Energy consumption anomaly monitoring method and device, electronic equipment and storage medium

Abstract

The application relates to an energy consumption anomaly monitoring method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring first data, which is the energy consumption data of a building within a first time. The first time is segmented to obtain a plurality of time periods, and each time period in the plurality of time periods has the same time interval. A first matrix is obtained based on the plurality of time periods, and the first matrix is an energy consumption data matrix arranged based on the plurality of time periods. The first matrix is supplemented to obtain a second matrix, and the second matrix is an energy consumption data matrix meeting the training requirements of a convolutional neural network. The second matrix is subjected to self-encoding training through the convolutional neural network, and the trained second matrix is clustered through a K-means algorithm to obtain a corresponding convolutional neural network model. The convolutional neural network model is called to classify newly input building energy consumption data to determine whether the newly input building energy consumption data is abnormal.

Description

Technical Field

[0001] This invention relates to the field of big data analytics, and in particular to a method, apparatus, electronic device, and storage medium for monitoring energy consumption anomalies. Background Technology

[0002] Buildings generate vast amounts of energy consumption data over long periods of operation. With advancements in computer technology and embedded systems, this data can be easily collected, ushering the construction industry into the era of big data. Therefore, identifying anomalous energy consumption patterns from this data has long been a focus for researchers and engineers.

[0003] Currently, two main approaches are used in the field of anomaly detection for energy consumption in office buildings: one is time-series clustering techniques driven by machine learning, such as K-means clustering algorithms; the other is anomaly identification methods based on general probabilistic statistical regression analysis. However, the operational patterns of office buildings vary significantly between holidays and weekdays, with weekdays accounting for a larger proportion. Furthermore, the proportions of transitional seasons, heating seasons, and cooling seasons differ, and office buildings exhibit variations in occupancy rates over the years. Therefore, these two approaches struggle to directly compare historical data, making it difficult to develop effective and reliable energy consumption models and thus hindering the effective addressing of imbalanced energy consumption data in office buildings.

[0004] Therefore, traditional methods of monitoring abnormal energy consumption data in office buildings are less effective in solving the problem of unbalanced energy consumption data in office buildings. Summary of the Invention

[0005] Therefore, it is necessary to provide an energy consumption anomaly monitoring method, device, electronic equipment, and storage medium that can effectively solve the problem of unbalanced energy consumption data in office buildings, addressing the aforementioned technical issues.

[0006] This invention provides a method for monitoring energy consumption anomalies, the method comprising:

[0007] Obtain the first data, which is the building's energy consumption data within a first time period;

[0008] By segmenting the first time period, multiple time periods are obtained, and each of the multiple time periods has the same time interval;

[0009] A first matrix is ​​obtained based on the multiple time periods, and the first matrix is ​​an energy consumption data matrix formed by arranging the first data based on the multiple time periods;

[0010] By padding the first matrix, a second matrix is ​​obtained, which is an energy consumption data matrix that meets the training requirements of a convolutional neural network.

[0011] The second matrix is ​​trained by autoencoding using the convolutional neural network, and the K-means algorithm is called to cluster the trained second matrix to obtain the corresponding convolutional neural network model.

[0012] The convolutional neural network model is invoked to classify the newly input building energy consumption data in order to determine whether the newly input building energy consumption data is abnormal.

[0013] In one embodiment, obtaining the second matrix by padding the first matrix includes:

[0014] Obtain the fill value, which is used to replace the missing values ​​in the first matrix;

[0015] The first matrix is ​​filled with the fill value to obtain the second matrix.

[0016] In one embodiment, the step of calling the convolutional neural network model to classify the newly input building energy consumption data to determine whether the newly input building energy consumption data is abnormal includes:

[0017] Obtain second data, which is the building energy consumption data input to the convolutional neural network model;

[0018] The convolutional neural network model is invoked to classify the second data in order to obtain the distance between the second data and the cluster center point.

[0019] In one embodiment, the step of calling the convolutional neural network model to classify the second data to obtain the distance between the second data and the cluster center point includes:

[0020] Based on the distance between the second data and the cluster center point, determine whether the second data is abnormal.

[0021] In one embodiment, determining whether the second data is abnormal based on the distance between the second data and the cluster center point includes:

[0022] Determine whether the distance between the second data and the cluster center point exceeds the first threshold; if so, then...

[0023] Output the result if the second data is abnormal energy consumption data; otherwise,

[0024] The output of the second data is the result of normal energy consumption data.

[0025] In one embodiment, the second data is the building's energy consumption data within a second time period, and the second time period has the same time interval as the first time period.

[0026] In one embodiment, the second matrix is ​​a square matrix that meets the training requirements of the convolutional neural network.

[0027] The present invention also provides an energy consumption anomaly monitoring device, the device comprising:

[0028] The first acquisition module is used to acquire the first data, which is the energy consumption data of the building within a first time period;

[0029] The time segmentation module is used to obtain multiple time periods by segmenting the first time period, wherein each time period has the same time interval.

[0030] The second acquisition module is used to acquire a first matrix based on the multiple time periods, wherein the first matrix is ​​an energy consumption data matrix formed by arranging the first data based on the multiple time periods;

[0031] The matrix completion module is used to complete the first matrix to obtain a second matrix, wherein the second matrix is ​​an energy consumption data matrix that meets the training requirements of a convolutional neural network.

[0032] The model training module is used to perform autoencode training on the second matrix through the convolutional neural network, and to call the K-means algorithm to cluster the trained second matrix to obtain the corresponding convolutional neural network model.

[0033] The anomaly detection module is used to call the convolutional neural network model to classify the newly input building energy consumption data in order to determine whether the newly input building energy consumption data is abnormal.

[0034] The present invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the energy consumption anomaly monitoring method as described above.

[0035] The present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, implements the energy consumption anomaly monitoring method as described above.

[0036] The aforementioned energy consumption anomaly monitoring method, device, electronic equipment, and storage medium acquire energy consumption data of office buildings within a specified time period. This data is then segmented into multiple time periods with identical intervals. Subsequently, an energy consumption data matrix is ​​obtained based on the distribution of the energy consumption data within these multiple time periods. This matrix is ​​then padded to obtain an energy consumption data matrix that meets the training requirements of a convolutional neural network. Finally, a convolutional neural network is used to autoencode and train the second matrix, and the K-means algorithm is called to cluster the trained energy consumption data to obtain the corresponding convolutional neural network model. This convolutional neural network is used to classify newly input building energy consumption data to determine whether the newly input building energy consumption data is abnormal. This method establishes a building energy consumption model through deep clustering analysis to discover abnormal energy consumption data. It adds a convolutional neural network autoencoder algorithm to the general clustering analysis, which improves the robustness of the energy consumption monitoring model and avoids the risk of model classification accuracy loss caused by various factors. It can better model and extract features for office building energy consumption models, realize the application of building energy conservation big data analysis technology in the real environment, and effectively solve the problem of unbalanced energy consumption data in office buildings. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0038] Figure 1 This is one of the flowcharts of the energy consumption anomaly monitoring method provided by the present invention;

[0039] Figure 2 This is the second schematic diagram of the energy consumption anomaly monitoring method provided by the present invention;

[0040] Figure 3 The third schematic diagram of the energy consumption anomaly monitoring method provided by the present invention;

[0041] Figure 4 The fourth flowchart illustrates the energy consumption anomaly monitoring method provided by this invention.

[0042] Figure 5 A flowchart illustrating the energy consumption anomaly monitoring method in a specific embodiment of the present invention;

[0043] Figure 6 An energy consumption data analysis effect diagram of the energy consumption anomaly monitoring method provided in a specific embodiment of the present invention;

[0044] Figure 7 This is a schematic diagram of the energy consumption anomaly monitoring device provided by the present invention;

[0045] Figure 8 An internal structural diagram of the computer device provided by the present invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] The following is combined with Figures 1-8 The present invention describes an energy consumption anomaly monitoring method, apparatus, electronic device, and storage medium.

[0048] like Figure 1 As shown, in one embodiment, an energy consumption anomaly monitoring method includes the following steps:

[0049] Step S110: Obtain the first data, which is the energy consumption data of the building within the first time period.

[0050] Specifically, the server obtains the building's energy consumption data within a specified time period, such as the energy consumption data of an office building for 24 hours a day.

[0051] Step S120: By segmenting the first time period, multiple time periods are obtained, and each time period has the same time interval.

[0052] Specifically, the server segmentes the first time in step S110 to obtain multiple time periods with the same time interval, such as dividing a 24-hour day into four time periods: 0-6, 7-12, 13-18, and 18-13.

[0053] Step S130: Obtain a first matrix based on multiple time periods. The first matrix is ​​an energy consumption data matrix composed of the first data arranged based on multiple time periods.

[0054] Specifically, the server obtains building energy consumption data based on multiple time periods obtained in step S120, and then arranges these multiple time periods to form an energy consumption data matrix, i.e., the first matrix. For example, a 4×6 matrix is ​​formed by building energy consumption data from four time periods: 0-6, 7-12, 13-18, and 18-13.

[0055] Step S140: By padding the first matrix, a second matrix is ​​obtained. The second matrix is ​​an energy consumption data matrix that meets the training requirements of a convolutional neural network.

[0056] Specifically, the server-side completes the energy consumption data matrix obtained in step S130 to obtain an energy consumption data matrix that meets the training requirements of the convolutional neural network, namely the second matrix. For example, six zeros are added to the first and last rows of the 4×6 matrix constructed in step S130 to complete the matrix, so that the building energy consumption data matrix for each day is a 6×6 square matrix.

[0057] Step S150: The second matrix is ​​trained by an autoencoder using a convolutional neural network, and the K-means algorithm is called to cluster the trained second matrix to obtain the corresponding convolutional neural network model.

[0058] Specifically, the server uses a convolutional neural network to autoencode the second matrix obtained after matrix completion in step S140, and calls the K-means algorithm to cluster the trained second matrix to obtain the corresponding convolutional neural network model.

[0059] Step S160: Call the convolutional neural network model to classify the newly input building energy consumption data in order to determine whether the newly input building energy consumption data is abnormal.

[0060] Specifically, the server calls the convolutional neural network model trained in step S150 to classify the newly input building energy consumption data to determine whether the newly input building energy consumption data is abnormal energy consumption data.

[0061] The aforementioned energy consumption anomaly monitoring method acquires energy consumption data of office buildings within a specified time period and segments this data into multiple time intervals with the same time interval. Subsequently, based on the distribution of office building energy consumption data within these multiple time intervals, a corresponding energy consumption data matrix is ​​obtained. This matrix is ​​then augmented to obtain an energy consumption data matrix that meets the training requirements of a convolutional neural network (CNN). Finally, a second matrix is ​​trained using a CNN with an autoencoder, and the K-means algorithm is used to cluster the trained energy consumption data to obtain a corresponding CNN model. This CNN is used to classify newly input building energy consumption data to determine whether the new input data is abnormal. This method establishes a building energy consumption model through deep clustering analysis to discover abnormal energy consumption data. By adding a CNN autoencoder algorithm to general clustering analysis, it improves the robustness of the energy consumption monitoring model, avoids the risk of model classification accuracy loss due to various factors, and better extracts features for building energy consumption modeling. This realizes the application of building energy conservation big data analysis technology in a real-world environment and effectively solves the problem of unbalanced energy consumption data in office buildings.

[0062] like Figure 2 As shown, in one embodiment, the energy consumption anomaly monitoring method provided by the present invention obtains a second matrix by completing a first matrix, including the following steps:

[0063] Step S142: Obtain the fill value, which is used to replace the missing values ​​in the first matrix.

[0064] Specifically, during the process of completing the energy consumption data matrix, the server first obtains the fill value, such as the value 0, to replace the missing value in the first matrix.

[0065] Step S144: Fill the first matrix with fill values ​​to obtain the second matrix.

[0066] Specifically, the server fills the energy consumption data matrix with the fill values ​​obtained in step S142 to obtain the second matrix. For example, the first and last rows of the energy consumption data matrix for the four time periods of 0-6, 7-12, 13-18, and 18-13 corresponding to 24 hours of a day are filled with 6 zero values ​​to complete the energy consumption data matrix and obtain a 6×6 square matrix.

[0067] like Figure 3 As shown, in one embodiment, the energy consumption anomaly monitoring method provided by the present invention calls a convolutional neural network model to classify newly input building energy consumption data in order to determine whether the newly input building energy consumption data is abnormal, including the following steps:

[0068] Step S162: Obtain the second data, which is the building energy consumption data input to the convolutional neural network model.

[0069] Specifically, the server obtains new building energy consumption data, i.e., the second data, which is used as input to the trained convolutional neural network model.

[0070] Step S164: Call the convolutional neural network model to classify the second data to obtain the distance between the second data and the cluster center point.

[0071] Specifically, the server calls the trained convolutional neural network model to classify the new input building energy consumption data obtained in step S162, and calculates the distance between the new input building energy consumption data and the cluster center point in the model.

[0072] Step S166: Based on the distance between the second data and the cluster center point, determine whether the second data is abnormal.

[0073] Specifically, the server determines whether the newly input building energy consumption data is abnormal based on the distance between the newly input building energy consumption data and the cluster center point in the model.

[0074] like Figure 4 As shown, in one embodiment, the energy consumption anomaly monitoring method provided by the present invention determines whether the second data is abnormal based on the distance between the second data and the cluster center point, including:

[0075] Step S410: Determine whether the distance between the second data and the cluster center point exceeds the first threshold.

[0076] Specifically, in the process of determining whether newly input building energy consumption data is abnormal, the server will determine whether the distance between the newly input building energy consumption data and the cluster center point in the model exceeds the set first threshold.

[0077] Step S420: Output the second data as the result of abnormal energy consumption data.

[0078] Specifically, if the judgment result in step S410 is that the distance between the newly input building energy consumption data and the cluster center point in the model exceeds the set first threshold, the server will output the result that the newly input building energy consumption data is abnormal energy consumption data.

[0079] Step S430: Output the second data as the result of normal energy consumption data.

[0080] Specifically, if the judgment result in step S410 is that the distance between the newly input building energy consumption data and the cluster center point in the model does not exceed the set first threshold, the server will output the result that the newly input building energy consumption data is normal energy consumption data.

[0081] See Figure 5 and Figure 6 As shown in the specific embodiment, this invention provides a method for monitoring energy consumption anomalies. First, energy consumption data of office buildings is acquired and preprocessed. Using a 24-hour day as the basic data sample, the day is divided into four time periods: 0-6, 7-12, 13-18, and 18-13, forming a 4×6 matrix. Six zeros are added to the first and last rows of this matrix to complete it, resulting in a 6×6 square matrix for each day's energy consumption data. Subsequently, a convolutional neural network is used to autoencode and train this 6×6 square matrix. The K-means algorithm is then used to cluster the trained energy consumption data matrix, ultimately obtaining the corresponding convolutional neural network model, i.e., the energy consumption monitoring model.

[0082] In this embodiment, after the convolutional neural network model is trained, the resulting energy consumption monitoring model needs to be evaluated. If the evaluation is successful, the energy consumption monitoring model is saved. When new building energy consumption data is input into the system, the saved energy consumption monitoring model is called to classify the new data and calculate the distance between the new data and the cluster center point in the model. This determines whether the distance exceeds a set threshold. If the distance exceeds the threshold, the new data is considered abnormal. If the distance does not exceed the threshold, the data is considered normal.

[0083] The aforementioned energy consumption anomaly monitoring method incorporates a convolutional neural network autoencoder algorithm on top of general clustering analysis. This enables better feature extraction for office building energy consumption modeling, including start-up time, peak energy consumption, and energy fluctuations. The autoencoder algorithm improves the robustness of the energy consumption monitoring model, effectively avoiding the loss of model classification accuracy caused by factors such as occupancy rates. Furthermore, by integrating deep clustering analysis with technologies related to office building energy consumption pattern recognition, office building energy consumption prediction, and building energy consumption anomaly data detection, a building energy consumption model is established to identify anomaly characteristics. BEMS can also record the time points of specific problems and provide corresponding management suggestions for different energy consumption anomalies, thus realizing the application and implementation of building energy conservation big data analysis technology in real-world environments. Secondly, this method can be embedded into building energy management platforms for large-scale application. Continuous and in-depth development towards building energy consumption prediction and demand-side management will fully explore the value of clustering analysis in optimizing the operation of building energy systems, providing more important support for achieving building energy conservation and carbon reduction.

[0084] The energy consumption anomaly monitoring device provided by the present invention is described below. The energy consumption anomaly monitoring device described below can be referred to in correspondence with the energy consumption anomaly monitoring method described above.

[0085] like Figure 7 As shown, in one embodiment, an energy consumption anomaly monitoring device includes a first acquisition module 710, a time segmentation module 720, a second acquisition module 730, a matrix completion module 740, a model training module 750, and an anomaly monitoring module 760.

[0086] The first acquisition module 710 is used to acquire the first data, which is the energy consumption data of the building within a first time period.

[0087] The time segmentation module 720 is used to segment the first time to obtain multiple time periods, each of which has the same time interval.

[0088] The second acquisition module 730 is used to acquire a first matrix based on multiple time periods. The first matrix is ​​an energy consumption data matrix composed of the first data arranged based on multiple time periods.

[0089] The matrix completion module 740 is used to complete the first matrix to obtain the second matrix, which is an energy consumption data matrix that meets the training requirements of the convolutional neural network.

[0090] The model training module 750 is used to autoencode the second matrix using a convolutional neural network and call the K-means algorithm to cluster the trained first matrix to obtain the corresponding convolutional neural network model.

[0091] The anomaly detection module 760 is used to call a convolutional neural network model to classify newly input building energy consumption data in order to determine whether the newly input building energy consumption data is abnormal.

[0092] In this embodiment, the matrix completion module of the energy consumption anomaly monitoring device provided by the present invention is specifically used for:

[0093] Get the filler values, which are used to replace the missing values ​​in the first matrix.

[0094] The first matrix is ​​filled with fill values ​​to obtain the second matrix.

[0095] In this embodiment, the energy consumption anomaly monitoring device provided by the present invention has an anomaly monitoring module specifically used for:

[0096] Obtain the second data, which is the building energy consumption data input into the convolutional neural network model.

[0097] The convolutional neural network model is invoked to classify the second data to obtain the distance between the second data and the cluster center point.

[0098] Based on the distance between the second data and the cluster center point, determine whether the second data is abnormal.

[0099] In this embodiment, the energy consumption anomaly monitoring device provided by the present invention further includes a judgment module, used for:

[0100] Determine if the distance between the second data point and the cluster center exceeds the first threshold. If so, then...

[0101] The second data point is determined to be abnormal energy consumption data. If not, then...

[0102] The second data point was determined to be normal energy consumption data.

[0103] Figure 8 This example illustrates a schematic diagram of the physical structure of an electronic device, which can be a smart terminal. Its internal structure diagram can be as follows: Figure 8 As shown. The electronic device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for monitoring abnormal energy consumption, which includes:

[0104] Obtain the first data, which is the building's energy consumption data within the first time period;

[0105] By segmenting the first time point, multiple time periods are obtained, and each time period has the same time interval.

[0106] The first matrix is ​​obtained based on multiple time periods. The first matrix is ​​an energy consumption data matrix composed of the first data arranged based on multiple time periods.

[0107] By padding the first matrix, a second matrix is ​​obtained. The second matrix is ​​an energy consumption data matrix that meets the training requirements of a convolutional neural network.

[0108] The second matrix is ​​trained by an autoencoder using a convolutional neural network, and the first matrix after training is clustered using the K-means algorithm to obtain the corresponding convolutional neural network model.

[0109] A convolutional neural network model is invoked to classify the newly input building energy consumption data in order to determine whether the newly input building energy consumption data is abnormal.

[0110] Those skilled in the art will understand that Figure 8 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the electronic device to which the present invention is applied. A specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0111] On the other hand, the present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, implements a method for monitoring abnormal power consumption, the method comprising:

[0112] Obtain the first data, which is the building's energy consumption data within the first time period;

[0113] By segmenting the first time point, multiple time periods are obtained, and each time period has the same time interval.

[0114] The first matrix is ​​obtained based on multiple time periods. The first matrix is ​​an energy consumption data matrix composed of the first data arranged based on multiple time periods.

[0115] By padding the first matrix, a second matrix is ​​obtained. The second matrix is ​​an energy consumption data matrix that meets the training requirements of a convolutional neural network.

[0116] The second matrix is ​​trained by an autoencoder using a convolutional neural network, and the first matrix after training is clustered using the K-means algorithm to obtain the corresponding convolutional neural network model.

[0117] A convolutional neural network model is invoked to classify the newly input building energy consumption data in order to determine whether the newly input building energy consumption data is abnormal.

[0118] In another aspect, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and when the processor executes the computer instructions, it implements a method for monitoring abnormal power consumption, the method comprising:

[0119] Obtain the first data, which is the building's energy consumption data within the first time period;

[0120] By segmenting the first time point, multiple time periods are obtained, and each time period has the same time interval.

[0121] The first matrix is ​​obtained based on multiple time periods. The first matrix is ​​an energy consumption data matrix composed of the first data arranged based on multiple time periods.

[0122] By padding the first matrix, a second matrix is ​​obtained. The second matrix is ​​an energy consumption data matrix that meets the training requirements of a convolutional neural network.

[0123] The second matrix is ​​trained by an autoencoder using a convolutional neural network, and the first matrix after training is clustered using the K-means algorithm to obtain the corresponding convolutional neural network model.

[0124] A convolutional neural network model is invoked to classify the newly input building energy consumption data in order to determine whether the newly input building energy consumption data is abnormal.

[0125] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory.

[0126] By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0127] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0128] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for monitoring energy consumption anomalies, characterized in that, The method includes: Obtain the first data, which is the building's energy consumption data within a first time period; By segmenting the first time period, multiple time periods are obtained, and each of the multiple time periods has the same time interval; A first matrix is ​​obtained based on the multiple time periods, and the first matrix is ​​an energy consumption data matrix formed by arranging the first data based on the multiple time periods; By padding the first matrix, a second matrix is ​​obtained, which is an energy consumption data matrix that meets the training requirements of a convolutional neural network. The second matrix is ​​trained by autoencoding using the convolutional neural network, and the K-means algorithm is called to cluster the trained second matrix to obtain the corresponding convolutional neural network model. The convolutional neural network model is invoked to classify the newly input building energy consumption data in order to determine whether the newly input building energy consumption data is abnormal. The step of obtaining the second matrix by completing the first matrix includes: Obtain the fill value, which is used to replace the missing values ​​in the first matrix; The first matrix is ​​filled with the fill value to obtain the second matrix; The step of calling the convolutional neural network model to classify the newly input building energy consumption data to determine whether the newly input building energy consumption data is abnormal includes: Obtain second data, which is the building energy consumption data input to the convolutional neural network model; The convolutional neural network model is invoked to classify the second data in order to obtain the distance between the second data and the cluster center point.

2. The energy consumption anomaly monitoring method according to claim 1, characterized in that, The step of calling the convolutional neural network model to classify the second data to obtain the distance between the second data and the cluster center point includes: Based on the distance between the second data and the cluster center point, determine whether the second data is abnormal.

3. The energy consumption anomaly monitoring method according to claim 2, characterized in that, The step of determining whether the second data is abnormal based on the distance between the second data and the cluster center point includes: Determine whether the distance between the second data and the cluster center point exceeds the first threshold; if so, then... Output the result if the second data is abnormal energy consumption data; otherwise, The output of the second data is the result of normal energy consumption data.

4. The energy consumption anomaly monitoring method according to any one of claims 1 to 3, characterized in that, The second data is the energy consumption data of the building during the second time period, and the second time period has the same time interval as the first time period.

5. The energy consumption anomaly monitoring method according to claim 1, characterized in that, The second matrix is ​​a square matrix that meets the training requirements of the convolutional neural network.

6. An energy consumption anomaly monitoring device, characterized in that, The device includes: The first acquisition module is used to acquire the first data, which is the energy consumption data of the building within a first time period; The time segmentation module is used to obtain multiple time periods by segmenting the first time period, wherein each time period has the same time interval. The second acquisition module is used to acquire a first matrix based on the multiple time periods, wherein the first matrix is ​​an energy consumption data matrix formed by arranging the first data based on the multiple time periods; The matrix completion module is used to complete the first matrix to obtain a second matrix, wherein the second matrix is ​​an energy consumption data matrix that meets the training requirements of a convolutional neural network. The model training module is used to perform autoencode training on the second matrix through the convolutional neural network, and to call the K-means algorithm to cluster the trained second matrix to obtain the corresponding convolutional neural network model. An anomaly detection module is used to call the convolutional neural network model to classify the newly input building energy consumption data in order to determine whether the newly input building energy consumption data is abnormal. The step of obtaining the second matrix by completing the first matrix includes: Obtain the fill value, which is used to replace the missing values ​​in the first matrix; The first matrix is ​​filled with the fill value to obtain the second matrix; The step of calling the convolutional neural network model to classify the newly input building energy consumption data to determine whether the newly input building energy consumption data is abnormal includes: Obtain second data, which is the building energy consumption data input to the convolutional neural network model; The convolutional neural network model is invoked to classify the second data in order to obtain the distance between the second data and the cluster center point.

7. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

8. A computer storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

Images