consumption evaluation shallow feature extraction module, an encoder connected with the first energy consumption evaluation shallow feature extraction module, and a decoder connected with the first energy consumption evaluation shallow feature extraction module and the encoder. The first set convolutional neural unit corresponding to the shallow feature extraction module includes at least one one-dimensional convolutional layer or a plurality of cascaded one-dimensional convolutional layers, which are used to extract first energy consumption evaluation shallow features of the historical use time, the historical power consumption data and / or the historical gas consumption data, and random noise under the conditions of the historical use time, the historical power consumption data and / or the historical gas consumption data. The encoder is used to extract first energy consumption evaluation deep features and second energy consumption evaluation deep features corresponding to the first energy consumption evaluation shallow features. The decoder is used to generate an energy consumption simulation evaluation based on the first energy consumption evaluation shallow features, the first energy consumption evaluation deep features and the second energy consumption evaluation deep features corresponding to the first energy consumption evaluation shallow features. The decoder includes a sixth set convolutional neural unit, a feature self-attention module connected with the sixth set convolutional neural unit, and a linear layer connected with the feature self-attention module. The one-dimensional convolutional layer or the plurality of cascaded one-dimensional convolutional layers corresponding to the sixth set convolutional neural unit are used to perform convolution processing on the first energy consumption evaluation shallow features to obtain second energy consumption evaluation shallow features. The feature self-attention module is used to perform self-attention processing on the second energy consumption evaluation shallow features, the first energy consumption evaluation deep features and the second energy consumption evaluation deep features to obtain self-attention features. The linear layer is used to perform linear processing on the self-attention features to generate the energy consumption simulation evaluation. The energy consumption simulation evaluation is configured as one of high energy consumption level, medium energy consumption level, low energy consumption level and normal energy consumption level. The simulation evaluation unit is configured to generate a corresponding energy consumption simulation evaluation by using current use time, current power consumption data and / or current gas consumption data corresponding to each set energy consumption device in a plurality of set energy consumption devices under a factory operation state based on the preset energy consumption simulation evaluation generator.

35. A plant operation state-based energy consumption simulation system, comprising: The electronic device is configured with a processor and a memory for storing processor-executable instructions; wherein the processor is configured to call the instructions stored in the memory to execute the energy consumption simulation method based on the factory operation state in any one of claims 1-15. The processor is configured to call the instructions stored in the memory to execute the energy consumption simulation method based on the factory operation state in any one of claims 1-15.

36. A plant operation state-based energy consumption simulation system, comprising: The processor is configured to call the instructions stored in the memory to execute the energy consumption simulation method based on the factory operation state in any one of claims 1-15. ​ ​ 37. A plant operation state-based energy consumption simulation system, comprising: The computer program product comprises a computer program / instruction which, when executed by a processor, implements the energy consumption simulation method based on the factory operation state according to any one of claims 1-15. The first training unit is configured to train a preset energy consumption simulation evaluation generator by using historical use time, historical power consumption data and / or historical gas consumption data corresponding to at least one acquisition time in a set time interval of each set energy consumption device in a plurality of set energy consumption devices under a factory operation state and historical energy consumption evaluation corresponding to the historical use time, the historical power consumption data and / or the historical gas consumption data, to obtain a preset energy consumption simulation evaluation generator; wherein the preset energy consumption simulation evaluation network comprises a first energy consumption evaluation shallow feature extraction module, an encoder connected to the first energy consumption evaluation shallow feature extraction module, and a decoder connected to the first energy consumption evaluation shallow feature extraction module and the encoder, respectively; the first set convolutional neural unit corresponding to the shallow feature extraction module comprises at least one one-dimensional convolutional layer or a plurality of cascaded one-dimensional convolutional layers, for extracting first energy consumption evaluation shallow features of the historical use time, the historical power consumption data and / or the historical gas consumption data and random noise under the condition of the historical use time, the historical power consumption data and / or the historical gas consumption data; the encoder is configured to extract first energy consumption evaluation deep features and second energy consumption evaluation deep features corresponding to the first energy consumption evaluation shallow features; the decoder is configured to generate an energy consumption simulation evaluation based on the first energy consumption evaluation shallow features, the first energy consumption evaluation deep features and the second energy consumption evaluation deep features corresponding to the first energy consumption evaluation shallow features; the decoder comprises a sixth set convolutional neural unit, a feature self-attention module connected to the sixth set convolutional neural unit, and a linear layer connected to the feature self-attention module; wherein the sixth set convolutional neural unit comprises one one-dimensional convolutional layer or a plurality of cascaded one-dimensional convolutional layers, for performing convolution processing on the first energy consumption evaluation shallow features to obtain second energy consumption evaluation shallow features; the feature self-attention module is configured to perform self-attention processing on the second energy consumption evaluation shallow features, the first energy consumption evaluation deep features and the second energy consumption evaluation deep features to obtain self-attention features; the linear layer is configured to perform linear processing on the self-attention features to generate the energy consumption simulation evaluation; the energy consumption simulation evaluation is configured as one of high energy consumption level, medium energy consumption level, low energy consumption level and normal energy consumption level; 38. An energy consumption simulation device based on a plant operating state, characterized by, The simulation evaluation unit is configured to generate a corresponding energy consumption simulation evaluation by using current use time, current power consumption data and / or current gas consumption data corresponding to each set energy consumption device in a plurality of set energy consumption devices under a factory operation state based on the preset energy consumption simulation evaluation generator. The computer program product comprises a computer program / instruction which, when executed by a processor, implements the energy consumption simulation method based on the factory operation state according to any one of claims 1-15. ​ 39. An energy consumption simulation device based on a plant operating state, comprising: ​ An electronic device configured with a processor and a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to perform the method for simulating energy consumption based on factory running state according to any one of claims 1-15.

40. An energy consumption simulation device based on a plant operating state, comprising: Comprising: A processor; A memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to perform the method for simulating energy consumption based on factory running state according to any one of claims 1-15.

41. A plant operation state-based energy consumption simulation device, comprising: Comprising: A computer program product provided with a computer program / instructions, which, when executed by a processor, implements the method for simulating energy consumption based on factory running state according to any one of claims 1-15.

42. A plant operating state based energy consumption prediction system, comprising: Comprising: A second training unit configured to train a preset energy consumption simulation evaluation generator by using historical usage time, historical power consumption data and / or historical gas consumption data corresponding to a first time in a set time interval and historical energy consumption evaluation corresponding to a second time after the first time of each set energy consumption device in a plurality of set energy consumption devices under factory running state, to obtain a preset energy consumption simulation evaluation generator; wherein the preset energy consumption simulation evaluation network comprises: a first energy consumption evaluation shallow feature extraction module, an encoder connected to the first energy consumption evaluation shallow feature extraction module, and a decoder connected to the first energy consumption evaluation shallow feature extraction module and the encoder, respectively; the first set of convolutional neural units corresponding to the shallow feature extraction module at least includes one one-dimensional convolutional layer or a plurality of cascaded one-dimensional convolutional layers, which are used to extract first energy consumption evaluation shallow features of the historical usage time, the historical power consumption data and / or the historical gas consumption data, and random noise under the conditions of the historical usage time, the historical power consumption data and / or the historical gas consumption data; the encoder is used to extract first energy consumption evaluation deep features and second energy consumption evaluation deep features corresponding to the first energy consumption evaluation shallow features; the decoder is used to generate an energy consumption simulation evaluation based on the first energy consumption evaluation shallow features, the first energy consumption evaluation deep features and the second energy consumption evaluation deep features corresponding thereto; the decoder comprises a sixth set of convolutional neural units, a feature self-attention module connected to the sixth set of convolutional neural units, and a linear layer connected to the feature self-attention module; wherein the one one-dimensional convolutional layer or the plurality of cascaded one-dimensional convolutional layers of the sixth set of convolutional neural units are used to perform convolution processing on the first energy consumption evaluation shallow features to obtain second energy consumption evaluation shallow features; the feature self-attention module is used to perform self-attention processing on the second energy consumption evaluation shallow features, the first energy consumption evaluation deep features and the second energy consumption evaluation deep features to obtain self-attention features; the linear layer is used to perform linear processing on the self-attention features to generate the energy consumption simulation evaluation; the energy consumption simulation evaluation is configured as one of high energy consumption level, medium energy consumption level, low energy consumption level and normal energy consumption level; The energy consumption evaluation prediction unit is configured to, based on the preset energy consumption simulation evaluation generator, utilize current use time, current power consumption data and / or current gas consumption data corresponding to each of the plurality of set energy consumption devices at the third time to predict energy consumption evaluation corresponding to a fourth time after the third time.

43. A plant operating state-based energy consumption prediction system, comprising: The computer program product comprises: An electronic device, configured with a processor and a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the energy consumption prediction method based on the factory operation state according to any one of claims 16-33.

44. A plant operating state based energy consumption prediction system, comprising: The computer program product comprises: A processor; A memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the energy consumption prediction method based on the factory operation state according to any one of claims 16-33.

45. A plant operating state-based energy consumption prediction system, comprising: The computer program product comprises: A computer program product, provided with a computer program / instruction, which, when executed by a processor, implements the energy consumption prediction method based on the factory operation state according to any one of claims 16-33.

46. A plant operation state-based energy consumption prediction device characterized by comprising: The computer program product comprises: The second training unit is configured to train a preset energy consumption simulation evaluation generator based on historical use time, historical power consumption data and / or historical gas consumption data of each of the plurality of set energy consumption devices at a first time in a set time interval and historical energy consumption evaluation at a second time after the first time under the factory operation state, to obtain the preset energy consumption simulation evaluation generator; wherein the preset energy consumption simulation evaluation network comprises a first energy consumption evaluation shallow feature extraction module, an encoder connected with the first energy consumption evaluation shallow feature extraction module, and a decoder connected with the first energy consumption evaluation shallow feature extraction module and the encoder, respectively; the first set of convolutional neural units corresponding to the shallow feature extraction module at least includes one one-dimensional convolutional layer or a plurality of cascaded one-dimensional convolutional layers, which are used to extract first energy consumption evaluation shallow features of the historical use time, the historical power consumption data and / or the historical gas consumption data, and random noise under the conditions of the historical use time, the historical power consumption data and / or the historical gas consumption data; the encoder is used to extract first energy consumption evaluation deep features and second energy consumption evaluation deep features corresponding to the first energy consumption evaluation shallow features; the decoder is used to generate an energy consumption simulation evaluation based on the first energy consumption evaluation shallow features, the first energy consumption evaluation deep features and the second energy consumption evaluation deep features corresponding thereto; the decoder comprises a sixth set of convolutional neural units, a feature self-attention module connected with the sixth set of convolutional neural units, and a linear layer connected with the feature self-attention module; wherein the one one-dimensional convolutional layer or the plurality of cascaded one-dimensional convolutional layers of the sixth set of convolutional neural units are used to perform convolution processing on the first energy consumption evaluation shallow features to obtain second energy consumption evaluation shallow features; the feature self-attention module is used to perform self-attention processing on the second energy consumption evaluation shallow features, the first energy consumption evaluation deep features and the second energy consumption evaluation deep features to obtain self-attention features; the linear layer is used to perform linear processing on the self-attention features to generate the energy consumption simulation evaluation; the energy consumption simulation evaluation is configured as one of high energy consumption level, medium energy consumption level, low energy consumption level and normal energy consumption level; An energy consumption evaluation prediction unit is configured to predict an energy consumption evaluation corresponding to a fourth time after a third time based on the preset energy consumption simulation evaluation generator and current use time, current power consumption data and / or current gas consumption data of each of the plurality of set energy consumption devices at the third time under the factory operation state.

47. A plant operation state-based energy consumption prediction device characterized by comprising: An electronic device is provided, which is configured with a processor and a memory for storing instructions executable by the processor; wherein the processor is configured to invoke the instructions stored in the memory to perform the energy consumption prediction method based on the factory operation state according to any one of claims 16-33. A processor is provided.

48. A plant operation state-based energy consumption prediction device, comprising: A processor is provided. ​ A memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored by the memory to perform the method of any one of claims 16-33 for predicting energy consumption based on a plant operating state.

49. A plant operation state-based energy consumption prediction device, comprising: comprising: A computer program product provided with a computer program / instructions which, when executed by a processor, implement the method of any one of claims 16-33 for predicting energy consumption based on a plant operating state.

50. A computer-readable storage medium having stored thereon computer program instructions, wherein, The computer program instructions, when executed by a processor, implement the method of any one of claims 1-15 for simulating energy consumption based on a plant operating state.

51. A computer-readable storage medium having stored thereon computer program instructions, wherein, The computer program instructions, when executed by a processor, implement the method of any one of claims 16-33 for predicting energy consumption based on a plant operating state.

Images