Building flexible load prediction method and system based on multi-source data fusion

Abstract

The present application provides a method and system for predicting building flexible load based on multi-source data fusion, which relates to the field of flexible load prediction. The method and system obtain flexible load comparison factors corresponding to multiple building comparison data groups respectively; perform flexible regression prediction based on the flexible load comparison factors and the building vehicle flow set to obtain the flexible load influence of the building vehicles; perform data fusion on multiple building comparison data groups according to the corresponding flexible load comparison factors to obtain a flexible load prediction data domain; perform data aggregation on the flexible load prediction data domain to obtain multiple flexible load prediction centers; and use multiple flexible load prediction centers to predict the building flexible load within the next flexible load prediction cycle. The present application can determine the degree of influence of the building vehicles on the flexible load based on the historical multi-source data of the building, thereby improving the accuracy of the building flexible load prediction under the condition of random interference of charging of building vehicles in the building.

Description

Technical Field

[0001] The present application relates to the technical field of flexible load prediction, and more specifically, to a method and system for predicting building flexible load based on multi-source data fusion. Background Art

[0002] Flexible loads in buildings refer to the part of the load that can be flexibly adjusted or managed in the energy demand of buildings. It mainly involves the energy consumption of buildings in terms of electricity, etc., and responds to fluctuations in power supply or balances the needs of energy systems by flexibly controlling these loads; flexible loads can be adjusted according to external conditions (such as changes in electricity prices, power supply shortages or fluctuations in renewable energy) to optimize energy efficiency, reduce building operating costs, and improve the stability of energy systems.

[0003] Building flexible load forecasting refers to the prediction of the adjustable energy demand or load of a building in the future through analysis and modeling. Among them, the charging of vehicles in the underground garage of the building is full of randomness. The charging time and charging mode of different vehicles are different, which greatly interferes with the accuracy of building flexible load forecasting. Therefore, how to improve the accuracy of building flexible load forecasting under the condition of random interference of building vehicle charging has become a problem that needs to be solved urgently. Summary of the invention

[0004] The present application provides a building flexible load prediction method and system based on multi-source data fusion, which can determine the degree of influence of construction vehicles on flexible load based on the historical multi-source data of the building, thereby improving the accuracy of building flexible load prediction when there is random interference from charging of construction vehicles in the building.

[0005] In a first aspect, the present application provides a building flexible load prediction method based on multi-source data fusion, and the building flexible load prediction method comprises the following steps:

[0006] Obtain multi-source data on historical buildings through Internet data services;

[0007] Separating the multi-source data of historical buildings based on different historical regression periods to obtain building comparison data groups corresponding to different historical regression periods;

[0008] According to the data distribution of each building comparison data group, a contrast test is performed respectively to obtain the flexible load contrast factors corresponding to the multiple building comparison data groups;

[0009] Acquire a construction vehicle flow set, perform flexible regression prediction based on the minimum value of the flexible load comparison factor and the construction vehicle flow set, and obtain the flexible load influence of the construction vehicles;

[0010] According to the corresponding flexible load comparison factors, data fusion is performed on multiple building comparison data groups to obtain a flexible load prediction data domain, and data aggregation is performed on the flexible load prediction data domain based on the flexible load influence of construction vehicles to obtain multiple flexible load prediction centers;

[0011] Multiple flexible load forecasting centers predict the building flexible load within the next flexible load forecasting cycle and dispatch the building power load based on the forecast results.

[0012] In this embodiment, the historical building multi-source data is separated based on different historical regression periods to obtain the building comparison data groups corresponding to different historical regression periods, which specifically include:

[0013] Get multiple historical regression periods;

[0014] Sampling the multi-source building data according to a first historical regression cycle to obtain a first building comparison data group;

[0015] The same method is used to obtain the building comparison data groups corresponding to other historical regression periods.

[0016] In this embodiment, contrast tests are performed respectively according to the data distribution of each building comparison data group, and the flexible load contrast factors corresponding to the multiple building comparison data groups are obtained, which specifically include:

[0017] Acquire a first building comparison data group, and acquire a historical regression period corresponding to the first building comparison data group; normalize each data dimension in the first building comparison data group to obtain a first normalized data group;

[0018] determining a flexible load comparison factor of the first building comparison data group according to the data distribution of each data dimension in the first normalized data group;

[0019] The same method is used to determine the flexible load comparison factors corresponding to other building comparison data groups.

[0020] In this embodiment, flexible regression prediction is performed based on the minimum value of the flexible load comparison factor and the construction vehicle flow set to obtain the flexible load influence of the construction vehicle, which specifically includes:

[0021] Obtaining a preset flexible load forecasting cycle, extracting signal distribution characteristics of the construction vehicle flow set according to the flexible load forecasting cycle, and obtaining monitoring flow distribution characteristics;

[0022] Determine the flexible regression prediction depth according to the monitored flow distribution characteristics, and predict the construction vehicle flow value of the next flexible load prediction cycle according to the flexible regression prediction depth;

[0023] The flow value of the nearest construction vehicle is obtained, and the flexible load influence of the construction vehicle is determined according to the flow value of the nearest construction vehicle, the flow value of the construction vehicle in the next flexible load prediction period, and the minimum value of the flexible load comparison factor.

[0024] In this embodiment, a moving average autoregressive model is used to predict the construction vehicle flow value of the next flexible load prediction period.

[0025] In this embodiment, data fusion is performed on multiple building comparison data groups according to corresponding flexible load comparison factors, and the flexible load prediction data domain obtained specifically includes:

[0026] Obtain multiple building comparison data groups and their corresponding flexible load comparison factors;

[0027] The flexible load contrast factors in each time zone in the multi-layer overlapping area are regionally integrated, and the flexible load influence characteristics corresponding to different time zones are obtained by regionally assigning values ​​to each time zone in the single-layer overlapping area.

[0028] According to the flexible load impact characteristics corresponding to different time zones, multiple building comparison data groups are fused to obtain the flexible load prediction data domain.

[0029] In this embodiment, the flexible load prediction data domain is aggregated based on the flexible load influence of the construction vehicles to obtain multiple flexible load prediction centers, specifically including:

[0030] Obtain the flexible load influence of the construction vehicle, perform influence mapping based on the flexible load influence of the construction vehicle, and obtain the predicted center dispersion quantity;

[0031] The flexible load forecasting data domain is clustered according to the dispersed number of the forecasting centers to obtain a plurality of flexible load forecasting centers.

[0032] In this embodiment, the InfluxDB time series database is used to store the multi-source data of the historical buildings.

[0033] In this embodiment, a camera on a construction road is used to collect the construction vehicle flow set.

[0034] In a second aspect, the present application provides a building flexible load prediction system based on multi-source data fusion, which is used to execute a building flexible load prediction method based on multi-source data fusion, and the building flexible load prediction system includes:

[0035] A multi-source data acquisition module is used to acquire multi-source data of historical buildings through Internet data services;

[0036] A data processing module, used for performing data separation on the multi-source data of historical buildings based on different historical regression periods, to obtain building comparison data groups corresponding to different historical regression periods;

[0037] The data processing module is further used to perform contrast tests according to the data distribution of each building comparison data group, and obtain the flexible load contrast factors corresponding to the multiple building comparison data groups;

[0038] The data processing module is further used to obtain a set of construction vehicle flow, perform flexible regression prediction based on the minimum value of the flexible load comparison factor and the set of construction vehicle flow, and obtain the flexible load influence of the construction vehicle;

[0039] The data processing module is further used to perform data fusion on multiple building comparison data groups according to corresponding flexible load comparison factors to obtain a flexible load prediction data domain, and perform data aggregation on the flexible load prediction data domain based on the flexible load influence of construction vehicles to obtain multiple flexible load prediction centers;

[0040] The execution module is used for multiple flexible load forecasting centers to predict the building flexible load within the next flexible load forecasting cycle and dispatch the building power load according to the prediction results.

[0041] The technical solution provided by the embodiments disclosed in this application has the following beneficial effects:

[0042] Obtain multi-source data of historical buildings through Internet data services; separate the multi-source data of buildings based on different historical regression periods to obtain building comparison data groups corresponding to different historical regression periods, which provides a basis for analyzing the electricity consumption patterns of buildings and identifying load changes in different historical regression periods; perform contrast tests based on the data distribution of each building comparison data group to obtain flexible load comparison factors corresponding to multiple building comparison data groups. The flexible load comparison factor reflects the degree of coupling between the building flexible load-related data within the building comparison data group, thereby identifying the regular fluctuations of building loads within different historical regression periods; obtain a short-term building vehicle flow set, and perform flexible regression based on the minimum value of the flexible load comparison factor and the short-term building vehicle flow set. Prediction is performed to obtain the flexible load influence of construction vehicles, that is, the influence of vehicle flow in the building on the building load is determined through regression analysis; then, data fusion is performed on multiple building comparison data groups according to the corresponding flexible load comparison factors to obtain a flexible load prediction data domain, and data aggregation is performed on the flexible load prediction data domain based on the flexible load influence of construction vehicles to obtain multiple flexible load prediction centers, and the flexible load prediction center is determined by fusing the influence of historical data and vehicle flow data, which can comprehensively consider the joint effects of multi-source factors and random interference of charging of construction vehicles, and reduce the prediction deviation caused by a single factor; multiple flexible load prediction centers predict the building flexible load in the next flexible load prediction period, thereby improving the accuracy of flexible load prediction.

[0043] In summary, the technical solution adopted in this application can determine the impact of construction vehicles on flexible loads based on the historical multi-source data of the building, thereby improving the accuracy of building flexible load prediction when random interference occurs in the charging of construction vehicles in the building. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0045] Figure 1 It is a flow chart of a building flexible load prediction method based on multi-source data fusion provided by the present application;

[0046] Figure 2 This is a module structure diagram of a building flexible load prediction system based on multi-source data fusion provided by this application. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0048] The embodiment of the present application provides a building flexible load prediction method and system based on multi-source data fusion, the core of which is to obtain multi-source data of historical buildings through Internet data services; to separate the multi-source data of historical buildings based on different historical regression periods to obtain building comparison data groups corresponding to different historical regression periods; to perform contrast tests based on the data distribution of each building comparison data group to obtain flexible load comparison factors corresponding to multiple building comparison data groups; to obtain a building vehicle flow set, to perform flexible regression prediction based on the minimum value of the flexible load comparison factor and the building vehicle flow set to obtain the flexible load influence of the building vehicles; to fuse the multiple building comparison data groups according to the corresponding flexible load comparison factors to obtain a flexible load prediction data domain, to aggregate the flexible load prediction data domain based on the flexible load influence of the building vehicles to obtain multiple flexible load prediction centers; to predict the building flexible load within the next flexible load prediction period by the multiple flexible load prediction centers, and to dispatch the building power load according to the prediction results.

[0049] Embodiment 1

[0050] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods. Figure 1 As shown in FIG. 1 , this figure is an exemplary flow chart of a building flexible load prediction method based on multi-source data fusion according to this embodiment of the present application, and the building flexible load prediction method includes the following steps:

[0051] In step S1, multi-source data of historical buildings are obtained through Internet data services.

[0052] Optionally, in this embodiment, the multi-source data of historical buildings are collected and recorded and then uploaded to the Internet cloud platform. Then, when conducting building flexible load forecasting, the multi-source data of historical buildings stored in the Internet cloud platform is obtained through the Internet cloud platform data storage service. In specific implementation, a time series database such as InfluxDB can be used to store the multi-source data of historical buildings.

[0053] It should be noted that the multi-source data of historical buildings includes multiple data sources from inside and outside the buildings. The multi-source data of historical buildings can comprehensively reflect the energy consumption, operating status and environmental impact of the buildings. Among them, the more important data is the building flexible load impact data. The building flexible load impact data has multiple data dimensions, each of which is coupled with each other and will independently have a corresponding impact on the building flexible load.

[0054] Optionally, in this embodiment, the historical building multi-source data includes: historical building energy consumption data, historical air conditioning equipment operation power consumption data, historical environmental temperature and humidity data, and historical user behavior data, which are respectively described as follows:

[0055] The energy consumption data of historical buildings include: Electricity usage data: the overall electricity consumption of the building, usually collected in real time through smart meters. Including electricity consumption, peak and valley electricity price response, etc.; Itemized energy consumption data: Energy consumption monitoring data for specific equipment (such as air conditioning, lighting, heating, etc.), which can describe the usage load of various types of equipment in detail;

[0056] The historical air-conditioning equipment operating power consumption data includes: air-conditioning and refrigeration equipment data: including temperature settings, operating hours, start and stop status, etc., which directly affect the building's power load; heating system data: heating equipment operating status, indoor temperature adjustment and other data, especially for winter heating needs.

[0057] The historical environmental temperature and humidity data include: outdoor temperature range and humidity range, and outdoor average light intensity;

[0058] The historical user behavior data includes: user occupancy rate and number of occupants, etc.

[0059] In step S2, the multi-source data of historical buildings are separated based on different historical regression periods to obtain building comparison data groups corresponding to different historical regression periods.

[0060] It should be noted that the historical regression period is a sampling period for selecting flexible load related data for data prediction. Different historical regression periods have different influencing factors on the flexible load. For example, when the historical regression period is one year, the factor affecting the flexible load is usually seasonal climate. When the historical regression period is one day, the factor affecting the flexible load is usually the weather on that day and the attendance of building personnel. Optionally, in this embodiment, the historical building multi-source data is separated based on different historical regression periods to obtain the building comparison data groups corresponding to different historical regression periods, which specifically include:

[0061] Get multiple historical regression periods;

[0062] Sampling the multi-source building data according to a first historical regression cycle to obtain a first building comparison data group;

[0063] The same method is used to obtain the building comparison data groups corresponding to other historical regression periods.

[0064] Optionally, in this embodiment, the multiple historical regression periods include: 24h, 48h, 72h, 30 days, 60 days, 90 days, 1 year, 2 years and 3 years, and data sampling is performed on the building multi-source data based on different historical regression periods to obtain building comparison data groups corresponding to different historical regression periods, wherein a time area that is sampled only once is a single-layer overlapping area, and a time area that is repeatedly sampled is a multi-layer overlapping area. For example, the building multi-source data of the past 144h will be sampled three times by the three building comparison data groups corresponding to 24h, 48h and 72h, and the past 144h is a time area in the multi-layer overlapping area.

[0065] In step S3, contrast tests are performed respectively according to the data distribution of each building comparison data group to obtain the flexible load contrast factors corresponding to the multiple building comparison data groups.

[0066] It should be noted that the flexible load contrast factor reflects the degree of coupling between the building flexible load related data in the building comparison data group. The larger the flexible load contrast factor is, the higher the correlation between the flexible load related data in the historical regression cycle is, and the stronger the periodic regularity of the data is. Therefore, the prediction accuracy when performing regression prediction in the historical regression cycle is better. Optionally, in this embodiment, contrast tests are performed respectively according to the data distribution of each building comparison data group, and the flexible load contrast factors corresponding to the multiple building comparison data groups are obtained, which specifically include:

[0067] Acquire a first building comparison data group, and acquire a historical regression period corresponding to the first building comparison data group; normalize each data dimension in the first building comparison data group to obtain a first normalized data group;

[0068] determining a flexible load comparison factor of the first building comparison data group according to the data distribution of each data dimension in the first normalized data group;

[0069] The same method is used to determine the flexible load comparison factors corresponding to other building comparison data groups.

[0070] Optionally, in this embodiment, the data dimensions in the first building comparison data group include: historical building energy consumption data, historical air-conditioning equipment operating power consumption data, historical ambient temperature and humidity data, and energy storage equipment energy storage data. The Pearson correlation coefficients between the four data dimensions can be obtained, and the average of the absolute values ​​of all Pearson correlation coefficients can be taken as the flexible load comparison factor of the first building comparison data group.

[0071] In step S4, a construction vehicle flow set is obtained, and a flexible regression prediction is performed based on the minimum value of the flexible load comparison factor and the construction vehicle flow set to obtain the flexible load influence of the construction vehicles.

[0072] Optionally, in this embodiment, the flexible regression prediction is performed according to the minimum value of the flexible load comparison factor and the construction vehicle flow set to obtain the flexible load influence of the construction vehicle, which specifically includes:

[0073] Obtaining a preset flexible load forecasting cycle, extracting signal distribution characteristics of the construction vehicle flow set according to the flexible load forecasting cycle, and obtaining monitoring flow distribution characteristics;

[0074] Determine the flexible regression prediction depth according to the monitored flow distribution characteristics, and predict the construction vehicle flow value of the next flexible load prediction cycle according to the flexible regression prediction depth;

[0075] The flow value of the nearest construction vehicle is obtained, and the flexible load influence of the construction vehicle is determined according to the flow value of the nearest construction vehicle, the flow value of the construction vehicle in the next flexible load prediction period, and the minimum value of the flexible load comparison factor.

[0076] In specific implementation, the variance value of the construction vehicle flow value in the construction vehicle flow set can be used as the monitoring flow distribution feature. The flexible regression prediction depth is the sample size when predicting the construction vehicle flow value of the next flexible load forecasting cycle. The larger the monitoring flow distribution feature, the more flow uncertainty factors there are in the construction vehicle flow set. At this time, it is necessary to increase the flexible regression prediction depth to improve the prediction accuracy when predicting the construction vehicle flow value of the next flexible load forecasting cycle. Optionally, in this embodiment, the flexible regression prediction depth can be determined according to the threshold interval where the monitoring flow distribution feature is located. For example, the flexible regression prediction depth can be 10, 20, 30, 40, and 50, a total of five flexible regression prediction depths, corresponding to the 5 threshold intervals where the monitoring flow distribution feature is located. The threshold interval can be calibrated according to multiple experiments, which will not be repeated here.

[0077] It should be noted that the flexible load influence of the construction vehicle represents the degree of influence of the random changes in the charging of the construction vehicle on the flexible load prediction of the building within the next flexible load prediction cycle. The present application uses the minimum value of the flexible load comparison factor to correct the flexible load influence of the construction vehicle, and can correct the flexible load influence of the construction vehicle based on the data coupling of the multi-source data of historical buildings. Therefore, when the data coupling and regularity of the multi-source data of historical buildings are strong, the flexible load influence of the construction vehicle is reduced, thereby improving the accuracy of the final building flexible load prediction. Optionally, in this embodiment, a moving average autoregressive model is used to predict the construction vehicle flow value of the next flexible load prediction cycle, wherein the flexible load influence of the construction vehicle = the minimum value of the flexible load comparison factor × (|construction vehicle flow value of the next flexible load prediction cycle - the nearest construction vehicle flow value| / the mean value of the monitored flow in the construction vehicle flow set).

[0078] Optionally, in this embodiment, the flexible load prediction period is a prediction period for performing flexible load prediction. In specific implementation, the flexible load prediction period can be preset to 24 hours.

[0079] In step S5, data fusion is performed on multiple building comparison data groups according to corresponding flexible load comparison factors to obtain a flexible load prediction data domain, and data aggregation is performed on the flexible load prediction data domain based on the flexible load influence of construction vehicles to obtain multiple flexible load prediction centers.

[0080] It should be noted that the flexible load prediction data domain is a data space composed of flexible load impact data and impact feature dimensions. Optionally, in this embodiment, data fusion is performed on multiple building comparison data groups according to corresponding flexible load comparison factors to obtain the flexible load prediction data domain, which specifically includes:

[0081] Obtain multiple building comparison data groups and their corresponding flexible load comparison factors;

[0082] The flexible load contrast factors in each time zone in the multi-layer overlapping area are regionally integrated, and the flexible load influence characteristics corresponding to different time zones are obtained by regionally assigning values ​​to each time zone in the single-layer overlapping area.

[0083] According to the flexible load impact characteristics corresponding to different time zones, multiple building comparison data groups are fused to obtain the flexible load prediction data domain.

[0084] In specific implementation, the area set of time areas that are sampled only once is a single-layer overlapping area, and the area set of time areas that are repeatedly sampled is a multi-layer overlapping area. In this embodiment, during the process of regional fusion of the flexible load comparison factors in each time area in the multi-layer overlapping area, for example, there are three samples in the first time area in the multi-layer overlapping area, and the average value of the flexible load comparison factors of the three building comparison data groups corresponding to the three samples is used as the flexible load influence feature of the first time area. When performing regional assignment for each time area in the single-layer overlapping area, the flexible load comparison factor of the building comparison data group corresponding to each time area when sampled is assigned data as the flexible load influence feature corresponding to the time area.

[0085] Optionally, in this embodiment, the process of fusing multiple building comparison data groups according to the flexible load influence characteristics corresponding to different time zones to obtain the flexible load prediction data domain can be to compose influence characteristic dimensions according to the flexible load influence characteristics corresponding to different time zones, thereby performing data dimensionality upgrade on the data space where the flexible load influence data in the multiple building comparison data groups are located to obtain the flexible load prediction data domain.

[0086] It should be noted that the flexible load influence of the construction vehicle indicates the influence of the charging change of the construction vehicle on the flexible load prediction of the building in the next flexible load prediction cycle. When the flexible load influence of the construction vehicle is large, it is necessary to increase the number of prediction centers for the flexible load prediction of the next flexible load prediction cycle, thereby increasing the amount of comparison data during the flexible load prediction, and avoiding the interference of the random charging of the construction vehicle on the flexible load prediction. Optionally, in this embodiment, the flexible load prediction data domain is aggregated based on the flexible load influence of the construction vehicle to obtain multiple flexible load prediction centers, which specifically include:

[0087] Obtain the flexible load influence of the construction vehicle, perform influence mapping based on the flexible load influence of the construction vehicle, and obtain the predicted center dispersion quantity;

[0088] The flexible load forecasting data domain is clustered according to the dispersed number of the forecasting centers to obtain a plurality of flexible load forecasting centers.

[0089] It should be noted that the flexible load prediction center described in the present application is a prediction data center determined based on historical multi-source data, and can predict the building power load of the next flexible load prediction cycle based on historical data experience. Optionally, in this embodiment, the corresponding number of dispersed prediction centers can be obtained according to the threshold interval of the flexible load influence of the construction vehicle according to the preset linear mapping table. The number of dispersed prediction centers is the initial number of clustering centers when clustering the data in the flexible load prediction data domain. In specific implementation, multiple initial clustering centers can be randomly selected in the flexible load prediction data domain according to the initial number of clustering centers, and then the K-means clustering method is used to cluster the data in the multiple initial clustering centers to obtain multiple flexible load prediction centers.

[0090] In step S6, a plurality of flexible load forecasting centers forecast the building flexible load within the next flexible load forecasting cycle, and dispatch the building power load according to the forecasting results.

[0091] In this embodiment, the prediction of the building flexible load in the next flexible load prediction cycle by multiple flexible load prediction centers specifically includes:

[0092] Collect the flexible load impact data of each dimension, perform data prediction on the flexible load impact data of each dimension respectively, and obtain the prediction data corresponding to each dimension in the next flexible load prediction cycle;

[0093] The load prediction data are composed according to the prediction data corresponding to each dimension, and the similarity between the load prediction data and each flexible load prediction center is calculated respectively. The building flexible load prediction result is determined according to the building power and similarity corresponding to each flexible load prediction center.

[0094] Optionally, in this embodiment, the flexible load impact data of each dimension include: historical building energy consumption data dimension, historical air-conditioning equipment operation power consumption data dimension, historical ambient temperature and humidity data dimension, and historical user behavior data dimension. In specific implementation, a moving average autoregressive model can be used to predict the flexible load impact data of each dimension to obtain the prediction data corresponding to each dimension in the next flexible load prediction cycle.

[0095] Optionally, in this embodiment, the ratio of the Euclidean distance between the load forecast data and each flexible load forecast center in the data space to the preset standard Euclidean distance can be used as the initial similarity, and the initial similarity can be corrected according to the flexible load influence characteristics of the influence characteristic dimension of the flexible load forecast center to obtain the similarity between the load forecast data and the flexible load forecast center. In specific implementation, the similarity between the load forecast data and the flexible load forecast center can be obtained by multiplying the flexible load influence characteristics of the influence characteristic dimension of the flexible load forecast center by the initial similarity.

[0096] Optionally, in this embodiment, based on the building power and similarity corresponding to each flexible load prediction center, determining the building flexible load prediction result specifically includes: using the similarity corresponding to each flexible load prediction center as a weight coefficient, weighted fusion of the building power corresponding to each flexible load prediction center, to obtain the building power prediction power within the next flexible load prediction cycle, and using the building power prediction power as the prediction result of the building flexible load.

[0097] Embodiment 2

[0098] This application provides a building flexible load prediction system based on multi-source data fusion, referring to Figure 2 As shown, this figure is a schematic diagram of a building flexible load prediction system according to this embodiment of the present application, and the building flexible load prediction system includes:

[0099] The multi-source data acquisition module 100 is used to acquire multi-source data of historical buildings through Internet data services;

[0100] The data processing module 200 is used to separate the multi-source data of historical buildings based on different historical regression periods to obtain building comparison data groups corresponding to different historical regression periods;

[0101] The data processing module 200 is further used to perform contrast tests according to the data distribution of each building comparison data group, and obtain the flexible load contrast factors corresponding to the multiple building comparison data groups;

[0102] The data processing module 200 is further used to obtain a set of construction vehicle flow, perform flexible regression prediction based on the minimum value of the flexible load comparison factor and the set of construction vehicle flow, and obtain the flexible load influence of the construction vehicle;

[0103] The data processing module 200 is further used to perform data fusion on multiple building comparison data groups according to corresponding flexible load comparison factors to obtain a flexible load prediction data domain, and perform data aggregation on the flexible load prediction data domain based on the flexible load influence of construction vehicles to obtain multiple flexible load prediction centers;

[0104] The execution module 300 is used to predict the building flexible load within the next flexible load prediction cycle by multiple flexible load prediction centers, and dispatch the building power load according to the prediction results.

[0105] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0106] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, the storage medium including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically-erasable programmable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0107] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

Claims

1. A building flexible load prediction method based on multi-source data fusion, characterized in that: The building flexible load prediction method comprises the following steps: Obtain multi-source data on historical buildings through Internet data services; Separating the multi-source data of historical buildings based on different historical regression periods to obtain building comparison data groups corresponding to different historical regression periods; According to the data distribution of each building comparison data group, a contrast test is performed respectively to obtain the flexible load contrast factors corresponding to the multiple building comparison data groups; Acquire a construction vehicle flow set, perform flexible regression prediction based on the minimum value of the flexible load comparison factor and the construction vehicle flow set, and obtain the flexible load influence of the construction vehicles; According to the corresponding flexible load comparison factors, data fusion is performed on multiple building comparison data groups to obtain a flexible load prediction data domain, and data aggregation is performed on the flexible load prediction data domain based on the flexible load influence of construction vehicles to obtain multiple flexible load prediction centers; Multiple flexible load forecasting centers forecast the building flexible load within the next flexible load forecasting cycle, and dispatch the building power load according to the forecasting results; Among them, the contrast test is performed according to the data distribution of each building comparison data group, and the flexible load comparison factors corresponding to the multiple building comparison data groups are obtained, which specifically include: Acquire a first building comparison data group, and acquire a historical regression period corresponding to the first building comparison data group; normalize each data dimension in the first building comparison data group to obtain a first normalized data group; determining a flexible load comparison factor of the first building comparison data group according to the data distribution of each data dimension in the first normalized data group; The flexible load comparison factors corresponding to other building comparison data groups are determined in the same way; Among them, the flexible regression prediction is performed according to the minimum value of the flexible load comparison factor and the construction vehicle flow set to obtain the flexible load influence of the construction vehicle, which specifically includes: Obtaining a preset flexible load forecasting cycle, extracting signal distribution characteristics of the construction vehicle flow set according to the flexible load forecasting cycle, and obtaining monitoring flow distribution characteristics; Determine the flexible regression prediction depth according to the monitored flow distribution characteristics, and predict the construction vehicle flow value of the next flexible load prediction cycle according to the flexible regression prediction depth; Obtaining the nearest construction vehicle flow value, and determining the flexible load influence of the construction vehicle according to the nearest construction vehicle flow value, the construction vehicle flow value of the next flexible load prediction period, and the minimum value of the flexible load comparison factor; Among them, the data of multiple building comparison data groups are fused according to the corresponding flexible load comparison factors, and the flexible load prediction data domain is obtained, which specifically includes: Obtain multiple building comparison data groups and their corresponding flexible load comparison factors; The flexible load contrast factors in each time zone in the multi-layer overlapping area are regionally integrated, and the flexible load influence characteristics corresponding to different time zones are obtained by regionally assigning values ​​to each time zone in the single-layer overlapping area. According to the flexible load impact characteristics corresponding to different time zones, multiple building comparison data groups are fused to obtain the flexible load prediction data domain; The single-layer overlapping region is a region set of time regions that are sampled only once, and the multi-layer overlapping region is a region set of time regions that are repeatedly sampled; Based on the flexible load influence of construction vehicles, data aggregation is performed on the flexible load prediction data domain to obtain multiple flexible load prediction centers, specifically including: Obtain the flexible load influence of the construction vehicle, perform influence mapping based on the flexible load influence of the construction vehicle, and obtain the predicted center dispersion quantity; Performing data clustering on the flexible load forecasting data domain according to the dispersed number of the forecasting centers to obtain a plurality of flexible load forecasting centers; Among them, the data dimensions in the first building comparison data group include: historical building energy consumption data, historical air-conditioning equipment operation power consumption data, historical ambient temperature and humidity data, and energy storage equipment energy storage data. The Pearson correlation coefficients between the four data dimensions are obtained, and the average of the absolute values ​​of all Pearson correlation coefficients is taken as the flexible load comparison factor of the first building comparison data group.

2. A building flexible load prediction method based on multi-source data fusion as claimed in claim 1, characterized in that: The historical building multi-source data is separated based on different historical regression periods to obtain the building comparison data groups corresponding to different historical regression periods, which specifically include: Get multiple historical regression periods; Sampling the multi-source building data according to a first historical regression cycle to obtain a first building comparison data group; The same method is used to obtain the building comparison data groups corresponding to other historical regression periods.

3. The method for predicting building flexible load based on multi-source data fusion according to claim 1, characterized in that: The moving average autoregressive model is used to predict the construction vehicle flow value in the next flexible load forecasting period.

4. The method for predicting building flexible load based on multi-source data fusion according to claim 1, characterized in that: The InfluxDB time series database is used to store the multi-source data of the historical buildings.

5. The method for predicting building flexible load based on multi-source data fusion according to claim 1, characterized in that: The construction vehicle flow set is collected using cameras on the construction roads.

6. A building flexible load prediction system based on multi-source data fusion, used to execute a building flexible load prediction method based on multi-source data fusion as claimed in any one of claims 1 to 5, characterized in that: The building flexible load prediction system comprises: A multi-source data acquisition module is used to acquire multi-source data of historical buildings through Internet data services; A data processing module, used for performing data separation on the multi-source data of historical buildings based on different historical regression periods, to obtain building comparison data groups corresponding to different historical regression periods; The data processing module is further used to perform contrast tests according to the data distribution of each building comparison data group, and obtain the flexible load contrast factors corresponding to the multiple building comparison data groups; The data processing module is further used to obtain a set of construction vehicle flow, perform flexible regression prediction based on the minimum value of the flexible load comparison factor and the set of construction vehicle flow, and obtain the flexible load influence of the construction vehicle; The data processing module is further used to perform data fusion on multiple building comparison data groups according to corresponding flexible load comparison factors to obtain a flexible load prediction data domain, and perform data aggregation on the flexible load prediction data domain based on the flexible load influence of construction vehicles to obtain multiple flexible load prediction centers; The execution module is used for multiple flexible load forecasting centers to predict the building flexible load within the next flexible load forecasting cycle and dispatch the building power load according to the prediction results.

Images