Public building energy efficiency diagnosis system and method

Abstract

The present invention belongs to the technical field of building energy conservation, and discloses a public building energy conservation diagnosis system and method. The method includes: obtaining historical energy consumption observation data of public buildings in a target area, grouping the historical energy consumption observation data according to preset influencing factors to obtain initial sample data for each group; determining valid sample data in the initial sample data for each group and generating incremental sample data for each group; establishing an energy consumption prediction model based on the incremental sample data and the valid sample data for each group; inputting the current building information, current environmental information, and current time information of a target building into the energy consumption prediction model to obtain energy consumption prediction data; obtaining an energy conservation diagnosis result based on the energy consumption prediction data and the current energy consumption observation data, and generating a public building energy conservation strategy for the target area. Through the above method, an energy consumption prediction model is established using limited data, energy conservation diagnosis in different scenarios is realized, the model establishment process is optimized, and the accuracy of diagnosis is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of building energy conservation, and particularly to an energy conservation diagnosis system and method for public buildings. Background Art

[0002] With the development of society, building energy consumption has been continuously increasing, and the energy management of public buildings, as major energy consumers, has received social attention. And the prediction of public building energy consumption is an important part in the process of public building management and building energy conservation, and is of great significance for improving building energy utilization efficiency and protecting the environment.

[0003] Currently, building energy consumption diagnosis methods are mainly divided into two methods: black box and grey box. Although the black box method can give the energy consumption anomaly detection results by algorithms based on a large amount of data, it requires a large amount of high-quality data as support, requires that building data has been recorded for a period of time and needs to include various types of anomalies, and lacks certain theoretical support, that is, it only compares with the past energy consumption situation of the building and cannot reflect the energy consumption that can be saved; while the simulation results of the grey box method have theoretical nature, but the construction of the white box therein is a key and difficult point, and the establishment of a building energy consumption model requires more building information. Therefore, it is necessary to establish a more scientific energy conservation diagnosis method for public buildings to effectively achieve building energy conservation.

[0004] The above content is only used to assist in understanding the technical solution of the present invention, and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main object of the present invention is to provide an energy conservation diagnosis system and method for public buildings, aiming to solve the technical problems in the prior art that traditional building energy consumption diagnosis methods require a large amount of high-quality and comprehensive data as support, and the establishment of a building energy consumption model is relatively complex.

[0006] To achieve the above object, the present invention provides an energy conservation diagnosis method for public buildings, and the method includes the following steps:

[0007] Obtain historical energy consumption observation data of public buildings in a target area, and group the historical energy consumption observation data according to preset influencing factors to obtain initial sample data for each group, where the preset influencing factors include building information, environmental information and time information;

[0008] Perform anomaly detection on the initial sample data for each group to determine valid sample data in the initial sample data for each group;

[0009] Generate incremental sample data for each group according to the valid sample data for each group;

[0010] Establish an energy consumption prediction model according to the incremental sample data for each group and the valid sample data for each group;

[0011] Input the current building information, current environmental information, and current time information of the target building into the energy consumption prediction model to obtain energy consumption prediction data;

[0012] Obtain the current energy consumption observation data of the target building, and perform energy conservation diagnosis based on the energy consumption prediction data and the current energy consumption observation data to obtain an energy conservation diagnosis result;

[0013] Generate an energy conservation strategy for public buildings in the target area according to the energy conservation diagnosis result.

[0014] Optionally, establishing an energy consumption prediction model according to the groups of incremental sample data and groups of valid sample data includes:

[0015] Determine an initial energy consumption prediction model according to a preset neural network model;

[0016] Determine input sample data and target sample data according to the groups of incremental sample data and groups of valid sample data;

[0017] Train the initial energy consumption prediction model according to the input sample data and the target sample data to obtain an energy consumption prediction model.

[0018] Optionally, training the initial energy consumption prediction model according to the input sample data and the target sample data to obtain an energy consumption prediction model includes:

[0019] Input the input sample data into the initial energy consumption prediction model to obtain sample output data;

[0020] Determine a loss function according to the sample output data and the target sample data;

[0021] Adjust the weights of the initial energy consumption prediction model according to the loss function to obtain an optimized energy consumption prediction model;

[0022] Input the input sample data into the optimized energy consumption prediction model to obtain optimized output data;

[0023] Determine a new loss function according to the optimized output data and the target sample data;

[0024] When the new loss function meets the preset optimization goal, determine an energy consumption prediction model according to the optimized energy consumption prediction model.

[0025] Optionally, after determining the new loss function according to the optimized output data and the target sample data, it further includes:

[0026] When the new loss function does not meet the preset optimization objective, adjust the weights of the optimized energy consumption prediction model according to the new loss function to obtain a new optimized energy consumption prediction model;

[0027] Input the input sample data into the new optimized energy consumption prediction model to obtain new optimized output data;

[0028] According to the new optimized output data, return to execute the step of determining a new loss function according to the optimized output data and the target sample data.

[0029] Optionally, the obtaining the current energy consumption observation data of the target building and performing energy conservation diagnosis according to the energy consumption prediction data and the current energy consumption observation data to obtain an energy conservation diagnosis result includes:

[0030] Obtain the current energy consumption observation data of the target building;

[0031] Compare the energy consumption prediction data with the current energy consumption observation data to obtain difference data;

[0032] When the difference data does not meet the preset data range, determine that the energy consumption of the target building is in an abnormal state, and determine that the energy conservation diagnosis result is that the target building needs energy conservation.

[0033] Optionally, after comparing the energy consumption prediction data with the current energy consumption observation data to obtain difference data, it further includes:

[0034] When the difference data meets the preset data range, determine that the energy consumption of the target building is in a normal state, and determine that the energy conservation diagnosis result is that the target building does not need energy conservation.

[0035] Optionally, the generating a public building energy conservation strategy in the target area according to the energy conservation diagnosis result includes:

[0036] According to the energy conservation diagnosis result, determine the public buildings in the target area that need energy conservation, and generate energy conservation summary information;

[0037] According to the energy conservation summary information, determine the public building energy conservation strategy in the target area.

[0038] Optionally, the performing anomaly detection on each group of initial sample data to determine the valid sample data in each group of initial sample data includes:

[0039] Perform anomaly detection on each group of initial sample data according to a preset anomaly detection algorithm to determine each group of abnormal sample data;

[0040] According to each group of initial sample data and each group of abnormal sample data, obtain each group of valid sample data.

[0041] Optionally, generating incremental sample data for each group according to the valid sample data of each group includes:

[0042] Generating the sample distribution of each group according to the valid sample data of each group;

[0043] Generating incremental sample data for each group according to the sample distribution of each group.

[0044] In addition, to achieve the above object, the present invention also proposes a public building energy conservation diagnosis system, and the public building energy conservation diagnosis system includes:

[0045] A sample acquisition module, configured to acquire historical energy consumption observation data of public buildings in a target area, and group the historical energy consumption observation data according to preset influencing factors to obtain initial sample data for each group, where the preset influencing factors include building type, energy consumption type, environmental type, and time type;

[0046] A sample generation module, configured to perform anomaly detection on the initial sample data of each group to determine the valid sample data in the initial sample data of each group;

[0047] The sample generation module is further configured to generate incremental sample data for each group according to the valid sample data of each group;

[0048] An energy consumption prediction module, configured to establish an energy consumption prediction model according to the incremental sample data of each group and the valid sample data of each group;

[0049] The energy consumption prediction module is further configured to input the current building type, current environmental information, and current time information of a target building into the energy consumption prediction model to obtain energy consumption prediction data;

[0050] An energy conservation diagnosis module, configured to acquire the current energy consumption observation data of the target building, and perform energy conservation diagnosis according to the energy consumption prediction data and the current energy consumption observation data to obtain an energy conservation diagnosis result;

[0051] The energy conservation diagnosis module is further configured to generate an energy conservation strategy for public buildings in the target area according to the energy conservation diagnosis result.

[0052] In addition, to achieve the above object, the present invention also proposes a public building energy conservation diagnosis device, and the public building energy conservation diagnosis device includes: a memory, a processor, and a public building energy conservation diagnosis program stored on the memory and executable on the processor, where the public building energy conservation diagnosis program is configured to implement the steps of the public building energy conservation diagnosis method as described above.

[0053] In addition, to achieve the above object, the present invention also provides a storage medium, on which a public building energy conservation diagnosis program is stored. When the public building energy conservation diagnosis program is executed by a processor, the steps of the public building energy conservation diagnosis method described above are implemented.

[0054] In the present invention, by obtaining the historical energy consumption observation data of public buildings in the target area, grouping the historical energy consumption observation data according to preset influencing factors to obtain initial sample data for each group, performing anomaly detection on the initial sample data for each group to determine the valid sample data in the initial sample data for each group, generating incremental sample data for each group according to the valid sample data for each group, establishing an energy consumption prediction model based on the incremental sample data for each group and the valid sample data for each group, inputting the current building information, current environmental information, and current time information of the target building into the energy consumption prediction model to obtain energy consumption prediction data, obtaining the current energy consumption observation data of the target building, obtaining an energy conservation diagnosis result based on the energy consumption prediction data and the current energy consumption observation data, and generating a public building energy conservation strategy in the target area according to the energy conservation diagnosis result. Compared with traditional building energy consumption diagnosis methods that require a large amount of high-quality and comprehensive data as support and the establishment of building energy consumption models is relatively complex, the present invention can generate appropriate new samples based on existing data, establish an energy consumption prediction model using limited data, optimize the model establishment process, and at the same time can achieve energy conservation diagnosis in different scenarios, effectively improving the accuracy and comprehensiveness of the diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 is a schematic structural diagram of a public building energy conservation diagnosis device in the hardware operating environment related to the embodiment of the present invention;

[0056] Figure 2 is a schematic flowchart of the first embodiment of the public building energy conservation diagnosis method of the present invention;

[0057] Figure 3 is a schematic flowchart of the second embodiment of the public building energy conservation diagnosis method of the present invention;

[0058] Figure 4 is a structural block diagram of the first embodiment of the public building energy conservation diagnosis system of the present invention.

[0059] The realization, functional characteristics, and advantages of the object of the present invention will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0060] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0061] Refer to Figure 1 , Figure 1This is a schematic structural diagram of a public building energy-saving diagnosis device for the hardware operating environment involved in the solution of the embodiment of the present invention.

[0062] As shown in Figure 1 , the public building energy-saving diagnosis device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display) and an input unit such as a keyboard (Keyboard). Optionally, the user interface 1003 may further include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless-fidelity (Wi-Fi) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM) or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk memory. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0063] Those skilled in the art can understand that Figure 1 the structure shown in

[0064] As shown in Figure 1 does not constitute a limitation on the public building energy-saving diagnosis device, and may include more or fewer components than shown in the figure, or combine some components, or arrange different components.

[0065] In Figure 1 the public building energy-saving diagnosis device shown, the network interface 1004 is mainly used for data communication with a network server; the user interface 1003 is mainly used for data interaction with a user; the processor 1001 and the memory 1005 in the public building energy-saving diagnosis device of the present invention may be arranged in the public building energy-saving diagnosis device. The public building energy-saving diagnosis device calls the public building energy-saving diagnosis program stored in the memory 1005 through the processor 1001 and executes the public building energy-saving diagnosis method provided by the embodiment of the present invention.

[0066] The embodiment of the present invention provides a public building energy-saving diagnosis method. Referring to Figure 2 , Figure 2 This is a schematic flowchart of the first embodiment of a public building energy-saving diagnosis method of the present invention.

[0067] In this embodiment, the public building energy conservation diagnosis method includes the following steps:

[0068] Step S10: Obtain the historical energy consumption observation data of public buildings in the target area, and group the historical energy consumption observation data according to preset influencing factors to obtain initial sample data for each group. The preset influencing factors include building information, environmental information, and time information.

[0069] It should be noted that the execution subject of this embodiment is a computer, which can be any computer capable of executing the public building energy conservation diagnosis program. This embodiment does not limit this. Through the public building energy conservation diagnosis program, the energy conservation diagnosis of public buildings is realized.

[0070] It can be understood that the target area is a preset area range. For example: public buildings within 1 km nearby, public buildings in a certain park. This embodiment does not limit this, and the energy conservation diagnosis of public buildings is carried out within this area range. The historical energy consumption observation data is the historical data of building energy consumption. In recent years, the energy consumption monitoring systems and data platforms for public buildings have been continuously developed, and relevant data has been collected through sensors and the Internet, and a certain data foundation has been obtained. The initial sample data is a sample set composed of each group of data after the historical data is grouped, and each group has corresponding initial sample data.

[0071] It should be understood that the preset influencing factors are external factors related to energy consumption, including building information, environmental information, and time information. The building information refers to information related to the building itself, including the type of building, the area of the building, etc. The type of building can be an office building (including office buildings, government department offices, etc.), a commercial building (including shopping malls, financial buildings, etc.), a tourism building (including hotels, entertainment venues, etc.), a science, education, culture, health and sports building (including cultural, educational, scientific research, medical, health, and sports buildings, etc.), a communication building (including post and telecommunications, communication, data centers, broadcasting rooms), a transportation building (including airports, high-speed railway stations, railway stations, subways, bus stations, etc.), etc. The environmental information refers to information related to the external environment of the building, including meteorological conditions, temperature conditions, and humidity conditions. The meteorological conditions can be sunny, cloudy, rainy, snowy, typhoon, etc. The temperature conditions can be represented by levels. For example, when the temperature is less than or equal to 0°C, it is the first-level temperature; when the temperature is greater than 0°C and less than or equal to 10°C, it is the second-level temperature; when the temperature is greater than 10°C and less than or equal to 20°C, it is the third-level temperature; when the temperature is greater than 20°C and less than or equal to 30°C, it is the fourth-level temperature; when the temperature is greater than 30°C and less than or equal to 40°C, it is the fifth-level temperature; when the temperature is greater than or equal to 40°C, it is the sixth-level temperature. The humidity conditions can also be represented by levels. For example, when the humidity is less than or equal to 20%, it is the first-level humidity; when the humidity is greater than 20% and less than or equal to 40%, it is the second-level humidity; when the humidity is greater than 40% and less than or equal to 60%, it is the third-level humidity; when the humidity is greater than 60% and less than or equal to 80%, it is the fourth-level humidity; when the humidity is greater than or equal to 80%, it is the fifth-level humidity. Temperature and humidity can also be described in other ways, and this embodiment does not limit this. The time information includes date attributes and time ranges. The date attributes can be holidays, rest days, working days, etc. The time ranges can be morning, noon, afternoon, and evening. In this embodiment, historical data is grouped according to different scenarios (different buildings, different environments, different times). For example, the energy consumption generated by an office building on a sunny morning on a working day, at the third-level temperature, and the second-level humidity condition, and the energy consumption generated by a commercial building on a rainy evening on a holiday, at the second-level temperature, and the third-level humidity condition. This embodiment does not limit this.

[0072] In specific implementation, historical data of the energy consumption of all public buildings in the target area is obtained, and the historical data is grouped according to different building information, different environmental information, and different time information to obtain initial sample data for each group under different scenarios.

[0073] Step S20: Perform anomaly detection on the initial sample data of each group to determine the valid sample data in the initial sample data of each group.

[0074] Further, the step S20 includes: performing anomaly detection on each group of initial sample data according to a preset anomaly detection algorithm to determine each group of anomalous sample data, and obtaining each group of valid sample data based on the initial sample data of each group and the anomalous sample data of each group.

[0075] It should be noted that the preset anomaly detection algorithm in this embodiment is the isolation forest algorithm. The isolation forest algorithm does not need to calculate indicators related to distance and density, which can greatly improve the speed. Other anomaly detection algorithms can also be used. This embodiment does not limit this and can be selected according to actual needs. The anomalous sample data are the anomalous data that deviate from other samples in the initial sample data, and the valid sample data are the remaining valid data after removing the anomalous data from the initial sample data. Each group of initial sample data has corresponding anomalous sample data and valid sample data.

[0076] It can be understood that the building energy consumption in different scenarios is different and there are differences. If the differences are large, normal data may be regarded as anomalous data during anomaly detection. Therefore, in this embodiment, the sample data are first grouped according to different buildings, different environments, and different times, and then anomaly detection is performed on each group of data respectively to improve the detection accuracy.

[0077] It should be understood that through the isolation forest algorithm, the anomaly score of the data can be calculated. If the anomaly score is close to 1, then it must be anomalous data. If the anomaly score is much less than 0.5, then it must not be anomalous data. If the scores of all data are around 0.5, then there are probably no anomalous data in the sample.

[0078] In specific implementation, in this embodiment, anomaly detection is performed on the grouped initial sample data in different scenarios. During this process, the isolation forest algorithm is used to find the anomalous data in each group according to the calculated data anomaly scores, and these anomalous data are removed from the initial sample data to obtain the valid sample data of each group.

[0079] Step S30: Generating each group of incremental sample data according to the valid sample data of each group.

[0080] Further, the step S30 includes: generating the sample distribution of each group according to the valid sample data of each group, and generating each group of incremental sample data according to the sample distribution of each group.

[0081] It should be noted that the sample distribution refers to the distribution function of the valid sample data, which can reflect the distribution of all data points. The incremental sample data are the new samples generated according to the valid sample data. Each group of valid sample data has corresponding incremental sample data.

[0082] It can be understood that the quantity of the incremental sample data can be set in the energy conservation diagnosis program for public buildings, for example: 200, 250, etc. This embodiment does not limit this, and it can be flexibly adjusted according to the actual situation. Since the quantity of each group of effective sample data is not necessarily the same, and the group with less effective sample data requires more new samples, therefore, the incremental sample data to be generated for different groups can be different, and can be set separately for different groups in the energy conservation diagnosis program for public buildings.

[0083] In specific implementation, according to the distribution law of the effective sample data, the corresponding distribution function is calculated, and according to the distribution function obtained for each group and the set generation quantity for each group, several new samples conforming to the distribution law of each group are generated for subsequent model training.

[0084] Step S40: Establish an energy consumption prediction model according to the incremental sample data and the effective sample data of each group.

[0085] It should be noted that the energy consumption prediction model refers to a model that can predict building energy consumption.

[0086] In specific implementation, the obtained incremental sample data and effective sample data of each group are used as samples, and an energy consumption prediction model is obtained through training.

[0087] Step S50: Input the current building information, current environmental information, and current time information of the target building into the energy consumption prediction model to obtain energy consumption prediction data.

[0088] It can be understood that the target building refers to a building selected from public buildings in the target area. The current building information is the building information of the target building, the current environmental information is the environmental information of the target building in the current state, and the current time information is the current time information. For example: if the target building is a shopping mall, and it is Saturday afternoon currently, the weather is sunny, the temperature is 26 degrees Celsius, and the humidity is 23%, then the current environmental information is a commercial building, the current environmental information is sunny, four-level temperature, and second-level humidity, and the current time information is a public holiday, afternoon. The energy consumption prediction data is the energy consumption prediction result obtained by using the energy consumption prediction model, which is a predicted value.

[0089] In specific implementation, input the current building information, current environmental information, and current time information corresponding to the target building into the trained energy consumption prediction model, and the energy consumption prediction result output by the model can be obtained.

[0090] Step S60: Obtain the current energy consumption observation data of the target building, and perform energy conservation diagnosis according to the energy consumption prediction data and the current energy consumption observation data to obtain an energy conservation diagnosis result.

[0091] Further, the step S60 includes: obtaining the current energy consumption observation data of the target building, comparing the energy consumption prediction data with the current energy consumption observation data to obtain difference data, and when the difference data does not meet the preset data range, determining that the energy consumption of the target building is in an abnormal state, and determining that the energy conservation diagnosis result is that the target building needs energy conservation.

[0092] It should be understood that when the difference data meets the preset data range, it is determined that the energy consumption of the target building is in a normal state, and it is determined that the energy conservation diagnosis result is that the target building does not need energy conservation.

[0093] It should be noted that the current energy consumption observation data is the current energy consumption data of the target building, which is the true value. The difference data refers to the difference between the predicted value and the true value. The preset data range is the error range allowed for the prediction model. Due to the inevitable errors in the prediction model, even when the building energy consumption is normal, there will be a certain error between the predicted value and the true value. Therefore, in this embodiment, a preset data range is set to allow a certain deviation between the predicted value and the true value, for example: 3%, 5%, which can be set in the energy conservation diagnosis program for public buildings. This embodiment does not limit this and can be adjusted according to actual needs. The energy conservation diagnosis result is the evaluation of the current energy consumption of the building, judging whether the building energy consumption is normal.

[0094] In a specific implementation, the energy consumption of the target building is subjected to energy conservation diagnosis, and the obtained energy consumption prediction result is compared with the actual energy consumption data. If the energy consumption prediction result is within the allowed error range, it is considered that the building energy consumption is normal and no further energy conservation is required. If the energy consumption prediction result is not within the allowed error range, it is considered that the difference between the true value and the predicted value is large. At this time, the building energy consumption is abnormal and energy conservation treatment is required.

[0095] Step S70: Generate an energy conservation strategy for public buildings in the target area according to the energy conservation diagnosis result.

[0096] Further, the step S70 includes: determining the public buildings that need energy conservation in the target area according to the energy conservation diagnosis result, generating energy conservation summary information, and determining the energy conservation strategy for public buildings in the target area according to the energy conservation summary information.

[0097] It can be understood that the energy conservation summary information refers to the information that summarizes all the public buildings that need energy conservation in the target area, which can be in the form of a table or described in other forms. The energy conservation strategy for public buildings is the energy conservation measures taken for all public buildings in the target area. For example, office buildings can turn off unused electrical equipment at night, including lights, elevators, air conditioners, etc. Commercial buildings can appropriately turn off some elevators in the morning and adjust the air conditioner temperature to the energy conservation state.

[0098] It should be understood that if energy conservation diagnosis is carried out on all public buildings in the target area, the energy conservation diagnosis results of all public buildings in the target area can be obtained.

[0099] In specific implementation, obtain the energy conservation diagnosis results of all public buildings in the target area, judge all buildings that need energy conservation treatment according to these diagnosis results, generate summary information, and formulate corresponding energy conservation measures according to the summary information, so that the public buildings in the target area can all achieve energy conservation.

[0100] In this embodiment, by obtaining the historical energy consumption observation data of public buildings in the target area, grouping the historical energy consumption observation data according to preset influencing factors to obtain initial sample data for each group, performing anomaly detection on the initial sample data for each group to determine the valid sample data in the initial sample data for each group, generating incremental sample data for each group according to the valid sample data for each group, establishing an energy consumption prediction model according to the incremental sample data for each group and the valid sample data for each group, inputting the current building information, current environment information and current time information of the target building into the energy consumption prediction model to obtain energy consumption prediction data, obtaining the current energy consumption observation data of the target building, obtaining the energy conservation diagnosis result according to the energy consumption prediction data and the current energy consumption observation data, and generating an energy conservation strategy for public buildings in the target area according to the energy conservation diagnosis result. This embodiment does not require a large amount of high-quality and comprehensive data as support, can generate appropriate new samples according to existing data, can establish an energy consumption prediction model using limited data, and optimizes the model establishment process. The energy consumption prediction model can achieve energy conservation diagnosis in different scenarios and improve the accuracy of diagnosis.

[0101] Reference Figure 3 , Figure 3 is a schematic flowchart of the second embodiment of a method for energy conservation diagnosis of public buildings according to the present invention.

[0102] Based on the above first embodiment, the step S40 includes:

[0103] Step S401: Determine an initial energy consumption prediction model according to a preset neural network model.

[0104] It should be noted that the preset neural network model in this embodiment is a BP (Back Propagation) neural network model, and other types of neural network models can also be used. This embodiment does not limit this and can be selected according to actual situations. The initial energy consumption prediction model refers to an initial model that has not been trained.

[0105] In specific implementation, use the BP neural network model as the initial energy consumption prediction model for subsequent model training.

[0106] Step S402: Determine the input sample data and the target sample data according to the groups of incremental sample data and the groups of valid sample data.

[0107] It can be understood that the input sample data refers to the data that needs to be input into the model during model training, and the target sample data is the target data that is expected to be output during model training.

[0108] It should be understood that in this embodiment, the building information, environmental information, and time information of each group are used as the input data of the model, and the incremental sample data and valid sample data corresponding to each group of building information, environmental information, and time information are used as the target data output by the model. As a result, for the subsequent established energy consumption prediction model, the input is the building information, environmental information, and time information, and the output is the predicted energy consumption data, and different energy consumption prediction results can be obtained according to different scenarios.

[0109] In a specific implementation, according to the obtained groups of incremental sample data and the groups of valid sample data, the building information, environmental information, and time information of each group are used as the model input, and the corresponding incremental sample data and valid sample data of each group are used as the output target to train the energy consumption prediction model.

[0110] Step S403: Train the initial energy consumption prediction model according to the input sample data and the target sample data to obtain an energy consumption prediction model.

[0111] Further, the step S403 includes: inputting the input sample data into the initial energy consumption prediction model to obtain sample output data, determining a loss function according to the sample output data and the target sample data, adjusting the weights of the initial energy consumption prediction model according to the loss function to obtain an optimized energy consumption prediction model, inputting the input sample data into the optimized energy consumption prediction model to obtain optimized output data, determining a new loss function according to the optimized output data and the target sample data, and when the new loss function meets the preset optimization goal, determining the energy consumption prediction model according to the optimized energy consumption prediction model.

[0112] It should be noted that the sample output data is the output data obtained after the input sample data is input into the initial energy consumption prediction model, the loss function is the error between the sample output data and the target sample data, the optimized energy consumption prediction model is the energy consumption prediction model during the optimization process, the optimized output data is the output data of the optimized energy consumption prediction model, and the preset optimization goal is the pre-set training completion state, for example: the loss function reaches the minimum.

[0113] In a specific implementation, the input sample data is input into the initial energy consumption prediction model to obtain output data. According to the difference between the sample output data and the target sample data, a loss function is calculated. The loss function optimizes and adjusts the weights of the initial energy consumption prediction model in the manner of error gradient descent to obtain an optimized energy consumption prediction model after preliminary optimization. The sample data is input again, and the optimized output data and the target sample data are used to calculate a new loss function. When the new loss function reaches the minimum, it is considered that the training is completed at this time, and the final energy consumption prediction model is generated.

[0114] It can be understood that when the new loss function does not meet the preset optimization target, according to the new loss function, the weights of the optimized energy consumption prediction model are adjusted to obtain a new optimized energy consumption prediction model. The input sample data is input into the new optimized energy consumption prediction model to obtain new optimized output data. According to the new optimized output data, the step of determining a new loss function according to the optimized output data and the target sample data is returned for execution.

[0115] In a specific implementation, if the obtained new loss function does not reach the minimum, it is necessary to continue training the model, optimize and adjust the weights of the model, and perform iterative training until the loss function reaches the minimum.

[0116] In this embodiment, according to a preset neural network model, an initial energy consumption prediction model is determined. According to each group of incremental sample data and each group of valid sample data, input sample data and target sample data are determined. The initial energy consumption prediction model is trained according to the input sample data and the target sample data to obtain an energy consumption prediction model. In this embodiment, the building information, environmental information, and time information of each group are used as model inputs, and the corresponding incremental sample data and valid sample data of each group are used as output targets to train an energy consumption prediction model with the minimum loss function, which can predict the energy consumption in different scenarios and improve the accuracy and comprehensiveness of the prediction.

[0117] In addition, an embodiment of the present invention also proposes a storage medium, on which a public building energy conservation diagnosis program is stored. When the public building energy conservation diagnosis program is executed by a processor, the steps of the public building energy conservation diagnosis method described above are implemented.

[0118] Refer to Figure 4 , Figure 4 which is the structural block diagram of the first embodiment of the public building energy conservation diagnosis system of the present invention.

[0119] As Figure 4 shown, the public building energy conservation diagnosis system proposed by the embodiment of the present invention includes:

[0120] A sample acquisition module 10, configured to obtain historical energy consumption observation data of public buildings in a target area, group the historical energy consumption observation data according to preset influencing factors, and obtain initial sample data for each group. The preset influencing factors include building type, energy consumption type, environmental type, and time type.

[0121] A sample generation module 20, configured to perform anomaly detection on the initial sample data for each group to determine valid sample data in the initial sample data for each group.

[0122] The sample generation module 20 is further configured to generate incremental sample data for each group according to the valid sample data for each group.

[0123] An energy consumption prediction module 30, configured to establish an energy consumption prediction model according to the incremental sample data for each group and the valid sample data for each group.

[0124] The energy consumption prediction module 30 is further configured to input the current building type, current environmental information, and current time information of a target building into the energy consumption prediction model to obtain energy consumption prediction data.

[0125] An energy conservation diagnosis module 40, configured to obtain the current energy consumption observation data of the target building, perform energy conservation diagnosis according to the energy consumption prediction data and the current energy consumption observation data, and obtain an energy conservation diagnosis result.

[0126] The energy conservation diagnosis module 40 is further configured to generate an energy conservation strategy for public buildings in the target area according to the energy conservation diagnosis result.

[0127] In this embodiment, by obtaining the historical energy consumption observation data of public buildings in the target area, grouping the historical energy consumption observation data according to preset influencing factors to obtain initial sample data for each group, performing anomaly detection on the initial sample data for each group to determine valid sample data in the initial sample data for each group, generating incremental sample data for each group according to the valid sample data for each group, establishing an energy consumption prediction model according to the incremental sample data for each group and the valid sample data for each group, inputting the current building information, current environmental information, and current time information of the target building into the energy consumption prediction model to obtain energy consumption prediction data, obtaining the current energy consumption observation data of the target building, obtaining an energy conservation diagnosis result according to the energy consumption prediction data and the current energy consumption observation data, and generating an energy conservation strategy for public buildings in the target area according to the energy conservation diagnosis result. This embodiment does not require a large amount of high-quality and comprehensive data as support, can generate appropriate new samples according to existing data, establish an energy consumption prediction model using limited data, and optimizes the model establishment process. The energy consumption prediction model can achieve energy conservation diagnosis in different scenarios and improve the accuracy of diagnosis.

[0128] In one embodiment, the energy consumption prediction module 30 is further configured to determine an initial energy consumption prediction model according to a preset neural network model;

[0129] According to each group of incremental sample data and each group of valid sample data, determine input sample data and target sample data;

[0130] Train the initial energy consumption prediction model according to the input sample data and the target sample data to obtain an energy consumption prediction model.

[0131] In one embodiment, the energy consumption prediction module 30 is further configured to input the input sample data into the initial energy consumption prediction model to obtain sample output data;

[0132] Determine a loss function according to the sample output data and the target sample data;

[0133] Adjust the weights of the initial energy consumption prediction model according to the loss function to obtain an optimized energy consumption prediction model;

[0134] Input the input sample data into the optimized energy consumption prediction model to obtain optimized output data;

[0135] Determine a new loss function according to the optimized output data and the target sample data;

[0136] When the new loss function meets a preset optimization target, determine an energy consumption prediction model according to the optimized energy consumption prediction model.

[0137] In one embodiment, when the new loss function does not meet the preset optimization target, the energy consumption prediction module 30 is further configured to adjust the weights of the optimized energy consumption prediction model according to the new loss function to obtain a new optimized energy consumption prediction model;

[0138] Input the input sample data into the new optimized energy consumption prediction model to obtain new optimized output data;

[0139] According to the new optimized output data, return to execute the step of determining a new loss function according to the optimized output data and the target sample data.

[0140] In one embodiment, the energy conservation diagnosis module 40 is further configured to obtain current energy consumption observation data of the target building;

[0141] Compare the energy consumption prediction data with the current energy consumption observation data to obtain difference data;

[0142] When the difference data does not conform to a preset data range, determine that the energy consumption of the target building is in an abnormal state, and determine that the energy conservation diagnosis result is that the target building needs energy conservation.

[0143] In one embodiment, the energy-saving diagnosis module 40 is further configured to determine that the energy consumption of the target building is in a normal state when the difference data meets a preset data range, and determine that the energy-saving diagnosis result is that the target building does not require energy saving.

[0144] In one embodiment, the energy-saving diagnosis module 40 is further configured to determine public buildings that require energy saving in the target area according to the energy-saving diagnosis result, and generate energy-saving summary information;

[0145] Determine the energy-saving strategies for public buildings in the target area according to the energy-saving summary information.

[0146] In one embodiment, the sample generation module 20 is further configured to perform anomaly detection on each group of initial sample data according to a preset anomaly detection algorithm to determine each group of abnormal sample data;

[0147] Obtain each group of effective sample data according to each group of initial sample data and each group of abnormal sample data.

[0148] In one embodiment, the sample generation module 20 is further configured to generate the sample distribution of each group according to each group of effective sample data;

[0149] Generate each group of incremental sample data according to the sample distribution of each group.

[0150] It should be understood that the above is only an example for illustration and does not constitute any limitation to the technical solution of the present invention. In specific applications, those skilled in the art can set according to needs, and the present invention does not limit this.

[0151] It should be noted that the above-described work process is only illustrative and does not limit the protection scope of the present invention. In actual applications, those skilled in the art can select some or all of them according to actual needs to achieve the purpose of the solution of this embodiment, and there is no limitation here.

[0152] In addition, for technical details not described in detail in this embodiment, reference can be made to the public building energy-saving diagnosis method provided in any embodiment of the present invention, which will not be elaborated here.

[0153] In addition, it should be noted that in this text, the terms "including", "comprising" or any other variants thereof are intended to cover non-exclusive inclusion, such that a process, method, article or system including a series of elements not only includes those elements but also other elements not explicitly listed, or further includes elements inherent to such a process, method, article or system. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, article or system including that element.

[0154] The serial numbers of the above embodiments of the present invention are only for description and do not represent the superiority or inferiority of the embodiments.

[0155] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-described embodiment methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium (such as a read-only memory (ROM) / RAM, magnetic disk, optical disk), and includes several instructions for causing a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0156] The above are only the preferred embodiments of the present invention and do not limit the patent scope of the present invention accordingly. Any equivalent structural or equivalent process transformation made by using the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall equally be included in the patent protection scope of the present invention.

Claims

1. A method for energy conservation diagnosis of public buildings, characterized in that, the method for energy conservation diagnosis of public buildings includes: Obtain the historical energy consumption observation data of public buildings in the target area, group the historical energy consumption observation data according to preset influencing factors to obtain initial sample data for each group. The preset influencing factors include building information, environmental information and time information. The building information includes the type of building, the area of the building. The types of buildings include office buildings, commercial buildings, tourism buildings, science, education, culture, health and communication buildings, and transportation buildings. Office buildings include office buildings and government department offices. Commercial buildings include shopping malls and financial buildings. Tourism buildings include hotels and entertainment venues. Science, education, culture, health and communication buildings include cultural, educational, scientific research, medical, health and sports buildings. Communication buildings include post and telecommunications, communication, data centers and broadcasting rooms. Transportation buildings include airports, high-speed railway stations, railway stations, subways and bus stations. The environmental information includes meteorological conditions, temperature conditions and humidity conditions. The time information includes date attributes and time ranges; Perform anomaly detection on the initial sample data for each group to determine the valid sample data in the initial sample data for each group; Generate incremental sample data for each group according to the valid sample data for each group; Establish an energy consumption prediction model according to the incremental sample data for each group and the valid sample data for each group. The energy consumption prediction model is applicable to different scenarios; Input the current building information, current environmental information and current time information of the target building into the energy consumption prediction model to obtain energy consumption prediction data; Obtain the current energy consumption observation data of the target building, and perform energy conservation diagnosis according to the energy consumption prediction data and the current energy consumption observation data to obtain an energy conservation diagnosis result; Generate an energy conservation strategy for public buildings in the target area according to the energy conservation diagnosis result; The generating incremental sample data for each group according to the valid sample data for each group includes: Generate the sample distribution for each group according to the valid sample data for each group; Generate incremental sample data for each group according to the sample distribution for each group.

2. The method according to claim 1, characterized in that, the establishing an energy consumption prediction model according to the incremental sample data for each group and the valid sample data for each group includes: Determine an initial energy consumption prediction model according to a preset neural network model; Determine input sample data and target sample data according to the incremental sample data for each group and the valid sample data for each group; Train the initial energy consumption prediction model according to the input sample data and the target sample data to obtain an energy consumption prediction model.

3. The method according to claim 2, characterized in that, the training the initial energy consumption prediction model according to the input sample data and the target sample data to obtain an energy consumption prediction model includes: Input the input sample data into the initial energy consumption prediction model to obtain sample output data; Determine a loss function according to the sample output data and the target sample data; Adjust the weights of the initial energy consumption prediction model according to the loss function to obtain an optimized energy consumption prediction model; Input the input sample data into the optimized energy consumption prediction model to obtain optimized output data; Determine a new loss function according to the optimized output data and the target sample data; When the new loss function meets the preset optimization objective, determine an energy consumption prediction model according to the optimized energy consumption prediction model.

4. The method according to claim 3, wherein, after determining the new loss function according to the optimized output data and the target sample data, further includes: When the new loss function does not meet the preset optimization objective, adjust the weights of the optimized energy consumption prediction model according to the new loss function to obtain a new optimized energy consumption prediction model; Input the input sample data into the new optimized energy consumption prediction model to obtain new optimized output data; According to the new optimized output data, return to execute the step of determining a new loss function according to the optimized output data and the target sample data.

5. The method according to claim 1, wherein, obtaining the current energy consumption observation data of the target building, and performing energy conservation diagnosis according to the energy consumption prediction data and the current energy consumption observation data to obtain an energy conservation diagnosis result, including: Obtain the current energy consumption observation data of the target building; Compare the energy consumption prediction data with the current energy consumption observation data to obtain difference data; When the difference data does not conform to the preset data range, determine that the energy consumption of the target building is in an abnormal state, and determine that the energy conservation diagnosis result is that the target building needs energy conservation.

6. The method according to claim 5, wherein, after comparing the energy consumption prediction data with the current energy consumption observation data to obtain difference data, further includes: When the difference data conforms to the preset data range, determine that the energy consumption of the target building is in a normal state, and determine that the energy conservation diagnosis result is that the target building does not need energy conservation.

7. The method according to claim 1, wherein, generating a public building energy conservation strategy in the target area according to the energy conservation diagnosis result, including: According to the energy conservation diagnosis result, determine the public buildings in the target area that need energy conservation, and generate energy conservation summary information; According to the energy conservation summary information, determine the public building energy conservation strategy in the target area.

8. The method according to claim 1, wherein, performing anomaly detection on each group of initial sample data to determine the valid sample data in each group of initial sample data, including: Performing anomaly detection on each group of initial sample data according to a preset anomaly detection algorithm to determine each group of abnormal sample data; According to each group of initial sample data and each group of abnormal sample data, obtain each group of valid sample data.

9. A public building energy conservation diagnosis system, wherein, the public building energy conservation diagnosis system includes: A sample acquisition module, configured to acquire historical energy consumption observation data of public buildings in a target area, group the historical energy consumption observation data according to preset influencing factors to obtain initial sample data for each group. The preset influencing factors include building information, environmental information, and time information. The building information includes the type of building and the area of the building. The types of buildings include office buildings, commercial buildings, tourism buildings, science, education, culture, health, and communication buildings, and transportation buildings. Office buildings include office buildings and government department offices. Commercial buildings include shopping malls and financial buildings. Tourism buildings include hotels and entertainment venues. Science, education, culture, health, and communication buildings include cultural, educational, scientific, medical, health, and sports buildings. Communication buildings include post and telecommunications, communication, data centers, and broadcasting rooms. Transportation buildings include airports, high-speed railway stations, railway stations, subways, and bus stations. The environmental information includes meteorological conditions, temperature conditions, and humidity conditions. The time information includes date attributes and time ranges; A sample generation module, configured to perform anomaly detection on the initial sample data for each group to determine the valid sample data in the initial sample data for each group; The sample generation module is further configured to generate incremental sample data for each group according to the valid sample data for each group; An energy consumption prediction module, configured to establish an energy consumption prediction model according to the incremental sample data for each group and the valid sample data for each group. The energy consumption prediction model is applicable to different scenarios; The energy consumption prediction module is further configured to input the current building information, current environmental information, and current time information of a target building into the energy consumption prediction model to obtain energy consumption prediction data; An energy conservation diagnosis module, configured to acquire the current energy consumption observation data of the target building, perform energy conservation diagnosis according to the energy consumption prediction data and the current energy consumption observation data to obtain an energy conservation diagnosis result; The energy conservation diagnosis module is further configured to generate an energy conservation strategy for public buildings in the target area according to the energy conservation diagnosis result; The sample generation module is further configured to generate the sample distribution for each group according to the valid sample data for each group; Generate incremental sample data for each group according to the sample distribution for each group.

Images