Industrial equipment carbon emission measuring and calculating method and device based on running state dynamic monitoring

Abstract

The invention discloses an industrial equipment carbon emission measuring and calculating method and device based on running state dynamic monitoring, and the method comprises the steps: obtaining user historical electricity consumption data of detected equipment, and scanning the user historical electricity consumption data through a time sliding window, inputting the data scanned by the time sliding window and the corresponding industrial equipment operation state code into a UNet network for model training, and obtaining an equipment operation state monitoring model after the training is completed; adaptively adjusting the size of the window according to the data correlation between the adjacent windows; acquiring user real-time electricity consumption data of the to-be-tested equipment, and inputting the user real-time electricity consumption data into the equipment running state monitoring model to obtain a real-time running state code of the to-be-tested equipment; and calculating the carbon emission of the tested equipment according to the real-time running state code of the tested equipment. The method has the advantages of being simple in implementation method, high in efficiency and precision, high in real-time performance and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial equipment monitoring, and in particular to a method and device for calculating carbon emissions of industrial equipment based on dynamic monitoring of operating status. Background Art

[0002] Carbon emissions from industrial equipment account for a large proportion of total carbon emissions, so accurate monitoring and effective control of carbon emissions from industrial equipment are very necessary. In the prior art, carbon emissions monitoring methods for industrial equipment usually adopt a carbon emission factor method based on fossil fuels. This type of method estimates carbon emissions based on the basic carbon accounting equation of greenhouse gas (GHG) emissions = activity data (AD) × emission factor (EF), where AD is the amount of production or consumption activities that lead to greenhouse gas emissions, such as the consumption of each fossil fuel, the consumption of limestone raw materials, electricity, steam, etc. This type of method is simple and easy to implement, but because it only estimates carbon emissions based on fossil fuel consumption from the perspective of time and users, it only roughly estimates carbon emission activities, and industrial equipment is usually electrically driven, and the carbon emissions of equipment under different driving states are different, and a carbon emission activity is usually completed by the collaborative work of multiple industrial equipment, and the carbon emissions of different industrial equipment under different operating states are different. Therefore, the actual accuracy of the traditional carbon emission factor method is not high and the scope of application is limited, and it is difficult to support refined carbon management. Summary of the invention

[0003] The technical problem to be solved by the present invention is: in response to the technical problems existing in the prior art, the present invention provides a method and device for calculating carbon emissions of industrial equipment based on dynamic monitoring of operating status, which has a simple implementation method, high efficiency and accuracy, and strong real-time performance.

[0004] In order to solve the above technical problems, the technical solution proposed by the present invention is:

[0005] A method for calculating carbon emissions of industrial equipment based on dynamic monitoring of operating status, comprising the following steps:

[0006] Obtain the user's historical electricity consumption data of the device under test, use a time sliding window to scan the user's historical electricity consumption data, input the data scanned by the time sliding window and the corresponding industrial equipment operation status code into the UNet network for model training, and obtain the equipment operation status monitoring model after the training is completed. In the process of using the time sliding window to scan the user's historical electricity consumption data, the size of the window is adaptively adjusted according to the data correlation between adjacent windows;

[0007] Acquire the real-time power consumption data of the user of the device under test at a specified time interval, input the acquired real-time power consumption data of the user into the device operation status monitoring model, and obtain the real-time operation status of the device under test;

[0008] The carbon emissions of the equipment under test are calculated based on the real-time operating status of the equipment under test.

[0009] Furthermore, the adaptively adjusting the size of the window according to the data correlation between adjacent windows includes:

[0010] Calculate the correlation coefficient between the power consumption data of the current window and the previous window:

[0011] The size of the current window is adjusted according to the size of the correlation strength coefficient. If the correlation strength coefficient is greater than the preset strong correlation threshold (the threshold is 0.70), that is, the correlation strength coefficient is greater than 0.70 for strong correlation, then the window width is increased; if the correlation strength coefficient is less than the preset weak correlation threshold (the threshold is 0.5), that is, the correlation coefficient is less than 0.5 for weak correlation, it is judged that there is a change in the operating status of the equipment, and the window width is reduced.

[0012] Furthermore, the correlation strength coefficient includes the Wilson correlation coefficient. The expression for calculating the Wilson correlation coefficient between the power consumption data of the current window and the previous window is:

[0013]

[0014] Among them, S i is the electricity consumption data sequence of the user in the i-th window, S i,j is the power consumption data of the jth segment in the i-th window, n is the number of segments in each window; μ i is the mean value of the electricity consumption data sequence in the i-th window; σ i is the variance of the electricity consumption data sequence in the ith window, ρ(S i , S i-1 ) is the Wilson correlation coefficient between the electricity consumption data sequence of the i-th window and the electricity consumption data sequence of the i-1-th window.

[0015] Furthermore, in the process of using the time sliding window to scan the user's historical electricity consumption data, the initial time window is set to W=A, where A is set according to the minimum unit time of device operation. If the correlation value is greater than the preset strong correlation threshold, the current window width is increased by A; if the correlation value is less than the preset weak correlation threshold, the current window width is reduced by A.

[0016] Furthermore, the model training process also includes using an adaptive sparse filtering optimization algorithm to optimize the electricity consumption characteristics extracted from the user's electricity consumption data. The adaptive sparse filtering optimization algorithm uses multiple deep sparse filters to filter the input data to calculate the optimized electricity consumption characteristics. Each deep sparse filter is composed of two layers, where the first layer normalizes the input feature data, and the output of the first layer is used as the input of the second layer. The second layer is normalized again to obtain sparse characteristics, and each deep sparse filter calculates the final output through a layer-by-layer greedy algorithm.

[0017] Furthermore, the calculation expression of the feature optimization using the adaptive sparse filtering optimization algorithm is:

[0018]

[0019]

[0020] Among them, A is the optimized user power consumption feature vector, W is the weight matrix, RS is the L2 normalized matrix of each row of A, CS is the result of L2 normalization of each column of RS, and P represents the sparsity adjustment parameter; by optimizing the sample characteristics and minimizing CS as the goal, the optimized user power consumption characteristics are solved.

[0021] Furthermore, the equipment operation status monitoring model extracts power consumption features from the input user power consumption data, imports them into a convolutional neural network after sparse filtering, and performs feature extraction through a fusion layer, extracts optimized power consumption features for measuring the equipment load status, and obtains equipment status recognition results after the optimized power consumption features pass through a Sigmoid activation function.

[0022] Furthermore, the step of calculating the carbon emissions of the device under test within the specified time period according to the real-time operating state code of the device under test and the unit carbon emissions factor under the corresponding operating state within the specified time period includes:

[0023] Obtain all real-time running status codes of the device under test identified within a specified time period, and form a running status vector from all the identified real-time running status codes, wherein the running status code is a one-hot code;

[0024] Obtaining a unit carbon emission factor corresponding to each real-time operating state code to form a carbon emission factor vector, wherein the unit carbon emission factor is a unit carbon emission under different operating states;

[0025] The operating state vector is multiplied by the carbon emission factor vector and the carbon emission factor vector to obtain the carbon emission of the device under test within a specified time period.

[0026] A device for calculating carbon emissions of industrial equipment based on dynamic monitoring of operating status includes a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to execute the computer program to perform the above method.

[0027] A computer-readable storage medium storing a computer program, wherein the computer program implements the above method when executed.

[0028] Compared with the prior art, the advantages of the present invention are: the present invention breaks down carbon emission activities into the operating status of user equipment within a unit time, uses the user's historical electricity consumption data to train the equipment operating status monitoring model based on the UNet network, and utilizes the characteristics of the UNet network to fully mine the user's electricity consumption characteristics from the user's historical electricity consumption data. The real-time user electricity consumption data can be quickly identified using the monitoring model, which can greatly shorten the data and time required for equipment carbon monitoring. At the same time, by adjusting the structure of the input data and adjusting the size of the monitoring window according to the correlation strength of the electricity consumption data, different time periods can be adaptively cut using the strength of the electricity consumption correlation, rather than in a fixed cutting manner, thereby being able to more effectively identify the equipment operation length and achieve more accurate real-time monitoring of carbon emissions. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 Schematic diagram of the implementation process of the method for calculating carbon emissions of industrial equipment based on dynamic monitoring of operating status in this embodiment.

[0030] Figure 2 The figure is a flow chart of the window adaptation rule adopted in this embodiment.

[0031] Figure 3 This is a schematic diagram of the structural principle of the convolutional neural network based on sparse filtering optimization adopted in this embodiment.

[0032] Figure 4 The present invention is a flowchart for realizing the calculation of carbon emissions of industrial equipment in a specific application embodiment of the present invention. DETAILED DESCRIPTION

[0033] The present invention is further described below in conjunction with the accompanying drawings and specific preferred embodiments, but the protection scope of the present invention is not limited thereby.

[0034] A carbon emission activity usually involves multiple industrial equipment working together. Each industrial equipment will generate carbon emissions during operation, and the carbon emissions of industrial equipment depend on the operating status of the industrial equipment, that is, the carbon emissions of the equipment under different operating status are different. For example, when the equipment is driven by a high proportion of fossil fuels, that is, when the equipment is operated at a high efficiency, the carbon emissions generated by the equipment are large, and correspondingly, when the equipment is driven by a low proportion of fossil fuels, that is, when the equipment is operated at a low efficiency, the carbon emissions generated by the equipment are also low. Taking the carbon emissions of the steelmaking process as an example, the steelmaking process needs to be completed through the coordinated work of blast furnaces, converters, refining furnaces, electric arc furnaces and auxiliary equipment. For each boiler, if the furnace is in a high power state such as a high temperature or a good conversion effect, the carbon emissions generated are also high, and if the operating efficiency of the furnace is low, the corresponding carbon emissions are also less. In the process of estimating carbon emissions, the traditional carbon emission factor method directly makes a rough estimate of a carbon emission activity as a whole, so the actual accuracy is not high and the scope of application is limited, making it difficult to support refined carbon management.

[0035] The present invention breaks down carbon emission activities into the operating status of user equipment within a unit time, monitors the operating status of each industrial equipment in real time, and determines the corresponding unit carbon emission factor based on the operating status of each industrial equipment, thereby accurately calculating the real-time carbon emissions of each industrial equipment, effectively improving the accuracy and scope of carbon emission monitoring, and realizing refined management of carbon emission activities. At the same time, combined with the characteristics that the correlation strength of power consumption is strongly correlated with the operating status of the equipment, the user's power consumption data is obtained by using a time sliding window, and the size of the monitoring window is adaptively optimized by using the correlation strength of power consumption between adjacent windows, which can improve the accuracy of user equipment status identification, so that data related to the estimated equipment carbon emissions can be more accurately determined based on the equipment operating status, so as to better perform carbon emission monitoring.

[0036] like Figure 1 As shown, the steps of the method for calculating carbon emissions of industrial equipment based on dynamic monitoring of operating status in this embodiment include:

[0037] Step S101. Obtain the user's historical electricity consumption data of the device under test, use a time sliding window to scan the user's historical electricity consumption data, input the data scanned by the time sliding window and the corresponding operating status code of the industrial equipment into the UNet network for model training, and obtain the equipment operating status monitoring model after the training is completed. In the process of using the time sliding window to scan the user's historical electricity consumption data, the size of the window is adaptively adjusted according to the data correlation between adjacent windows;

[0038] Step S102. Acquire the real-time power consumption data of the user of the device under test at a specified time interval, input the acquired real-time power consumption data of the user into the device operation status monitoring model, and obtain the real-time operation status code of the device under test;

[0039] Step S103: Calculate the carbon emissions of the device under test within the specified time period according to the real-time operating status code of the device under test and the unit carbon emissions factor under the corresponding operating status.

[0040] This embodiment uses the user's historical electricity consumption data to train the equipment operation status monitoring model based on the UNet network. The UNet network can effectively capture the spatial characteristics of electricity consumption data, fully mine the user's electricity consumption characteristics from the user's historical electricity consumption data, and dynamically adjust the network's focus on high identification accuracy from the perspective of time series, thereby achieving high-precision user equipment status identification, and real-time user electricity consumption data can be quickly identified using the monitoring model, which can greatly shorten the data and time required for equipment carbon monitoring. At the same time, by adjusting the structure of the input data, the size of the monitoring window is adaptively optimized according to the strength of the correlation of the electricity consumption data, so that different time periods can be adaptively cut using the strength of the electricity consumption correlation, rather than in a fixed cutting manner, so that the equipment operation length can be more effectively identified, and more accurate real-time monitoring of carbon emissions can be achieved in combination with the monitoring of the equipment operation status.

[0041] In this embodiment, after obtaining the user's historical electricity consumption data of the device under test in step S101, the user's historical electricity consumption data is also preprocessed. The preprocessing may include using Lagrange interpolation method to interpolate missing values ​​and other processing to form a user electricity consumption time series.

[0042] In this embodiment, the adaptively adjusting the size of the window according to the data correlation between adjacent windows in step S101 includes:

[0043] Calculate the correlation coefficient between the power consumption data of the current window and the previous window:

[0044] The size of the current window is adjusted according to the size of the correlation strength coefficient. If the correlation strength coefficient is greater than a preset strong correlation threshold (for example, the threshold can be taken as 0.70, that is, the correlation strength coefficient is greater than 0.70 for strong correlation), the window width is increased; if the correlation strength coefficient is less than a preset weak correlation threshold (for example, the threshold can be taken as 0.5, that is, the correlation strength coefficient is less than 0.5 for weak correlation), it is determined that there is a change in the operating status of the equipment, and the window width is reduced.

[0045] In step S102 of this embodiment, by using a time sliding window to adaptively cut data of different time periods, after setting the window size, the window size is adjusted according to the difference in power consumption correlation within the window size, so that the minimum unit of the device operation status is the window size. For example, setting 30 minutes as the window size can make the minimum unit of the device operation status half an hour.

[0046] Optionally, the correlation strength coefficient may be a Wilson correlation coefficient, etc. For example, the Wilson correlation coefficient between the power consumption data of the current window and the previous window may be calculated according to the following formula:

[0047]

[0048]

[0049] Among them, S i is the electricity consumption data sequence of the user in the i-th window, S i,j is the power consumption data of the jth segment in the i-th window, n is the number of segments in each window; μ i is the mean value of the electricity consumption data sequence in the i-th window; σ i is the variance of the electricity consumption data sequence in the ith window, ρ(S i , S i-1 ) is the Wilson correlation coefficient between the electricity consumption data sequence of the i-th window and the electricity consumption data sequence of the i-1-th window.

[0050] For example, when the window size W is set to 30 minutes, the power consumption data within 30 minutes are arranged, and then the mean, variance and Wilson correlation coefficient between the power consumption data of adjacent windows are calculated in the above manner, where S i represents the user's 30-minute electricity consumption data sequence, S i,j represents the power consumption data of the jth segment (e.g., five minutes) of the i-th 30-minute period of the user, ρ(S i , S i-1 ) represents the Wilson correlation coefficient between the electricity consumption data of each 30 minutes and the electricity consumption data sequence of the previous 30 minutes, and then the window size is adjusted according to the size of the Wilson correlation coefficient.

[0051] This embodiment optimizes the window size according to the correlation strength of the power consumption data between adjacent windows, such as Figure 2As shown, if the correlation coefficient between the two windows is less than the preset weak correlation threshold (specifically 0.5), that is, the correlation is weak, indicating that the equipment operating state has changed. Reducing the sliding window W can improve the monitoring accuracy; if the correlation coefficient between the two windows is greater than the preset strong correlation threshold, that is, the correlation is extremely strong, it is considered that equipment change behavior is impossible. At this time, expanding the sliding window W can increase the monitoring speed. If the correlation is between the preset strong correlation threshold and the preset weak correlation threshold, the sliding window W will not change. Through the above method, the dynamic monitoring of the equipment operating state can be fully utilized to improve the monitoring accuracy, and the monitoring efficiency can also be effectively improved to achieve carbon monitoring that takes into account both monitoring accuracy and efficiency. The size of the above preset strong correlation threshold and the preset weak correlation threshold can be configured according to actual needs.

[0052] For example, when adjusting the window size according to the Wilson correlation coefficient, the variable DIF i It represents the correlation strength coefficient of the i-th window compared with the previous window, recorded as:

[0053]

[0054] Take the absolute value of the correlation coefficient to determine the degree of correlation. For example, 0-0.09 can be taken as no correlation, 0.1-0.5 as weak correlation, 0.3-0.7 as medium correlation, and 0.7-1.0 as strong correlation. Then set the step threshold: weak correlation threshold P1 = 0.5, strong correlation threshold P2 = 0.7. If DIF i If DIF is less than P1, it indicates that the correlation is greatly weakened, and the correlation between the two windows is considered to be weak, then the sliding window W should be reduced to improve the monitoring accuracy. i If it is greater than P2, it indicates that the correlation is extremely strong. It is considered that the correlation between the two windows is extremely strong, and it is considered that device change behavior is impossible. The sliding window W is expanded to increase the monitoring speed. If the correlation is between P1 and P2, the sliding window W is not changed.

[0055] In this embodiment, when using a sliding window to scan the user's historical electricity consumption data, the initial time window is set to W=A, where A is set according to the minimum unit time of the device operation. If the correlation value is greater than the preset strong correlation threshold (specifically, the threshold can be taken as 0.70), the current window width is increased by A. If the correlation value is less than the preset weak correlation threshold (specifically, the threshold can be taken as 0.5), the current window width is reduced by A.

[0056] For example, based on the user's electricity usage habits, the window can be initially set to W=30, and the adjustment details are: if the correlation is very strong, the window is increased by 30 to improve the monitoring speed; if the correlation is weak, the window is reduced by 30; if the correlation is average, the window remains unchanged.

[0057] In this embodiment, the process of model training also includes using an adaptive sparse filtering optimization algorithm to optimize the power consumption features extracted from the user's power consumption data. In the adaptive sparse filtering optimization algorithm, multiple deep sparse filters are used to filter the input data to calculate the optimized power consumption features. Each deep sparse filter is composed of two layers, wherein the first layer normalizes the input feature data, and the output of the first layer is used as the input of the second layer. The second layer is normalized again to obtain sparse characteristics, and each deep sparse filter calculates the final output through a layer-by-layer greedy algorithm. Through the above method, the use of an adaptive sparse filter can accelerate training, achieve rapid convergence, and have better generalization capabilities. Combining the sparse filtering algorithm (SF) and the adaptive time window to optimize the features, it can also more fully explore the power consumption characteristics of the equipment, optimize the feature extraction method required for carbon monitoring, and further improve the accuracy of carbon monitoring.

[0058] Optionally, the following expression can be used to implement the optimization feature of the adaptive sparse filtering optimization algorithm:

[0059]

[0060] Among them, A is the optimized user power consumption feature vector, W is the weight matrix, RS is the L2 normalized matrix of each row of A, CS is the result of L2 normalization of each column of RS, and P represents the sparsity adjustment parameter; by optimizing the sample characteristics and minimizing CS as the goal, the optimized user power consumption characteristics are solved.

[0061] like Figure 3 As shown, in this embodiment, the equipment operation status monitoring model extracts power consumption features from the input user power consumption data, imports them into the convolutional neural network after sparse filtering, and extracts features through the fusion layer, extracts optimized power consumption features to measure the equipment load status, and obtains the equipment status recognition result after the optimized power consumption features are activated by the Sigmoid function. For example, the model can output 0 / 1, 0 for not working, 1 for normal operation, and the output layer Sigmoid activation function can be in the following form:

[0062]

[0063] This embodiment combines window adaptive optimization and device identification based on the strength of electricity consumption correlation to set a corresponding window size (for example, 30 minutes, one electricity consumption data per minute) as the initial time window, so that the minimum granularity of the collected data of the entire device can reach the minute level (for example, 1 minute), so that the minute granularity and window detection can be combined with the carbon emission intensity monitored at a larger scale (for example, unit hour), thereby realizing more accurate real-time monitoring of carbon emissions.

[0064] In step S103 of this embodiment, the carbon emissions of the device under test can be calculated according to the real-time operating status of the device under test in combination with the carbon factor method, and the steps include:

[0065] Step S131. Acquire all real-time running status codes of the device under test identified within a specified time period, and form a running status vector from all the identified real-time running status codes;

[0066] Step S132. Obtain the unit carbon emission factor corresponding to each real-time operating state code to form a carbon emission factor vector, where the unit carbon emission factor is the unit carbon emission under different operating states;

[0067] Step S133: Multiply the operating state vector and the carbon emission factor vector to obtain the carbon emissions of the device under test within a specified time period.

[0068] Specifically, different unit carbon emission factors EF can be set for different operating states of the device under test in advance according to the carbon emissions set under different operating states, that is, the unit carbon emissions under different operating states. After identifying the real-time operating state of the device under test, the carbon emissions of the device within a certain period of time are calculated in combination with the unit carbon emission factors EF corresponding to each operating state. Finally, the carbon emissions of multiple collaborative devices can be combined to obtain the accurate carbon emissions of a carbon emission activity. For example, the operating state of the device under test can be divided into high power state, normal power state and low power state, and a unit carbon emission factor EF (EF1, EF2 and EF3) is configured for each of the three operating states. If it is identified that the current device is in a high power state, the corresponding unit carbon emission factor is EF1. If the current device is in a normal power state, the corresponding unit carbon emission factor is EF2. The real-time carbon emissions of each device involved in a carbon emission activity are superimposed to obtain the accurate carbon emissions of the carbon emission activity.

[0069] Preferably, a unique hot code can be used to encode different operating states of the device. For example, a three-digit code is used to represent the operating state of the device, 001 represents the first operating state of the first device a (for example, high power), and 010 represents the second operating state of the first device a (for example, low power). Different codes of each operating state are pre-set with different unit emission factors. For example, the first device a corresponds to unit emission factors a1, a2, and a3 when they are 001 (first state), 010 (second state), and 100 (third state), respectively. The vector A formed by the unit emission factors a1, a2, and a3 is the carbon emission factor vector. All operating states of the first device a within the specified time period to be monitored are encoded to form an operating state vector. For example, 001010 can be used to indicate that the first device a has two operating states, 001 and 002, within the specified time period. Multiplying the vector A by the operating state vector can obtain the carbon emissions of the first device a within the specified time period. Correspondingly, a carbon activity may include multiple industrial equipment. In order to obtain the carbon emissions of the entire carbon activity, the operating status of other equipment within a specified time period is further obtained. For example, the operating status of the second device b within the specified time period is coded as 001001, that is, it is in a high-power operating state. The carbon emission factor vector B of the second device b in different operating states is obtained. The vector B contains the unit carbon emission factor of the second device b in each operating state. Correspondingly, the operating status vector of the second device within the specified time period is multiplied by the vector B to obtain the carbon emissions of the second device b within the specified time period. Finally, the carbon emissions of the first device a, the second device b and all other related equipment are added to obtain the accurate carbon emissions of a complete carbon activity.

[0070] The present invention is further described below by taking the carbon monitoring of a certain industrial equipment by the above method of the present invention in a specific application embodiment as an example. Figure 4 As shown, the detailed steps are:

[0071] Step S201, data collection: collecting historical electricity consumption data of the industrial equipment in the past year;

[0072] Step S202, data preprocessing: Lagrange interpolation method is used to interpolate missing values ​​of user historical data;

[0073]

[0074] In the formula, S i,j is the electricity consumption of the i-th user on the j-th day.

[0075] During the data collection and data preprocessing process, for users whose data missing ratio exceeds 5% of the historical data, the user's data is considered to be incorrect and the user is directly eliminated; for users whose data missing ratio is within 5% of the historical data, the Lagrange interpolation method is used to interpolate and supplement the default value.

[0076] Step 203, set 30 minutes as the window size, optimize the window based on the correlation difference between half-hour power consumption, so that the minimum unit time of equipment operation is half an hour. That is, set the initial time window size to 30, build the monitoring window adaptive optimization rule based on the correlation strength of power consumption data between adjacent windows, if the correlation between the power consumption of the window and the power consumption of the previous window is greater than 0.8, increase the window width W=W+30, if the correlation is less than 0.20, then the window bandwidth is W=W-30.

[0077] Step 204, model training: extract the sample features of user electricity consumption, and use sparse filtering to optimize the sample features, and use them as the input of the convolutional neural network. Use the operating status mark of the industrial equipment as the model output, and then use the convolutional neural network algorithm to train the model to obtain a trained model.

[0078] Specifically, an adaptive sparse filtering algorithm is used to optimize features. A deep sparse filter consisting of two layers is used. Layer 1 is trained with data blocks of size (n-29)*30. First, normalization is performed to obtain normalized feature data as the input of subsequent layers; the second layer is normalized again to obtain sparse characteristics, and the optimized features are calculated by two layers of deep sparse filtering. The final output is calculated by a layer-by-layer greedy algorithm.

[0079] Step 205, real-time monitoring of carbon emissions of user industrial equipment: using the trained model obtained in step 204, the equipment is monitored in real time based on real-time acquisition of user electricity consumption data in groups of 30 minutes, and the carbon emission factor method is used to multiply the code of the operating equipment status with the corresponding unit emission factor vector to obtain the carbon emissions of the industrial equipment, thereby realizing carbon monitoring of industrial equipment. By combining the carbon emissions of various related industrial equipment, the accurate carbon emissions of a complete carbon emission activity can be obtained.

[0080] This embodiment also provides an industrial equipment carbon emission measurement device based on dynamic monitoring of operating status, including a processor and a memory, the memory is used to store computer programs, and the processor is used to execute the computer program to perform the above method.

[0081] It is understandable that the above method of this embodiment can be executed by a single device, such as a computer or server, etc., and can also be applied to a distributed scenario and completed by multiple devices in cooperation with each other. In the case of a distributed scenario, one of the multiple devices can only execute one or more steps in the above method of this embodiment, and multiple devices interact to complete the above method. The processor can be implemented in the form of a general-purpose CPU, a microprocessor, an application-specific integrated circuit, or one or more integrated circuits, etc., for executing related programs to implement the above method of this embodiment. The memory can be implemented in the form of a read-only memory ROM, a random access memory RAM, a static storage device, and a dynamic storage device. The memory can store an operating system and other applications. When the above method of this embodiment is implemented by software or firmware, the relevant program code is stored in the memory and called and executed by the processor.

[0082] This embodiment also provides a computer-readable storage medium storing a computer program, and the computer program implements the above method when executed.

[0083] Those skilled in the art will appreciate that the above-mentioned embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may take the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, may be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the functions in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing device to work in a specific way, so that the instructions stored in the computer-readable memory produce a product including an instruction device, which implements the functions specified in the process. Figure 1 A process or multiple processes and / or boxes Figure 1These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0084] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A method for calculating carbon emissions of industrial equipment based on dynamic monitoring of operating status, characterized in that the steps include: Obtain the user's historical electricity consumption data of the device under test, use a time sliding window to scan the user's historical electricity consumption data, input the data scanned by the time sliding window and the corresponding industrial equipment operation status code into the UNet network for model training, and obtain the equipment operation status monitoring model after the training is completed. In the process of using the time sliding window to scan the user's historical electricity consumption data, the size of the window is adaptively adjusted according to the data correlation between adjacent windows; Acquire the real-time power consumption data of the user of the device under test at a specified time interval, input the acquired real-time power consumption data of the user into the device operation status monitoring model, and obtain the real-time operation status code of the device under test; The carbon emissions of the device under test within the specified time period are calculated based on the real-time operating status code of the device under test and the unit carbon emissions factor under the corresponding operating status.

2. The method for calculating carbon emissions of industrial equipment based on dynamic monitoring of operating status according to claim 1 is characterized in that: The adaptively adjusting the size of the window according to the data correlation between adjacent windows includes: Calculate the correlation coefficient between the power consumption data of the current window and the previous window: The size of the current window is adjusted according to the size of the correlation strength coefficient. If the correlation strength coefficient is greater than the preset strong correlation threshold, the window width is increased. If the correlation strength coefficient is less than the preset weak correlation threshold, it is determined that there is a change in the operating status of the equipment, and the window width is reduced.

3. The method for calculating carbon emissions of industrial equipment based on dynamic monitoring of operating status according to claim 2 is characterized in that: The correlation strength coefficient includes the Wilson correlation coefficient. The expression for calculating the Wilson correlation coefficient between the power consumption data of the current window and the previous window is: Among them, S i is the electricity consumption data sequence of the user in the i-th window, S i,j is the power consumption data of the jth segment in the i-th window, n is the number of segments in each window; μ i is the mean value of the electricity consumption data sequence in the i-th window; σ i is the variance of the electricity consumption data sequence in the ith window, ρ(S i , S i-1 ) is the Wilson correlation coefficient between the electricity consumption data sequence of the i-th window and the electricity consumption data sequence of the i-1-th window.

4. The method for calculating carbon emissions of industrial equipment based on dynamic monitoring of operating status according to claim 2 is characterized in that: In the process of scanning the user's historical electricity consumption data using the time sliding window, the initial time window is set to W=A, where A is the value obtained according to the minimum unit time setting of the equipment operation. If the correlation value is greater than the preset strong correlation threshold, the current window width is increased by A; if the correlation value is less than the preset weak correlation threshold, the current window width is reduced by A.

5. The method for calculating carbon emissions of industrial equipment based on dynamic monitoring of operating status according to any one of claims 1 to 4, characterized in that: The model training process also includes using an adaptive sparse filtering optimization algorithm to optimize the power consumption features extracted from the user's power consumption data. The adaptive sparse filtering optimization algorithm uses multiple deep sparse filters to filter the input data to obtain the optimized power consumption features. Each deep sparse filter is composed of two layers, wherein the first layer normalizes the input feature data, and the output of the first layer is used as the input of the second layer. The second layer is normalized again to obtain sparse characteristics, and each deep sparse filter calculates the final output through a layer-by-layer greedy algorithm.

6. The method for calculating carbon emissions of industrial equipment based on dynamic monitoring of operating status according to claim 5 is characterized in that: The calculation expression of the feature optimization using the adaptive sparse filtering optimization algorithm is: Among them, A is the optimized user power consumption feature vector, W is the weight matrix, RS is the L2 normalized matrix of each row of A, CS is the result of L2 normalization of each column of RS, and P represents the sparsity adjustment parameter; by optimizing the sample characteristics and minimizing CS as the goal, the optimized user power consumption characteristics are solved.

7. The method for calculating carbon emissions of industrial equipment based on dynamic monitoring of operating status according to any one of claims 1 to 4, characterized in that: The equipment operation status monitoring model extracts power consumption features from the input user power consumption data, imports them into the convolutional neural network after sparse filtering, and performs feature extraction through the fusion layer. The optimized power consumption features are extracted to measure the equipment load status, and the optimized power consumption features are passed through the Sigmoid activation function to obtain the equipment status recognition result.

8. The method for calculating carbon emissions of industrial equipment based on dynamic monitoring of operating status according to any one of claims 1 to 4, characterized in that: The step of calculating the carbon emissions of the device under test within the specified time period according to the real-time operating status code of the device under test and the unit carbon emissions factor under the corresponding operating status within the specified time period includes: Obtain all real-time running status codes of the device under test identified within a specified time period, and form a running status vector from all the identified real-time running status codes, wherein the running status code is a one-hot code; Obtaining a unit carbon emission factor corresponding to each real-time operating state code to form a carbon emission factor vector, wherein the unit carbon emission factor is a unit carbon emission under different operating states; The operating state vector is multiplied by the carbon emission factor vector and the carbon emission factor vector to obtain the carbon emission of the device under test within a specified time period.

9. A device for calculating carbon emissions of industrial equipment based on dynamic monitoring of operating status, comprising a processor and a memory, wherein the memory is used to store a computer program, characterized in that: The processor is configured to execute the computer program to perform the method according to any one of claims 1 to 8.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed, the method according to any one of claims 1 to 8 is implemented.

Images