Building energy consumption prediction method, system, equipment and medium

Abstract

The invention provides a building energy consumption prediction method, system and device and a medium, and the method comprises the steps: carrying out the feature mapping association of the overall energy consumption of a target building and the local energy consumption of each functional region, obtaining the association identification features of the energy consumption between the overall energy consumption and each functional region, building an energy consumption prediction model based on all the association identification features, and carrying out the prediction of the energy consumption. Determining a suppression gain coefficient of a feature channel corresponding to each functional region in the energy consumption prediction model; performing fuzzy evaluation on the energy consumption fluctuation level of the energy consumption abnormal time period according to the load fluctuation characteristics of each functional area to obtain a fuzzy evaluation value of the energy consumption fluctuation level; and when the fuzzy evaluation value exceeds a preset energy consumption fluctuation threshold value, performing cost constraint on energy consumption feature attention of each functional region in the energy consumption prediction model according to all suppression gain coefficients, and predicting the building energy consumption based on the energy consumption prediction model after the cost constraint. By adopting the scheme of the invention, the cost constraint of different feature attention in the energy consumption prediction model can be realized.

Description

Technical Field

[0001] The present application relates to the technical field of energy consumption prediction, and more specifically, to a building energy consumption prediction method, system, device, and medium. Background Art

[0002] In building energy management, building energy consumption detection, as a basic link in intelligent energy regulation, is of great significance for achieving building energy conservation, load response regulation, and energy utilization optimization. Traditional energy consumption monitoring methods mainly focus on the collection and presentation of total energy consumption data, lacking the refined identification and dynamic response capabilities for the energy consumption characteristics of each functional area, and it is difficult to support the early identification of abnormal energy consumption behaviors and targeted regulation. With the complexity of building structures and functional layouts, refined energy consumption perception and high-timeliness prediction capabilities have gradually become the core requirements of building energy management systems.

[0003] Existing mainstream building energy consumption prediction methods generally have static problems in model structure design, making it difficult to adapt to the complex and changeable energy consumption characteristics of internal building functional areas. Especially in scenarios with frequent load fluctuations or frequent local energy consumption anomalies, it is difficult to effectively extract the non-linear correlation between overall energy consumption data and local energy consumption behaviors, resulting in the lack of response ability of the model to regional abnormal energy consumption. The fundamental reason for this defect is that traditional methods fail to introduce a feature channel mechanism that can reflect the dynamic correlation between overall and local energy consumption, lack means to suppress the energy consumption characteristics of irrelevant areas and strategies to strengthen the characteristics of key areas, thereby restricting the model's analytical ability and prediction stability for abnormal energy consumption behaviors. Therefore, how to achieve cost constraints on different feature attentions in the energy consumption prediction model to improve the robustness of building energy consumption prediction has become a difficult problem faced by the industry. Summary of the Invention

[0004] The present application provides a building energy consumption prediction method, system, device, and medium, which can achieve cost constraints on different feature attentions in the energy consumption prediction model.

[0005] In a first aspect, the present application provides a building energy consumption prediction method. First, low-frequency energy consumption monitoring devices are deployed outside the target building in advance, and high-frequency energy consumption monitoring devices are deployed in each functional area inside the target building. The building energy consumption prediction method includes the following steps: In the current energy consumption monitoring period, collect the overall energy consumption data of the target building through the low-frequency energy consumption monitoring device, and collect the local energy consumption data of each functional area of the target building through the high-frequency energy consumption monitoring device; Perform characteristic mapping correlation on the overall energy consumption of the target building and the local energy consumption of each functional area according to the overall energy consumption data and all local energy consumption data, obtain the correlation identification characteristics of the energy consumption between the whole and each functional area, and then construct an energy consumption prediction model based on all the correlation identification characteristics, and determine the suppression gain coefficient of the corresponding characteristic channels of each functional area in the energy consumption prediction model; Predict the energy consumption abnormal period in the next energy consumption monitoring period based on the historical energy consumption records of the target building, and then perform a fuzzy evaluation on the energy consumption fluctuation level of the energy consumption abnormal period according to the load fluctuation characteristics of each functional area to obtain the fuzzy evaluation value of the energy consumption fluctuation level in the energy consumption abnormal period; In the process of predicting the energy consumption in the next energy consumption monitoring period, when the fuzzy evaluation value exceeds the preset energy consumption fluctuation threshold, impose a cost constraint on the energy consumption characteristic attention of each functional area in the energy consumption prediction model according to all the suppression gain coefficients, and then predict the building energy consumption in the next energy consumption monitoring period based on the energy consumption prediction model after the cost constraint.

[0006] Preferably, performing characteristic mapping correlation on the overall energy consumption of the target building and the local energy consumption of each functional area according to the overall energy consumption data and all local energy consumption data to obtain the correlation identification characteristics of the energy consumption between the whole and each functional area specifically includes: Align the time of the overall energy consumption data and the local energy consumption data of each functional area, and extract the correlation mapping characteristics of the energy consumption between the whole and each functional area within the same time window; Determine the dynamic correlation coefficient between the local energy consumption data of each functional area and the overall energy consumption data; Determine the correlation identification characteristics of the energy consumption between the whole and each functional area through all the dynamic correlation coefficients and all the correlation mapping characteristics.

[0007] Preferably, constructing an energy consumption prediction model based on all the correlation identification characteristics specifically includes: Construct a prediction model including multiple characteristic input channels, where each channel corresponds to a functional area; Use the correlation identification characteristics corresponding to each functional area as the input parameters of the corresponding channels respectively, and set the initial parameters of the prediction model; Train the prediction model by using a supervised training method combined with the historical energy consumption data of the target building, and iteratively optimize the initial parameters to train and obtain an energy consumption prediction model.

[0008] Preferably, determining the suppression gain coefficient of the corresponding characteristic channels of each functional area in the energy consumption prediction model specifically includes: Extract the deviation distribution characteristics between the energy consumption change rate of each functional area in the historical energy consumption monitoring period and the overall energy consumption change trend; Determine the response intensity of the feature channels corresponding to each functional area in the energy consumption prediction model based on the deviation distribution characteristics; Set the initial suppression factors of the feature channels corresponding to each functional area according to all the response intensities; Dynamically adjust the initial suppression factors of each feature channel through the error feedback during the training process of the energy consumption prediction model to obtain the suppression gain coefficients of the feature channels corresponding to each functional area in the energy consumption prediction model.

[0009] Preferably, predicting the energy consumption abnormal period in the next energy consumption monitoring cycle based on the historical energy consumption records of the target building specifically includes: Obtain the historical overall energy consumption data from the historical energy consumption records of the target building; Extract the periodic characteristics of the historical overall energy consumption data, identify the time periods with energy consumption mutations and frequent anomalies, and predict the energy consumption abnormal period in the next energy consumption monitoring cycle through the identified time periods.

[0010] Preferably, performing a fuzzy evaluation on the energy consumption fluctuation level of the energy consumption abnormal period according to the load fluctuation characteristics of each functional area to obtain the fuzzy evaluation value of the energy consumption fluctuation level in the energy consumption abnormal period specifically includes: Construct a fuzzy membership function describing energy consumption fluctuations based on the load fluctuation characteristics of each functional area in the historical period, and then determine the fuzzy membership degrees of each functional area in different fluctuation intervals; Perform fuzzy reasoning on the energy consumption fluctuation levels of each functional area in the energy consumption abnormal period according to all the fuzzy membership degrees to obtain the energy consumption fluctuation credibility of each functional area at different fluctuation levels; Determine the fuzzy evaluation value of the energy consumption fluctuation level in the energy consumption abnormal period through all the energy consumption fluctuation credibilities.

[0011] Preferably, the low-frequency energy consumption monitoring device is specifically a multi-functional power meter.

[0012] In a second aspect, the present application provides a building energy consumption prediction system, including: An acquisition module, configured to collect the overall energy consumption data of the target building through a low-frequency energy consumption monitoring device and collect the local energy consumption data of each functional area of the target building through a high-frequency energy consumption monitoring device in the current energy consumption monitoring cycle; A processing module, configured to perform feature mapping association on the overall energy consumption of the target building and the local energy consumption of each functional area according to the overall energy consumption data and all the local energy consumption data to obtain the associated identification features of the energy consumption between the whole and each functional area, and then construct an energy consumption prediction model based on all the associated identification features and determine the suppression gain coefficients of the feature channels corresponding to each functional area in the energy consumption prediction model; The processing module is further configured to predict the energy consumption abnormal period in the next energy consumption monitoring period based on the historical energy consumption records of the target building, and then perform a fuzzy evaluation on the energy consumption fluctuation level of the energy consumption abnormal period according to the load fluctuation characteristics of each functional area to obtain the fuzzy evaluation value of the energy consumption fluctuation level within the energy consumption abnormal period; The execution module is configured to, during the energy consumption prediction process of the next energy consumption monitoring period, when the fuzzy evaluation value exceeds the preset energy consumption fluctuation threshold, perform cost constraint on the energy consumption feature attention of each functional area in the energy consumption prediction model according to all the suppression gain coefficients, and then predict the building energy consumption in the next energy consumption monitoring period based on the energy consumption prediction model after cost constraint.

[0013] In a third aspect, the present application provides a computer device, which includes a memory and a processor. The memory stores code, and the processor is configured to obtain the code and execute the above-mentioned building energy consumption prediction method.

[0014] In a fourth aspect, the present application provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the above-mentioned building energy consumption prediction method is implemented.

[0015] The technical solutions provided by the disclosed embodiments of the present application have the following beneficial effects: In the embodiment of the present application, during the current energy consumption monitoring period, the overall energy consumption data of the target building is collected by the low-frequency energy consumption monitoring device, and the local energy consumption data of each functional area of the target building is collected by the high-frequency energy consumption monitoring device; the overall energy consumption and the local energy consumption of each functional area of the target building are subjected to feature mapping association according to the overall energy consumption data and all local energy consumption data to obtain the associated identification features of the energy consumption between the whole and each functional area, and then an energy consumption prediction model is constructed based on all the associated identification features, and the suppression gain coefficients of the corresponding feature channels of each functional area in the energy consumption prediction model are determined; the energy consumption abnormal period in the next energy consumption monitoring period is predicted based on the historical energy consumption records of the target building, and then the energy consumption fluctuation level of the energy consumption abnormal period is subjected to fuzzy evaluation according to the load fluctuation characteristics of each functional area to obtain the fuzzy evaluation value of the energy consumption fluctuation level within the energy consumption abnormal period; during the energy consumption prediction process of the next energy consumption monitoring period, when the fuzzy evaluation value exceeds the preset energy consumption fluctuation threshold, perform cost constraint on the energy consumption feature attention of each functional area in the energy consumption prediction model according to all the suppression gain coefficients, and then predict the building energy consumption in the next energy consumption monitoring period based on the energy consumption prediction model after cost constraint.

[0016] It can be seen that in this application, the cost constraints are imposed on the energy consumption feature attention of each functional area in the energy consumption prediction model by all the suppression gain coefficients, and the building energy consumption is predicted based on the energy consumption prediction model after the cost constraints. First, by performing feature mapping and correlation on the overall energy consumption of the target building and the local energy consumption of each functional area, it is possible to reveal the response degree and behavioral characteristics of different functional areas to the overall energy consumption change on the premise of time consistency, thereby establishing an energy consumption feature expression mechanism with regional discrimination ability. This feature mapping process realizes the high-dimensional correlation expression of energy consumption data by extracting the coupling features between the overall and local energy consumption, and constructs an energy consumption prediction model on this basis, enhancing the model's perception ability of the energy consumption differences between regions from the source. Second, by determining the suppression gain coefficients of the corresponding feature channels of each functional area in the energy consumption prediction model, the expression intensity of the corresponding feature channels of each functional area can be dynamically adjusted, effectively suppressing the interference of abnormal area features on the overall prediction result, highlighting the dominant energy consumption features of the key areas, realizing the controllable adjustment of feature attention, and providing a clearer semantic basis for the prediction of abnormal periods. Then, based on the historical energy consumption data, the abnormal energy consumption periods in the next monitoring cycle are predicted, and combined with the load fluctuation characteristics of the functional areas, a fuzzy evaluation method is used to quantitatively fuse and evaluate the energy consumption fluctuation level, so as to output a fuzzy evaluation value with the ability to accommodate uncertainty, significantly improving the adaptability of the energy consumption prediction model to complex fluctuation behaviors. Finally, when the fuzzy evaluation value exceeds the set threshold, the cost constraint mechanism activates the readjustment process of the feature attention in the energy consumption prediction model, so that the energy consumption prediction model further strengthens the energy consumption feature expression of the key areas in the abnormal prediction task, and minimizes the misleading effect of local noise features on the judgment result of the energy consumption prediction model. This mechanism constructs an adaptive adjustment path of the energy consumption prediction model in the energy consumption fluctuation scenario through a dynamic cost constraint method, effectively overcoming the problem of insufficient robustness caused by the fixed feature expression mode of traditional prediction methods. In summary, the solution of this application can realize the cost constraints of different feature attentions in the energy consumption prediction model, thereby improving the robustness of building energy consumption prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is an exemplary flowchart of a building energy consumption prediction method according to some embodiments of the present application; Figure 2 is a schematic diagram of an application scenario of a building energy consumption prediction system according to some embodiments of the present application; Figure 3 is a schematic flowchart of determining a fuzzy evaluation value according to some embodiments of the present application; Figure 4 is a schematic diagram of the structure of a building energy consumption prediction system according to some embodiments of the present application; Figure 5It is a schematic structural diagram of a computer device for implementing a building energy consumption prediction method according to some embodiments of the present application. Detailed implementation manners

[0018] To better understand the technical solutions of the present application, the technical solutions of the present application will be described in detail below in conjunction with the accompanying drawings of the specification and specific implementation manners.

[0019] Refer to Figure 1 , this figure is an exemplary flowchart of a building energy consumption prediction method according to some embodiments of the present application. The building energy consumption prediction method 100 mainly includes the following steps: In step 101, within the current energy consumption monitoring period, the overall energy consumption data of the target building is collected through low-frequency energy consumption monitoring devices, and the local energy consumption data of each functional area of the target building is collected through high-frequency energy consumption monitoring devices.

[0020] It should be noted that the low-frequency energy consumption monitoring devices can be pre-deployed outside the target building and the high-frequency energy consumption monitoring devices can be deployed in each functional area inside the target building in the present application, which can be implemented in the following manner: that is, a low-frequency energy consumption monitoring device can be installed at the main building inlet or the position of the main power distribution meter to record the overall energy consumption situation of the building. The sampling frequency of this type of device is generally set to once every 5 to 15 minutes to meet the needs of overall energy consumption trend monitoring; high-frequency energy consumption monitoring devices can be installed in each functional area inside the building (such as the office area, computer room, kitchen) respectively. This type of device is installed in the area branch circuit or local power distribution unit, and the sampling frequency is usually once every 1 to 5 minutes to capture the rapid changes in regional energy consumption.

[0021] In specific implementation, the overall energy consumption data of the target building can be collected through low-frequency energy consumption monitoring devices in the following manner: that is, the energy consumption data of the whole building can be collected through a multi-functional power meter, and its sampling frequency is set to once every 5 to 15 minutes, and it is connected to the background data platform or edge computing gateway through a standard communication protocol (such as Modbus) to collect and upload data regularly; the local energy consumption data of each functional area of the target building can be collected through high-frequency energy consumption monitoring devices in the following manner: that is, the energy consumption data of each functional area of the building can be collected through a multi-functional power meter, and its sampling frequency is set to once every 1 to 5 minutes, and it is connected to the background data platform or edge computing gateway through a standard communication protocol (such as Modbus) to collect and upload data regularly; it should be noted that the energy consumption data specifically includes voltage, current, power, and total electricity consumption.

[0022] In some embodiments, refer to Figure 2As shown in the figure, it is a schematic diagram of the application scenario of the building energy consumption prediction system in some embodiments of the present application. In the monitoring scenario 110, multiple monitoring devices 120 are installed. The energy consumption data of the building is monitored by the monitoring devices, and then the collected energy consumption data is transmitted to the data processor 130. The data processor analyzes and processes the energy consumption data, and predicts the building energy consumption. Finally, the prediction result is transmitted to the data storage 140, and the prediction result is stored by the data storage.

[0023] In step 102, based on the overall energy consumption data and all local energy consumption data, feature mapping association is performed on the overall energy consumption of the target building and the local energy consumption of each functional area, to obtain the associated identification features of the energy consumption between the whole and each functional area. Furthermore, an energy consumption prediction model is constructed based on all the associated identification features, and the suppression gain coefficients of the corresponding feature channels of each functional area in the energy consumption prediction model are determined.

[0024] In some embodiments, the feature mapping association of the overall energy consumption of the target building and the local energy consumption of each functional area based on the overall energy consumption data and all local energy consumption data to obtain the associated identification features of the energy consumption between the whole and each functional area can be implemented by the following steps: Perform time alignment on the overall energy consumption data and the local energy consumption data of each functional area, and extract the associated mapping features of the energy consumption between the whole and each functional area within the same time window; Determine the dynamic correlation coefficient between the local energy consumption data of each functional area and the overall energy consumption data; Determine the associated identification features of the energy consumption between the whole and each functional area through all the dynamic correlation coefficients and all the associated mapping features.

[0025] It should be noted that the associated mapping features in the present application are features that measure the temporal synchronization between the overall energy consumption and the local energy consumption of each functional area; the dynamic correlation coefficient in the present application refers to a statistic that measures the local linear correlation intensity generated by two time series changing with time within a sliding time window; the associated identification features in the present application are features that characterize the associated response intensity of each functional area in the overall energy consumption structure.

[0026] In specific implementation, first, time alignment is performed on the overall energy consumption data and the local energy consumption data of each functional area. The extraction of the correlation mapping features of the energy consumption between the overall and each functional area within the same time window can be achieved in the following manner, i.e., time alignment processing can be performed on the overall energy consumption data and the local energy consumption data of each functional area, that is, both are resampled and interpolated to complete according to a unified sampling period (such as every 5 minutes), an energy consumption data matrix under a unified time stamp is constructed, and then based on each time window (such as the sliding window length is 60 minutes), a mapping feature pair between the overall energy consumption and the local energy consumption of each functional area is extracted. For example, statistical features such as normalized difference, sliding mean difference, and energy consumption ratio can be used to form a feature vector, and this feature vector is used as the correlation mapping feature; second, the determination of the dynamic correlation coefficient between the local energy consumption data of each functional area and the overall energy consumption data can be achieved in the following manner, i.e., for each functional area, the Pearson correlation coefficient of the sliding window is used to analyze the dynamic correlation between the overall energy consumption and the energy consumption of the functional area over time, and the analysis result is used as the dynamic correlation coefficient between the local energy consumption data of the functional area and the overall energy consumption data, and then the dynamic correlation coefficient between the local energy consumption data of each functional area and the overall energy consumption data is obtained; then, the determination of the correlation identification feature of the energy consumption between the overall and each functional area through all the dynamic correlation coefficients and all the correlation mapping features can be achieved in the following manner, i.e., for each functional area, a vector composed of the dynamic correlation coefficient between the local energy consumption data of the functional area and the overall energy consumption data and the correlation mapping feature of the energy consumption between the overall and the functional area within the same time window is used as the correlation identification feature of the energy consumption between the overall and the functional area, and then the correlation identification feature of the energy consumption between the overall and each functional area is obtained.

[0027] In some embodiments, the construction of an energy consumption prediction model based on all the correlation identification features can be achieved through the following steps: Construct a prediction model including multiple feature input channels, where each channel corresponds to a functional area; Respectively use the correlation identification features corresponding to each functional area as the input parameters of the corresponding channels, and set the initial parameters of the prediction model; Use a supervised training method to train the prediction model in combination with the historical energy consumption data of the target building, and iteratively optimize the initial parameters, and then train to obtain an energy consumption prediction model.

[0028] It should be noted that the initial parameters in this application specifically include network structure parameters, activation function types, loss functions, optimizer configuration parameters, and training control parameters.

[0029] In specific implementation, the prediction model with multiple feature input channels can be constructed in the following way: First, divide the input channels by functional area dimension to construct a model structure with multiple input branches, such as using a multi-channel LSTM network or, inputting the associated identification features of each functional area into each channel; adopt a supervised training method to train the prediction model in combination with the historical energy consumption data of the target building, and iteratively optimize the initial parameters, and then the energy consumption prediction model can be trained in the following way: First, sort out the energy consumption data of the target building in the historical period to construct a supervised training data set, where the input is the associated identification features of each functional area at multiple consecutive time steps, and the label is the overall energy consumption value after the corresponding time step. Then divide the supervised training data set into a training set and a validation set according to a certain proportion and input them into the constructed multi-channel prediction model. During the training process, use the mean squared error (MSE) as the loss function, and use an optimizer (such as Adam) to execute the backpropagation algorithm to iteratively optimize the model parameters such as the initially set weights and biases. In each round of iteration, the model calculates the predicted value according to the current parameters, compares it with the true value to generate an error, and the error is backpropagated to update the parameters of each layer to minimize the loss function value. A preset iteration stop strategy can be introduced during the training process to prevent overfitting, and the generalization performance of the model can be monitored through the validation set. When the model reaches the convergence standard or the early termination condition on the validation set, output the trained energy consumption prediction model, which has the ability to predict the future overall energy consumption based on the energy consumption characteristics of each functional area.

[0030] In some embodiments, determining the suppression gain coefficients of the corresponding feature channels of each functional area in the energy consumption prediction model can be implemented by the following steps: Extract the deviation distribution characteristics between the energy consumption change rate of each functional area and the overall energy consumption change trend during the historical energy consumption monitoring period; Based on the deviation distribution characteristics, determine the response intensity of the corresponding feature channels of each functional area in the energy consumption prediction model; Set the initial suppression factors of the corresponding feature channels of each functional area according to all the response intensities; Dynamically adjust the initial suppression factors of each feature channel through the error feedback during the training process of the energy consumption prediction model to obtain the suppression gain coefficients of the corresponding feature channels of each functional area in the energy consumption prediction model.

[0031] It should be noted that the deviation distribution characteristics in this application are the characteristics that measure the degree of deviation between the energy consumption changes of each functional area and the overall energy consumption change trend; the initial suppression factor in this application refers to a preset weight coefficient used to limit the response intensity of the functional area feature channel; the suppression gain coefficient in this application is an index that quantifies the suppression degree of the contribution of the functional area feature channel to the energy consumption prediction model.

[0032] In specific implementation, first, the deviation distribution characteristics between the energy consumption change rate of each functional area and the overall energy consumption change trend within the historical energy consumption monitoring period can be achieved in the following manner, that is: within multiple historical energy consumption monitoring periods, calculate the energy consumption change rate sequence of each functional area, and synchronously extract the overall energy consumption change trend of the corresponding period. Use a sliding time window to calculate the deviation value between the two point by point, and statistically obtain the deviation distribution characteristics (such as mean, standard deviation) of each functional area; Secondly, determining the response intensity of the corresponding feature channels of each functional area in the energy consumption prediction model based on the deviation distribution characteristics can be achieved in the following manner, that is: use a mapping function (such as the Sigmoid function) to map the deviation distribution characteristics to the response values of the corresponding feature channels of each functional area. The mapping function can convert the deviation magnitude into a numerical expression of the response intensity, and use the calculated response value as the response intensity of the corresponding feature channels of each functional area; Then, setting the initial suppression factor of the corresponding feature channels of each functional area according to all the response intensities can be achieved in the following manner, that is: according to the response intensities of all functional areas, reasonably set the initial suppression factors of different channels. Usually, after normalization, a higher suppression factor is given to the larger deviation to reduce its excessive influence on the model; Finally, dynamically adjusting the initial suppression factor of each feature channel through the error feedback during the training process of the energy consumption prediction model to obtain the suppression gain coefficient of the corresponding feature channels of each functional area in the energy consumption prediction model can be achieved in the following manner, that is: during the training process of the energy consumption prediction model, combined with the model error feedback mechanism, use optimization algorithms such as gradient descent to dynamically adjust the initial suppression factor, continuously optimize the suppression factor according to the error backpropagation, and then use the optimized suppression factor as the suppression gain coefficient of the corresponding feature channels of the corresponding functional area.

[0033] In step 103, based on the historical energy consumption records of the target building, predict the energy consumption abnormal period in the next energy consumption monitoring period, and then perform a fuzzy evaluation on the energy consumption fluctuation level of the energy consumption abnormal period according to the load fluctuation characteristics of each functional area to obtain the fuzzy evaluation value of the energy consumption fluctuation level within the energy consumption abnormal period.

[0034] In some embodiments, predicting the energy consumption abnormal period in the next energy consumption monitoring period based on the historical energy consumption records of the target building can be achieved through the following steps: Obtain the historical overall energy consumption data from the historical energy consumption records of the target building; Extract the periodic characteristics of the historical overall energy consumption data, identify the time periods with energy consumption mutations and frequent abnormalities, and predict the energy consumption abnormal period in the next energy consumption monitoring period through the identified time periods.

[0035] It should be noted that the historical energy consumption records in this application refer to the energy consumption records within multiple historical monitoring periods; the energy consumption abnormal time period in this application refers to the time interval during which the overall energy consumption of the building shows drastic fluctuations or deviates from the normal mode within a specific time period.

[0036] In specific implementation, first, the daily overall energy consumption data is extracted from the historical energy consumption records of the target building to construct an energy consumption curve arranged in time series. Then, a periodic analysis method (such as Fourier transform) is used to extract the periodic components in the data for identifying the energy consumption fluctuation rules at different time scales. Secondly, a mutation detection algorithm (such as the cumulative sum algorithm based on moving average difference) is used to identify the mutation points such as sudden increases and sudden decreases in energy consumption in the curve. Further, by statistically analyzing the time positions and frequencies of these mutation points in previous cycles, high-frequency abnormal sections are identified, such as morning rush hours and season transition periods. Finally, based on the distribution patterns of these historical high-frequency abnormal time periods, time series clustering (such as K-means clustering of the start time of the mutation section) is used to predict the potential abnormal time period interval in the next energy consumption monitoring cycle.

[0037] In some embodiments, refer to Figure 3 As shown, this figure is a schematic flowchart for determining the fuzzy evaluation value in some embodiments of this application. In this embodiment, the energy consumption fluctuation level of the energy consumption abnormal time period is fuzzily evaluated according to the load fluctuation characteristics of each functional area, and the fuzzy evaluation value of the energy consumption fluctuation level within the energy consumption abnormal time period can be implemented by the following steps: In step 1031, a fuzzy membership function describing the energy consumption fluctuation is constructed based on the load fluctuation characteristics of each functional area in the historical period, and then the fuzzy membership of each functional area in different fluctuation intervals is determined. In step 1032, fuzzy reasoning is performed on the energy consumption fluctuation levels of each functional area within the energy consumption abnormal time period according to all the fuzzy memberships to obtain the energy consumption fluctuation credibility of each functional area at different fluctuation levels. In step 1033, the fuzzy evaluation value of the energy consumption fluctuation level within the energy consumption abnormal time period is determined through all the energy consumption fluctuation credibilities.

[0038] It should be noted that the fuzzy membership in this application is a measurement index used to represent the degree of belonging of an object to a specific fuzzy set; the energy consumption fluctuation credibility in this application refers to the degree of credibility indicating that the energy consumption fluctuation of a functional area belongs to a certain fluctuation level during the fuzzy evaluation process; the fuzzy evaluation value in this application is an index for measuring the overall energy consumption fluctuation degree within the energy consumption abnormal time period.

[0039] In specific implementation, first, a fuzzy membership function for describing energy consumption fluctuations is constructed based on the load fluctuation characteristics of each functional area in the historical period. Furthermore, the fuzzy membership degrees of each functional area in different fluctuation intervals can be determined in the following way: First, collect the load power data of each functional area in multiple historical periods, and calculate its load fluctuation characteristic indicators, such as standard deviation, variance or coefficient of variation. These indicators reflect the fluctuation amplitude and instability of the load. Based on the statistical results, divide the intervals of the load fluctuation amplitude, such as low fluctuation interval, medium fluctuation interval and high fluctuation interval. Usually, the interval boundaries are set in combination with expert experience or statistical distribution characteristics. Then, construct a fuzzy membership function for each interval. The commonly used function forms include triangular function and trapezoidal function. These functions map the actual load fluctuation values to membership values in the range of [0, 1], indicating the degree to which the value belongs to the corresponding fluctuation interval. Further, substitute the load fluctuation value of each functional area at the current time into the corresponding membership function to calculate its fuzzy membership degree in each fluctuation interval, and obtain a fuzzy set reflecting the fluctuation state of the area; According to all the fuzzy membership degrees, fuzzy inference is carried out on the energy consumption fluctuation levels of each functional area during the energy consumption abnormal period to obtain the energy consumption fluctuation credibility of each functional area at different fluctuation levels, which can be achieved in the following way: Take the fuzzy membership degrees of each functional area as input, construct a fuzzy rule base. The rule form usually adopts the "if-then" structure. For example, "if the load fluctuation membership degree is high, then the energy consumption fluctuation level is high". These rules can be determined based on expert experience or historical data statistics, and the Mamdani fuzzy inference method is used to perform fuzzy matching between the input fuzzy membership degree values and the rule base, calculate the activation intensity through the intersection and union operations of fuzzy sets, obtain the output fuzzy sets of each rule, and then synthesize the output fuzzy sets of all rules to form a comprehensive output membership function. Defuzzify the comprehensive output membership function through the centroid method or the maximum membership degree method to obtain the energy consumption fluctuation credibility of the functional area at different fluctuation levels; Finally, the fuzzy evaluation value of the energy consumption fluctuation level during the energy consumption abnormal period can be determined through all the energy consumption fluctuation credibilities, which can be achieved in the following way: Fuzzy fusion can be carried out on all the energy consumption fluctuation credibilities to obtain the membership function of the comprehensive energy consumption fluctuation of the entire building during this period, and then it is converted into a specific fuzzy evaluation value through the defuzzification method, which is used to quantitatively reflect the magnitude and credibility of the energy consumption fluctuation, and realize the overall fuzzy evaluation of the energy consumption fluctuation during the abnormal period.

[0040] It should be noted that, by introducing a fuzzy membership function, the present application quantitatively models the historical characteristics of load fluctuations in each functional area, realizes the conversion from quantitative data to fuzzy levels, effectively improves the inclusiveness and flexibility of the fluctuation range. Secondly, through a fuzzy inference mechanism, the fuzzy membership information of multiple functional areas is integrated to obtain the credibility distribution at each fluctuation level, enhancing the expression ability of the coupling relationship of multi-source data. Finally, based on the credibility of each area, the fuzzy evaluation value of the abnormal energy consumption period is comprehensively output, providing a path for dynamically, flexibly, and interpretable characterization of the energy consumption fluctuation level, thereby improving the accuracy of abnormal risk identification and the forward-looking response in building energy consumption management. This solution not only improves the accuracy of energy consumption fluctuation level evaluation but also enhances the adaptability of the evaluation system to uncertainties and regional differences.

[0041] In step 104, during the energy consumption prediction process of the next energy consumption monitoring period, when the fuzzy evaluation value exceeds the preset energy consumption fluctuation threshold, the cost constraint is imposed on the energy consumption feature attention of each functional area in the energy consumption prediction model according to all the suppression gain coefficients, and then the building energy consumption of the next energy consumption monitoring period is predicted based on the energy consumption prediction model after the cost constraint.

[0042] It should be noted that the judgment of "when the fuzzy evaluation value exceeds the preset energy consumption fluctuation threshold" in the solution of the present application mainly aims to realize the dynamic regulation of the behavior of the prediction model to cope with potential high-risk energy consumption fluctuations. The specific reasons are as follows: The fuzzy evaluation value is a comprehensive index for measuring the overall fluctuation level during the abnormal energy consumption period. Its increase represents that the building energy consumption is about to enter a high-fluctuation and unstable state. At this time, if we continue to rely on the conventional prediction model, it is easy to cause a significant increase in the overall prediction deviation due to the abnormal amplification effect of the energy consumption in local areas. By introducing the preset energy consumption fluctuation threshold as the decision boundary, it helps to divide the energy consumption prediction task into two categories: the normal state and the high-fluctuation state, realizing the differential processing of the model in different operating scenarios and improving the stability and safety of the prediction. Only when the evaluation value exceeds this threshold, the "cost constraint mechanism" is activated, and the feature attention of the abnormal dominant area is adjusted by introducing the suppression gain coefficient, so as to avoid the excessive interference of the abnormal area on the overall prediction result and enhance the robustness of the energy consumption prediction model.

[0043] It should also be noted that the energy consumption fluctuation threshold in the present application refers to the numerical boundary for judging whether there is a significant abnormal energy consumption fluctuation in a certain energy consumption monitoring period. Its setting scheme can be determined based on the statistical distribution characteristics of historical energy consumption data. Specifically, the percentile method can be used, and the upper bound value located in the 90% or 95% confidence interval is set as the threshold according to the fluctuation amplitude distribution to ensure that the identified fluctuations truly belong to the abnormal fluctuation area.

[0044] In some embodiments, the cost constraint on the energy consumption feature attention of each functional area in the energy consumption prediction model according to all suppression gain coefficients can be implemented by the following steps: Extract the energy consumption feature attention of each functional area in the energy consumption prediction model; Construct a cost function that reflects the deviation between the feature attention and its expected contribution according to the suppression gain coefficients of the corresponding feature channels of each functional area; Introduce the cost function into the parameter update process of the energy consumption prediction model, iteratively constrain the energy consumption feature attention of each functional area, and suppress the features of the functional areas with excessive contributions.

[0045] It should be noted that the energy consumption feature attention in this application is an index that measures the contribution degree of the energy consumption features of the functional area to the overall energy consumption prediction result.

[0046] In specific implementation, first, the extraction of the energy consumption feature attention of each functional area in the energy consumption prediction model can be implemented in the following way, that is: the attention distribution value of each functional area in the current prediction process can be obtained by traversing the attention distribution layer (such as the channel attention module) of the energy consumption prediction model; second, the construction of a cost function that reflects the deviation between the feature attention and its expected contribution according to the suppression gain coefficients of the corresponding feature channels of each functional area can be implemented in the following way, that is: take the suppression gain coefficient as the reference value of the expected attention weight of the corresponding functional area. Usually, the functional area with a smaller suppression gain coefficient should have a lower feature weight. To reflect the deviation between the actual attention and the expected contribution, a cost function can be constructed, which can be expressed in the common L2 norm way as: for the i-th functional area, the cost term is , where is the current attention of the model, is its suppression gain coefficient. This cost function is introduced into the total loss function of the model as a regularization term and participates in the gradient descent optimization process together with the main loss (such as MSE), so as to continuously suppress the attention distribution that does not conform to the expectation in model training, enhance the prediction stability and robustness; introducing the cost function into the parameter update process of the energy consumption prediction model and iteratively constraining the energy consumption feature attention of each functional area to suppress the features of the functional areas with excessive contributions can be implemented in the following way, that is: use the conventional gradient descent algorithm (such as the Adam optimizer) in the process of training the energy consumption prediction model, perform backpropagation and weight update on the total loss function, so as to dynamically adjust the attention value in each iteration to make it tend to match the suppression gain coefficient, effectively suppress the excessive feature contributions in the functional area, avoid overfitting of the model to some areas, and improve the robustness and generalization ability of the overall prediction.

[0047] It should be noted that, in the solution of this application, by imposing a cost constraint on the energy consumption feature attention in the energy consumption prediction model based on the suppression gain coefficients of all functional areas, it effectively avoids the over - contribution of some functional area features in the model, improves the generalization ability and prediction stability of the model. Compared with the method in the prior art where the feature weights of each area are fixed or adjusted without constraints, this solution dynamically suppresses abnormally high - contribution features, reduces the influence of noise and abnormal fluctuations on the prediction results, thereby improving the overall prediction accuracy and robustness, and meeting the modeling requirements for the diversity and non - linear fluctuations of complex building energy consumption.

[0048] It should also be noted that the prediction of the building energy consumption in the next energy consumption monitoring period by the energy consumption prediction model based on the cost constraint in this application means that the energy consumption prediction model after the cost constraint is used to predict the building energy consumption in the next energy consumption monitoring period.

[0049] On the other hand, in some embodiments, this application provides a building energy consumption prediction system. Refer to Figure 4 , this figure is a schematic structural diagram of the building energy consumption prediction system shown according to some embodiments of this application. The building energy consumption prediction system 400 includes: a collection module 401, a processing module 402, and an execution module 403, which are described as follows: The collection module 401. In this application, the collection module 401 is mainly used to collect the overall energy consumption data of the target building through low - frequency energy consumption monitoring devices and collect the local energy consumption data of each functional area of the target building through high - frequency energy consumption monitoring devices during the current energy consumption monitoring period; The processing module 402. In this application, the processing module 402 is used to perform feature mapping and association on the overall energy consumption and the local energy consumption of each functional area of the target building according to the overall energy consumption data and all local energy consumption data, obtain the associated identification features of the energy consumption between the whole and each functional area, and then construct an energy consumption prediction model based on all the associated identification features, and determine the suppression gain coefficients of the corresponding feature channels of each functional area in the energy consumption prediction model; In this application, the processing module 402 is also used to predict the energy consumption abnormal time period in the next energy consumption monitoring period based on the historical energy consumption records of the target building, and then perform a fuzzy evaluation on the energy consumption fluctuation level during the energy consumption abnormal time period according to the load fluctuation characteristics of each functional area to obtain the fuzzy evaluation value of the energy consumption fluctuation level during the energy consumption abnormal time period; The execution module 403. In this application, the execution module 403 is mainly used to, during the energy consumption prediction process in the next energy consumption monitoring period, when the fuzzy evaluation value exceeds the preset energy consumption fluctuation threshold, impose a cost constraint on the energy consumption feature attention of each functional area in the energy consumption prediction model according to all the suppression gain coefficients, and then predict the building energy consumption in the next energy consumption monitoring period based on the energy consumption prediction model after the cost constraint.

[0050] In addition, the present application also provides a computer device, which includes a memory and a processor. The memory stores code, and the processor is configured to obtain the code and execute the above-mentioned building energy consumption prediction method.

[0051] In some embodiments, referring to Figure 5 , this figure is a schematic structural diagram of a computer device for implementing the building energy consumption prediction method according to some embodiments of the present application. The building energy consumption prediction method in the above embodiments can be implemented by Figure 5 the computer device shown. The computer device 500 includes at least one processor 501, a communication bus 502, a memory 503, and at least one communication interface 504.

[0052] The processor 501 can be a general-purpose central processing unit (CPU) or an application-specific integrated circuit (ASIC).

[0053] The communication bus 502 can be used to transmit information between the above components.

[0054] The memory 503 can be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM), or other types of dynamic storage devices that can store information and instructions. It can also be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM), or other optical disc storage, optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disks, or other magnetic storage devices, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 503 can exist independently and be connected to the processor 501 through the communication bus 502. The memory 503 can also be integrated with the processor 501.

[0055] Among them, the memory 503 is used to store the program code for executing the solution of this application, and is controlled by the processor 501 for execution. The processor 501 is used to execute the program code stored in the memory 503. The program code may include one or more software modules. The building energy consumption prediction method in the above embodiment can be implemented by one or more software modules in the processor 501 and the program code in the memory 503.

[0056] The communication interface 504, using any device such as a transceiver, is used to communicate with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0057] In a specific implementation, as an embodiment, the computer device may include multiple processors, and each of these processors may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. The processor here may refer to one or more devices, circuits, and / or processing cores for processing data (such as computer program instructions).

[0058] The above computer device may be a general-purpose computer device or a special-purpose computer device. In a specific implementation, the computer device may be a desktop computer, a laptop computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. The embodiments of this application do not limit the type of the computer device.

[0059] In addition, this application also provides a computer-readable storage medium, and the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above building energy consumption prediction method is implemented.

[0060] Although the preferred embodiments of this application have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concepts. Therefore, the appended claims are intended to be construed to include the preferred embodiments and all changes and modifications falling within the scope of this application.

[0061] Obviously, those skilled in the art can make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalent technologies, this application is also intended to include these changes and modifications.

Claims

1. A method for predicting building energy consumption, wherein, Pre - deploy low - frequency energy consumption monitoring devices outside the target building in advance and high - frequency energy consumption monitoring devices in each functional area inside the target building. It is characterized in that the building energy consumption prediction method includes the following steps: In the current energy consumption monitoring period, collect the overall energy consumption data of the target building through the low - frequency energy consumption monitoring device, and collect the local energy consumption data of each functional area of the target building through the high - frequency energy consumption monitoring device; Based on the overall energy consumption data and all local energy consumption data, perform feature mapping correlation on the overall energy consumption of the target building and the local energy consumption of each functional area to obtain the correlation identification features of the energy consumption between the whole and each functional area. Then, based on all the correlation identification features, construct an energy consumption prediction model and determine the suppression gain coefficients of the corresponding feature channels of each functional area in the energy consumption prediction model; Based on the historical energy consumption records of the target building, predict the energy consumption abnormal time period in the next energy consumption monitoring period. Then, according to the load fluctuation characteristics of each functional area, perform a fuzzy evaluation on the energy consumption fluctuation level during the energy consumption abnormal time period to obtain the fuzzy evaluation value of the energy consumption fluctuation level during the energy consumption abnormal time period; In the process of predicting the energy consumption in the next energy consumption monitoring period, when the fuzzy evaluation value exceeds the preset energy consumption fluctuation threshold, impose a cost constraint on the energy consumption feature attention of each functional area in the energy consumption prediction model according to all the suppression gain coefficients. Then, based on the energy consumption prediction model after cost constraint, predict the building energy consumption in the next energy consumption monitoring period.

2. The method according to claim 1, characterized in that, Performing feature mapping correlation on the overall energy consumption of the target building and the local energy consumption of each functional area based on the overall energy consumption data and all local energy consumption data to obtain the correlation identification features of the energy consumption between the whole and each functional area specifically includes: Perform time alignment on the overall energy consumption data and the local energy consumption data of each functional area, and extract the correlation mapping features of the energy consumption between the whole and each functional area within the same time window; Determine the dynamic correlation coefficients between the local energy consumption data of each functional area and the overall energy consumption data; Determine the correlation identification features of the energy consumption between the whole and each functional area through all the dynamic correlation coefficients and all the correlation mapping features.

3. The method according to claim 1, wherein Constructing an energy consumption prediction model based on all the correlation identification features specifically includes: Construct a prediction model including multiple feature input channels, where each channel corresponds to a functional area; Respectively use the correlation identification features corresponding to each functional area as the input parameters of the corresponding channel, and set the initial parameters of the prediction model; Adopt a supervised training method to train the prediction model in combination with the historical energy consumption data of the target building, and iteratively optimize the initial parameters, and then train to obtain an energy consumption prediction model.

4. The method according to claim 1, wherein Determining the suppression gain coefficients of the corresponding feature channels of each functional area in the energy consumption prediction model specifically includes: Extract the deviation distribution features between the energy consumption change rate of each functional area in the historical energy consumption monitoring period and the overall energy consumption change trend; Based on the deviation distribution features, determine the response intensity of the corresponding feature channels of each functional area in the energy consumption prediction model; Set the initial suppression factors of the corresponding feature channels of each functional area according to all the response intensities; Dynamically adjust the initial suppression factors of each feature channel through error feedback during the training process of the energy consumption prediction model, and obtain the suppression gain coefficients of the feature channels corresponding to each functional area in the energy consumption prediction model.

5. The method according to claim 1, characterized in that, Predicting the energy consumption abnormal period in the next energy consumption monitoring period based on the historical energy consumption records of the target building specifically includes: Obtain the historical overall energy consumption data from the historical energy consumption records of the target building; Extract periodic features from the historical overall energy consumption data, identify the time periods with energy consumption mutations and frequent anomalies, and predict the energy consumption abnormal period in the next energy consumption monitoring period through the identified time periods.

6. The method according to claim 1, wherein Performing a fuzzy evaluation on the energy consumption fluctuation level of the energy consumption abnormal period according to the load fluctuation characteristics of each functional area to obtain the fuzzy evaluation value of the energy consumption fluctuation level within the energy consumption abnormal period specifically includes: Construct a fuzzy membership function describing energy consumption fluctuations based on the load fluctuation characteristics of each functional area in the historical period, and then determine the fuzzy membership degrees of each functional area in different fluctuation intervals; Perform fuzzy reasoning on the energy consumption fluctuation levels of each functional area within the energy consumption abnormal period according to all the fuzzy membership degrees to obtain the energy consumption fluctuation credibility of each functional area at different fluctuation levels; Determine the fuzzy evaluation value of the energy consumption fluctuation level within the energy consumption abnormal period through all the energy consumption fluctuation credibilities.

7. The method according to claim 1, wherein The low-frequency energy consumption monitoring device is specifically a multi-functional power meter.

8. An energy consumption prediction system for buildings, characterized in that, It includes: A collection module, configured to collect the overall energy consumption data of the target building through a low-frequency energy consumption monitoring device and collect the local energy consumption data of each functional area of the target building through a high-frequency energy consumption monitoring device during the current energy consumption monitoring period; A processing module, configured to perform feature mapping association on the overall energy consumption of the target building and the local energy consumption of each functional area according to the overall energy consumption data and all the local energy consumption data to obtain the associated identification features of the energy consumption between the whole and each functional area, and then construct an energy consumption prediction model based on all the associated identification features, and determine the suppression gain coefficients of the feature channels corresponding to each functional area in the energy consumption prediction model; The processing module is further configured to predict the energy consumption abnormal period in the next energy consumption monitoring period based on the historical energy consumption records of the target building, and then perform a fuzzy evaluation on the energy consumption fluctuation level of the energy consumption abnormal period according to the load fluctuation characteristics of each functional area to obtain the fuzzy evaluation value of the energy consumption fluctuation level within the energy consumption abnormal period; An execution module, configured to, during the energy consumption prediction process in the next energy consumption monitoring period, when the fuzzy evaluation value exceeds the preset energy consumption fluctuation threshold, perform cost constraints on the energy consumption feature attention of each functional area in the energy consumption prediction model according to all the suppression gain coefficients, and then predict the building energy consumption in the next energy consumption monitoring period based on the energy consumption prediction model after cost constraint.

9. A computer device, the computer device comprising a memory and a processor, the memory storing code, characterized in that, The processor is configured to obtain the code and execute the building energy consumption prediction method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the building energy consumption prediction method according to any one of claims 1 to 7.

Images