Property energy consumption anomaly monitoring method and system based on big data analysis

Abstract

The application discloses a property energy consumption anomaly monitoring method and system based on big data analysis, and relates to the technical field of energy consumption anomaly management. The property energy consumption anomaly monitoring method based on big data analysis comprises the following steps: in a specified energy consumption monitoring period, first, a specified property service area energy consumption power group is acquired, and energy consumption anomalies are screened by comparing the energy consumption power group with an energy consumption prediction interval; then, according to the comparison result, repeated inspection is performed on the energy consumption anomaly data of the abnormal energy consumption period, transient normal fluctuations caused by normal working condition switching of equipment or sudden changes in external environment are excluded, and the repeated inspection comprises peak value anomaly detection and rechecking verification; finally, after the repeated inspection is completed, correlation verification is performed in combination with an abnormal energy consumption feature set. The method provides a scientific basis for property energy consumption management, helps to accurately locate energy consumption anomaly points, effectively solves the problem of misjudgment of anomaly recognition caused by transient changes in the prior art, and realizes energy saving, consumption reduction and cost optimization.

Description

Technical Field

[0001] This invention relates to the field of energy consumption anomaly management technology, and in particular to a method and system for monitoring property energy consumption anomalies based on big data analysis. Background Technology

[0002] Driven by the widespread adoption of the Internet of Things (IoT) and smart sensors, high-precision, multi-dimensional real-time data collection of property energy consumption and related environmental data has been achieved. Meanwhile, the maturity of big data processing technology has facilitated the deep integration of energy consumption anomaly monitoring and intelligent systems. By deploying sensors and smart devices in facilities, energy consumption data, including electricity, water, and gas consumption, is continuously collected and integrated with information such as equipment operating status, environmental parameters, and time characteristics. This data is transmitted via network to a central platform for centralized storage. The system then performs data cleaning and standardization to eliminate noise and ensure quality. Big data analytics are then used to compare the pre-processed real-time data with historical benchmarks to identify abnormal consumption events that deviate from normal ranges, such as sudden increases or decreases. Once an anomaly is detected, the system automatically generates an alarm and pushes it to property management personnel, triggering preset response mechanisms such as adjusting equipment operation or notifying the maintenance team to intervene, thereby optimizing energy use and preventing potential malfunctions in a timely manner.

[0003] For example, Chinese invention patent with publication number CN119539762A discloses a smart park property operation and maintenance management system based on big data, including: a big data acquisition and processing module for collecting various categories of smart park data, including energy consumption data, population density data, equipment operation and maintenance data, and park environmental maintenance data; an intelligent lighting and energy management module for acquiring intelligent lighting energy saving and energy optimization operation data; a building automation and security module for monitoring the operation of park cameras, access control and elevators; and a park environmental maintenance module for real-time updating of park environmental maintenance data.

[0004] For example, Chinese invention patent CN118333247B discloses a method and system for assessing property energy consumption based on deep learning, including: obtaining a query operation for property energy consumption; obtaining the corresponding data and energy consumption attributes of the target park equipment based on the identity verification of the query operation; obtaining the park energy consumption assessment result of each energy consumption type assessment indicator based on the obtained data and energy consumption attributes; and obtaining the target property energy consumption assessment based on the obtained new park energy budget data and energy consumption assessment result.

[0005] The above-mentioned technology has at least the following technical problems:

[0006] In existing technologies, property energy consumption is diverse, encompassing various types such as electricity consumption, water usage, gas supply, lighting demand, and air conditioning operation. The data volume is large, highly dynamic, and comes from diverse sources. Monitoring equipment cannot provide real-time abnormal feedback on sudden increases or decreases in energy consumption within milliseconds to seconds caused by sudden equipment startup, large motors, elevator sudden stops, or rapid changes in the external environment such as sudden temperature changes or strong winds. This means that when monitoring the time characteristics of transient energy consumption fluctuations, energy consumption anomaly monitoring systems may lose true transient anomaly characteristics. This transient delay and failure can cause the system to easily misjudge normal but drastic changes in operating conditions as abnormal events, resulting in the problem of misidentification of anomalies caused by transient changes in energy consumption during the energy consumption monitoring process. Summary of the Invention

[0007] To address the technical problem of misidentification of anomalies caused by transient changes in energy consumption during energy consumption monitoring in existing technologies, this invention provides a method and system for monitoring property energy consumption anomalies based on big data analysis. The technical solution is as follows:

[0008] On the one hand, a method for monitoring abnormal energy consumption in properties based on big data analysis is provided. This method includes: S1. Obtaining energy consumption power sets for designated property service areas within each specified energy consumption monitoring period, and comparing them with energy consumption prediction intervals. The energy consumption power sets are used to quantify the real-time energy consumption and dynamic trends of the designated property service areas within the corresponding energy consumption monitoring period. The energy consumption prediction interval comparison is used to determine whether the instantaneous energy consumption peak is within the energy consumption prediction interval, thereby screening for energy consumption anomalies within the designated property service areas; S2. Based on the results of the energy consumption prediction interval comparison, repeatedly verifying the abnormal energy consumption data corresponding to the abnormal energy consumption periods monitored by the smart property integrated service management platform in each energy consumption monitoring period, in order to eliminate anomalies caused by the smart property management platform. The energy consumption anomaly data refers to the abnormal data that is determined after comparison with the prediction interval within the corresponding energy consumption monitoring period, which is not in the corresponding energy consumption prediction interval. The repeated verification includes peak anomaly detection and review verification. The abnormal energy consumption period includes the abnormal energy consumption period to be detected and the abnormal energy consumption period to be verified. S3. After completing the repeated verification, the abnormal energy consumption feature set is combined with the association verification. The abnormal energy consumption feature set is used to visualize the multi-dimensional features of the energy consumption anomaly of the corresponding property in the specified property service area. The association verification is used to input the energy consumption anomaly data obtained after review verification into the smart property integrated service management platform for matching analysis.

[0009] On the other hand, a property energy consumption anomaly monitoring system based on big data analysis is provided. This system includes: a property energy consumption monitoring and prediction module, an energy consumption anomaly data processing module, and an anomaly energy consumption correlation verification and update module. The property energy consumption monitoring and prediction module is used to acquire the energy power groups of a specified property service area in each specified energy consumption monitoring period and compare the energy consumption prediction intervals. The energy consumption anomaly data processing module is used to repeatedly check the energy consumption anomaly data corresponding to the anomaly energy consumption period in each energy consumption monitoring period based on the results of the energy consumption prediction interval comparison, so as to eliminate transient normal fluctuations caused by equipment normal operating condition switching or sudden changes in the external environment. The anomaly energy consumption correlation verification and update module is used to perform correlation verification in combination with the anomaly energy consumption feature set after completing the repeated verification.

[0010] The beneficial effects of the technical solutions provided by the embodiments of the present invention include at least the following:

[0011] 1. Within a specified and uniformly long energy consumption monitoring period, acquire energy consumption power sets for designated property service areas. By comparing these sets with energy consumption prediction intervals, screen for energy consumption anomalies. These power sets directly reflect the energy consumption status of the area. Based on the comparison results, repeatedly verify the abnormal energy consumption data during abnormal energy consumption periods to eliminate transient fluctuations caused by equipment switching between normal operating conditions or sudden changes in the external environment. Repeated verification, including peak anomaly detection and review, reduces the false detection rate. After review and verification, perform correlation verification using an abnormal energy consumption feature set. This set visualizes the multi-dimensional characteristics of property energy consumption anomalies, confirming anomalies from multiple dimensions, improving the accuracy and reliability of anomaly identification, providing a scientific basis for property energy consumption management, and contributing to energy conservation, cost reduction, and efficiency optimization.

[0012] 2. By using various types of high-frequency acquisition terminals, energy consumption data in designated property service areas is collected in real time to obtain continuous energy consumption-power curves. This allows for feedback and control of energy consumption within the property management service area, reducing the monitoring gaps caused by sensor malfunctions or interference. It also prevents insensitivity to anomalies due to missing energy consumption monitoring data. During energy consumption data monitoring, the range of energy consumption prediction intervals is defined based on historical energy consumption data from the same time period, maximizing the acceptable range of energy consumption data. This dynamic range avoids limiting the power curve to only small fluctuations over time. In a coordinate system with power as the vertical axis and time as the horizontal axis, each time point has a corresponding prediction range, preventing errors in overall anomaly assessment caused by deviations in a single prediction value.

[0013] 3. By using the maximum and minimum values ​​of the energy consumption prediction interval, the time periods corresponding to the currently acquired energy consumption data are classified into abnormal energy consumption periods to be verified and abnormal energy consumption periods to be detected. In the abnormal energy consumption periods to be verified, interpolation of peak values ​​and neighborhoods is used to avoid misjudgment of single-point peak values. The trend distinguishes between real anomalies and normal fluctuations, and the window step size is automatically adjusted according to the acquired deviation to quickly respond to extreme anomalies similar to sudden equipment failures, improve detection timeliness, reduce redundant calculations, and suppress low-risk false alarms. In the abnormal energy consumption periods to be detected, the total energy consumption area of ​​the period to be verified is integrated to verify the rationality of the period from the perspective of overall energy consumption, reduce the impact of local noise or single-point errors on the judgment, and more comprehensively assess the energy consumption status. In the process of wavelet denoising, the time domain resolution is improved to filter noise interference and avoid misjudgment.

[0014] 4. By verifying the overlap of multiple features, the abnormal energy consumption period is overlapped with the abnormal energy consumption feature set to identify change events within the same period, avoiding the one-sidedness of a single data source; in the feature comparison and quantification of anomalies, the balance between the degree of anomaly and the efficiency of instruction response is quantified to avoid misjudgment due to the deviation of a single indicator; in the anomaly handling, resource allocation is optimized in a hierarchical manner, and operation and maintenance resources are allocated according to the anomaly level to improve the response speed of problems.

[0015] 5. By combining equipment operating status, energy consumption data, and operation logs, the system accurately locates anomalies such as single-device failures, multi-device coupling anomalies, or environmental interference, avoiding misjudgments caused by isolated data. Through a three-level anomaly classification—equipment-level, system-level, and environment-level—the system clarifies the scope of impact and priority of handling anomalies: Level 1 anomalies with minor impact require priority repair; Level 2 anomalies involving multiple devices require system adjustments; and Level 3 anomalies related to the environment do not require equipment intervention, optimizing the allocation of operation and maintenance resources and reducing blind troubleshooting. By feeding the classification results back to the smart property management platform to update the reference set and predictions, the system achieves self-learning, improving its ability to identify similar anomalies in the future, thereby enhancing the precision and efficiency of energy consumption management. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart illustrating a property energy consumption anomaly monitoring method based on big data analysis provided in an embodiment of the present invention;

[0018] Figure 2 A flowchart for property energy consumption monitoring and prediction and partial energy consumption anomaly data processing provided in an embodiment of the present invention;

[0019] Figure 3 A flowchart for the verification and validation of energy consumption anomaly data processing provided in this embodiment of the invention;

[0020] Figure 4 A flowchart for abnormal energy consumption correlation verification and updating provided in an embodiment of the present invention;

[0021] Figure 5 This is a schematic diagram of the structure of a property energy consumption anomaly monitoring system based on big data analysis provided in an embodiment of the present invention. Detailed Implementation

[0022] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0023] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0024] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0025] This invention provides a method for monitoring abnormal energy consumption in properties based on big data analysis, such as... Figure 1 The flowchart shown is for a property energy consumption anomaly monitoring method based on big data analysis. The processing flow of this method may include the following steps:

[0026] S1. Within each specified energy consumption monitoring period, the duration of each monitoring period is the same. Energy consumption power groups for the designated property service area are obtained and energy consumption prediction intervals are compared. The energy consumption power groups are used to quantify the real-time energy consumption and dynamic trend of the designated property service area within the corresponding energy consumption monitoring period, and to intuitively reflect the energy output status of public facilities such as elevators, central air conditioning and residents / merchants in the area. The energy consumption prediction interval comparison is used to determine whether the instantaneous energy consumption peak is within the energy consumption prediction interval, thereby screening for energy consumption anomalies in the designated property service area.

[0027] S2. Based on the results of the energy consumption prediction interval comparison, the abnormal energy consumption data corresponding to the abnormal energy consumption periods in each energy consumption monitoring period are repeatedly checked to exclude transient normal fluctuations caused by equipment normal operating condition switching such as elevator start-up, motor start-up and shutdown, or sudden changes in the external environment such as a sharp rise in temperature or a sudden strong wind. Abnormal energy consumption data refers to the abnormal data that is determined after the prediction interval comparison within the corresponding energy consumption monitoring period and is not in the corresponding energy consumption prediction interval, including the amplitude of transient rise / fall and waveform characteristics. Repeated checks include peak abnormality detection and verification to reduce the false detection rate of abnormal energy consumption periods. Abnormal energy consumption periods include abnormal energy consumption periods to be detected and abnormal energy consumption periods to be verified.

[0028] S3. After completing the duplicate verification, the abnormal energy consumption feature set is used for correlation verification. The abnormal energy consumption feature set is used to visualize the multi-dimensional characteristics of the energy consumption anomalies of the corresponding properties in the specified property service area. It usually includes time distribution, equipment correlation, environmental impact factors, etc., and intuitively presents the correlation between abnormal energy consumption and the operating status of public facilities such as elevator failure, abnormal start and stop of air conditioning, energy consumption behavior of residents / merchants such as the use of high-power equipment during non-business hours, and environmental parameters such as extreme weather. The correlation verification is used to input the energy consumption anomaly data obtained after verification into the smart property integrated service management platform for matching analysis, confirm the causes of anomalies such as equipment failure, pipeline leakage, illegal energy use, etc., and generate a comprehensive report including the anomaly level, impact range and handling suggestions. The report is pushed to the platform at the same time to trigger subsequent automated response mechanisms such as dispatching repairs, remote control or manual intervention processes.

[0029] In this embodiment, the smart property integrated service management platform is a platform for property management and grassroots governance based on big data and other technologies. It is an urban management platform for numerous property services. It comprehensively applies technologies such as the Internet of Things, big data, and artificial intelligence, and integrates with various sensors to monitor, analyze, and process data such as temperature, humidity, light, and air quality in the living environment. Through real-time intervention, it comprehensively improves the quality of life.

[0030] Property services cover a wide range, including public buildings such as sports centers, public parks and high-speed railways, academies, hospitals and colleges, financial institutions such as banks, energy office buildings and offices, and residential communities. In terms of public services, smart property integrated service management is unavoidable. The smart property integrated service management platform includes building management, equipment management, and has data statistics, analysis and real-time monitoring and early warning.

[0031] This method effectively improves the accuracy and efficiency of detecting abnormal energy consumption in properties through phased monitoring and multi-dimensional verification. First, it compares energy consumption prediction intervals at fixed time periods to quickly screen for potential anomalies. Then, it repeatedly verifies transient normal fluctuations to reduce false detection rates. Finally, it combines anomaly feature set correlation verification to accurately locate the cause and generate a response plan. It achieves intelligent processing throughout the entire process, from anomaly identification to cause analysis and response triggering, reducing labor costs, promptly detecting equipment failures and illegal energy use, contributing to energy conservation and efficient property management, and improving service response speed and management precision.

[0032] like Figure 2 The diagram shows a flowchart of property energy consumption monitoring and prediction and partial energy consumption anomaly data processing provided in an embodiment of the present invention. Instantaneous energy consumption peak values ​​are obtained and compared with energy consumption intervals. When the peak value is greater than the maximum value of the energy consumption prediction interval, peak anomaly detection is performed. The time window step size is determined by whether the obtained energy consumption fluctuation intensity deviation is greater than the historical maximum value. If the energy consumption is within the energy consumption prediction interval, there is no abnormal energy consumption period at this time. When the energy consumption is less than the minimum value of the energy consumption interval, a verification is performed.

[0033] Furthermore, within each defined energy consumption monitoring period, there exists an energy consumption-power curve. Specifically, energy consumption data within a designated property service area is collected in real time using various types of high-frequency acquisition terminals (such as power transformers, smart water meters, gas flow meters, and equipment condition sensors). This forms a continuous energy consumption-power curve with seconds as the horizontal axis and corresponding energy consumption power values ​​as the vertical axis. Each data point on the continuous energy consumption-power curve contains structured information, including timestamps, energy consumption types, instantaneous power, and associated device IDs. The complete curve within a single energy consumption monitoring period of equal duration on the continuous energy consumption-power curve is defined as an energy consumption power group (i.e., an ordered set of all energy consumption data within that period). The structured information of each data point on the continuous energy consumption-power curve is called energy consumption data.

[0034] The specific analysis method for energy consumption prediction interval comparison is as follows: Obtain the instantaneous energy consumption peak value corresponding to each energy consumption power group; determine the matching relationship between the instantaneous energy consumption peak value and the maximum and minimum values ​​of the energy consumption prediction interval: Q1, if the obtained instantaneous energy consumption peak value is greater than the maximum value of the energy consumption prediction interval, then the time period of the energy consumption power group corresponding to the instantaneous energy consumption peak value is recorded as the abnormal energy consumption period to be detected, and peak anomaly detection is performed; Q2, if the obtained instantaneous energy consumption peak value is greater than the minimum value of the energy consumption prediction interval but not greater than the maximum value of the energy consumption prediction interval, then the time period of the energy consumption power group corresponding to the instantaneous energy consumption peak value is recorded as the period with no abnormal energy consumption.

[0035] In this embodiment, the instantaneous energy consumption peak of each time period's energy consumption power group is quickly matched with a preset energy consumption prediction interval. When the peak exceeds the maximum value of the interval, the time period is automatically marked as an abnormal energy consumption period to be detected and peak anomaly detection is triggered. This ensures that transient normal fluctuations such as elevator start-stop will not interfere with the initial screening, while capturing potential real anomalies. When the peak is between the minimum and maximum values ​​of the prediction interval, it is determined to be a time period without abnormal energy consumption, avoiding misjudgments caused by reasonable energy consumption fluctuations such as gradual changes in air conditioning load. Combined with the structured energy consumption-power curve data formed by high-frequency acquisition at the second level, automated initial screening of abnormal time periods is achieved, reducing the amount of invalid detections and allowing subsequent repeated inspection resources to focus on highly suspicious time periods. This not only ensures the rapid location of real anomalies such as equipment failure or illegal energy use, but also reduces the false alarm rate and computational load of the overall system.

[0036] Furthermore, the specific steps for peak detection are as follows: obtain the instantaneous energy consumption peak value within the period of abnormal energy consumption to be detected, and obtain the average energy consumption fluctuation value according to the set neighborhood time period. The instantaneous energy consumption peak value represents the highest energy consumption power value appearing in the energy consumption power group within the period of abnormal energy consumption to be detected. This value is the instantaneous maximum value of energy consumption power within this period, which can intuitively reflect the extreme energy consumption state at a certain moment (such as the power spike when the equipment suddenly starts). The set neighborhood time period represents the continuous and equal-length neighborhood time intervals defined before and after the occurrence of the peak value, based on the time point corresponding to the instantaneous energy consumption peak value. The neighborhood time interval includes the left neighborhood and the right neighborhood. For example, 10 seconds before the peak value to the peak value is the left neighborhood, and 10 seconds after the peak value is the right neighborhood. This is used to capture the energy consumption change trend before and after the instantaneous energy consumption peak value. The average energy consumption fluctuation value represents the average value of the difference between the average energy consumption power of the left neighborhood and the instantaneous energy consumption peak value (left fluctuation value) and the difference between the average energy consumption power of the right neighborhood and the instantaneous energy consumption peak value.

[0037] The sum of the instantaneous peak energy consumption and the average energy consumption fluctuation value is recorded as the energy consumption fluctuation intensity index. The energy consumption fluctuation intensity index is used to quantify the degree of instantaneous power fluctuation in a specified property service area during abnormal energy consumption periods. Based on the energy consumption fluctuation intensity index deviation and the mapping relationship between the historical energy consumption fluctuation and window step size adjustment values, the actual time window step size adjustment value is obtained to dynamically adjust the time window step size. If the energy consumption fluctuation intensity index is greater than the maximum value of the historical energy consumption fluctuation intensity index, the actual window step size adjustment value is recorded as the time window step size increase value, increasing the time window step size to improve the response rate to instantaneous energy consumption peaks. If the energy consumption fluctuation intensity index is greater than the reference energy consumption fluctuation intensity index but not greater than the maximum value of the historical energy consumption fluctuation intensity index, the actual window step size adjustment value is recorded as the time window step size decrease value, decreasing the time window step size to reduce the energy consumption fluctuation calculation workload and suppress false alarms. During the time window step size adjustment process, the associated operation log of the time period corresponding to the instantaneous energy consumption peak is retrieved. If the operating power of the specified operating equipment recorded in the associated operation log is greater than the corresponding set rated power, a power over-limit warning is sent; otherwise, the window step size adjustment is completed.

[0038] In this embodiment, the energy consumption fluctuation-window step size mapping relationship is based on the XGBoost model. The mapping relationship between energy consumption fluctuation and window step size is obtained by training an algorithm based on the obtained historical energy consumption fluctuation intensity index deviation and historical energy consumption fluctuation-window step size adjustment value, using a precise greedy criterion. Inputting the currently obtained energy consumption fluctuation intensity index deviation into this mapping relationship yields the corresponding window step size adjustment value. The set domain time period is represented by the summation and averaging of historically selected domain time periods during the historical peak anomaly detection process. In practical applications, fine-tuning can be performed on this basis. The energy consumption fluctuation intensity index deviation represents the difference between the obtained energy consumption fluctuation intensity and the set energy consumption fluctuation intensity. The set energy consumption fluctuation intensity is represented by the summation and averaging of historical energy consumption fluctuation intensities during the historical peak anomaly detection process.

[0039] An energy consumption fluctuation intensity index is generated by calculating the instantaneous peak energy consumption and the average energy consumption fluctuation value of the neighborhood, which quantifies the degree of power fluctuation during abnormal periods. Based on the mapping relationship between this index and historical fluctuation data, the time window step size is dynamically adjusted: when the index exceeds the historical maximum value, the step size is increased to improve the capture speed of extreme peaks such as sudden elevator starts; when the index is between the reference value and the historical maximum value, the step size is decreased to reduce the amount of invalid calculation under stable operating conditions. This adaptive step size adjustment avoids over-response to transient normal fluctuations and ensures rapid detection of real anomalies such as sudden increases in equipment power. Simultaneously, a rated power verification mechanism based on associated operation logs is used to trigger a power over-limit warning when the equipment actually exceeds the limit.

[0040] like Figure 3The flowchart shown is a verification process for energy consumption anomaly data processing provided in this embodiment of the invention. It determines whether interpolation processing exists. If so, interpolation is performed to supplement the data and verify whether there is missing data. If missing data exists, an energy consumption log missing warning is issued. If not, an energy consumption area comparison is performed. The energy consumption area to be verified is compared with a reference value. If the value is greater than the reference value, the operating status is qualified. If the value is less than the reference value, the feature extraction frequency needs to be adjusted. It is then determined whether the newly obtained energy consumption area to be verified is greater than the reference value. If so, the data is qualified and correlation verification is performed. Otherwise, an abnormal energy consumption warning is sent.

[0041] Furthermore, if the obtained instantaneous energy consumption peak value is not greater than the minimum value of the energy consumption prediction interval, the time period corresponding to the energy consumption power group of the instantaneous energy consumption peak value is recorded as the abnormal energy consumption period to be verified and is then checked. The number of verification checks is determined based on the degree of deviation between the instantaneous energy consumption peak value and the energy consumption prediction interval. Specifically, based on the degree of deviation, the value is input into the reference mapping set to obtain the adjustment value of the maximum number of verification checks. The adjustment value is used to quantify the degree of increase of the maximum number of verification checks. If the instantaneous energy consumption peak value status is not confirmed when the number of verification checks exceeds the maximum number of verification checks, an energy consumption data anomaly warning information is generated to trigger a further manual verification process. At the same time, the energy consumption data, degree of deviation, verification records, and other information of the abnormal period are automatically stored in the energy consumption anomaly database to provide data support for subsequent energy consumption pattern analysis and anomaly cause tracing. The warning information is also pushed to the terminal devices of property management personnel to ensure that relevant personnel are aware of the situation and can intervene in a timely manner.

[0042] The verification process includes interpolation adjustment and energy consumption area comparison. Interpolation adjustment is used to handle missing data in the energy consumption log, while energy consumption area comparison is used to verify whether the operating status during abnormal energy consumption periods is qualified. Specifically, interpolation adjustment involves reading the energy consumption log of a specified property service area. If missing data exists, interpolation is performed to supplement it. Interpolation supplementation means interpolating based on the energy power groups corresponding to the left and right neighboring domains to obtain the reading time length of the corresponding energy consumption log after interpolation supplementation. If no missing data exists, energy consumption area comparison is performed. After completing the interpolation adjustment, the energy consumption log of the specified property service area is reread based on the obtained reading time length. If missing data exists, an energy consumption log missing warning is issued; otherwise, energy consumption area comparison is performed.

[0043] The specific process of energy consumption area comparison is as follows: Obtain the energy consumption-power curve (i.e., the power change over time) corresponding to the abnormal energy consumption period to be verified; perform integration processing on the abnormal energy consumption period to be verified to obtain the energy consumption area to be verified (quantitative data representing the total power during that period); compare the obtained energy consumption area to be verified with the reference energy consumption area to be verified; if the energy consumption area to be verified is greater than the reference energy consumption area to be verified, the operating status of the abnormal energy consumption period to be verified is determined to be qualified; otherwise, based on the deviation of the energy consumption area to be verified, input it into the historical area deviation-feature extraction frequency mapping relationship to obtain the increase adjustment value of the feature extraction frequency in the denoising process, which is used to prompt the feature extraction frequency to increase to this adjustment value to improve the temporal resolution of the signal and eliminate misjudgment of anomaly identification; after the feature extraction frequency is adjusted, perform correlation verification, and generate suggested correction measures instructions based on the deviation of the energy consumption area to be verified according to the correction measures stored in the historical reference mapping, which is used to prompt the preset personnel to take corresponding correction measures, such as reading the code fault in the operation log and prioritizing the investigation of components not covered in the most recent maintenance record.

[0044] In this embodiment, the deviation of the energy consumption area to be verified represents the difference between the acquired energy consumption area to be verified and the reference energy consumption area to be verified. The area deviation-feature extraction frequency mapping relationship is based on the LightGBM algorithm. By acquiring the historical energy consumption area deviation and historical feature extraction frequency adjustment values, the mapping relationship is obtained by inputting them into the recurrent neural network model for training. The current acquired energy consumption area deviation to be verified is input into this mapping relationship, and the corresponding feature extraction frequency can be output. The reference energy consumption area to be verified is represented by the summation and averaging of the historical energy consumption areas to be verified during the historical review and verification process.

[0045] When the instantaneous energy consumption peak value is lower than the minimum value of the prediction interval, the system automatically marks it as an abnormal energy consumption period to be verified. Based on the deviation of the instantaneous energy consumption peak value from the energy consumption prediction interval, the number of verifications is adjusted. If there is a decimal point, it is rounded up. For example, if the current deviation is 40% and the number of verifications is 4, then the number of verifications is determined as 4. (1+40%)=5.6, meaning the number of reviews is adjusted to 6, avoiding excessive consumption of data for minor fluctuations while ensuring in-depth verification of suspicious anomalies. Interpolation processing adjusts and repairs missing data in the energy consumption log, using neighborhood energy consumption power groups to intelligently fill in null values, ensuring the integrity of subsequent analysis; energy consumption area comparison quantifies the total energy consumption over a period of time through integration, and if it fails the first comparison with the reference value, it triggers adaptive adjustment of the feature extraction frequency to improve the time domain resolution and eliminate noise interference; if the second review still fails, the system combines the historical correction measures library to generate precise handling instructions, such as troubleshooting water pump controller faults.

[0046] like Figure 4As shown in the flowchart of abnormal energy consumption association verification and update provided in this embodiment of the invention, the increase of the event processing thread is changed by obtaining the deviation of the abnormal duration score. At the same time, the abnormality is classified according to its source, and the results of classification and association verification are input into the smart property integrated service management platform for update. The sources are divided into first level, second level and third level, and warnings are issued for first level energy consumption abnormality, second level energy consumption abnormality and third level energy consumption abnormality according to the level, and the preset personnel are prompted to take the next step.

[0047] Furthermore, the association verification is performed by combining the abnormal energy consumption feature set. The specific steps are as follows: overlap the abnormal energy consumption period to be verified with the abnormal energy consumption feature set, identify the change events in the abnormal energy consumption feature set, and perform feature comparison based on the abnormal energy consumption feature set in the reference mapping set. Specifically, this includes: obtaining the abnormal duration of the abnormal energy consumption period to be verified, comparing it with the length of the time interval corresponding to the abnormal energy consumption period to be verified, and obtaining the abnormal duration score. The abnormal duration score is used to quantify the persistence of the abnormal state and represents the ratio of the obtained abnormal duration to the length of the time interval.

[0048] The abnormal duration is defined as follows: if, within the time interval corresponding to the abnormal energy consumption period to be verified, the energy consumption power is within the predicted energy consumption range, the energy consumption power value is changed from not being within the predicted energy consumption range to being within the predicted energy consumption range, and this time interval is taken as the abnormal duration. If, within this time interval, the energy consumption power is not within the predicted energy consumption range, the end time corresponding to the abnormal energy consumption period is taken as the abnormal duration. The obtained abnormal duration score deviation is input into the historical duration-processing thread mapping relationship to obtain the increase value of the processing thread for the change event. This is used to prompt the thread for processing the change time to be increased to the obtained increase value of the processing thread for the change event, thereby accelerating the feature comparison speed and classifying the abnormal source. The abnormal source classification is used to classify the energy consumption anomaly level.

[0049] The duration-processing thread mapping relationship is based on a supervised learning algorithm using support vector machines. By obtaining the historical anomaly duration score deviation and the processing thread adjustment value of historical change events, the input values ​​are fed into the neural network model for training to obtain the duration-processing thread mapping relationship. By inputting the currently obtained anomaly duration score deviation into this mapping relationship, the corresponding processing thread adjustment value can be output.

[0050] The specific process for classifying anomaly sources is as follows: Anomalies are located by analyzing the operating status and energy consumption anomaly data of specified devices. If a single device or a group of similar devices experiences energy consumption anomalies, the corresponding operating logs are read. If no relevant maintenance records are found, it is determined to be a device malfunction. Such malfunctions are limited and can be isolated, with a small impact range, and a Level 1 energy consumption anomaly command is sent. If the operating status and energy consumption of two or more types of devices rise and fall simultaneously during the same time period, and all of these devices' operating logs show Level 1 energy consumption anomalies, it indicates that the coupling effect of device collaborative operation has triggered an overall anomaly, i.e., a system malfunction. This malfunction spans multiple devices. If the impact is large, a second-level energy consumption anomaly command is sent. If the operating status and energy consumption of multiple devices and systems show periodic fluctuations, and the energy consumption recovers after the periodic fluctuations end, it is determined to be an environmental anomaly caused by changes in the external environment. This anomaly is periodic and unrelated to the operating status of the devices, and a third-level energy consumption anomaly command is sent. After the anomaly source is classified, the energy consumption anomaly data, corresponding time period, abnormal energy consumption feature set data, data source and related feature classification results are input into the smart property integrated service management platform to update the reference mapping set and energy consumption prediction interval for subsequent property energy consumption anomaly monitoring.

[0051] In this embodiment, the reference map set stores abnormal energy consumption data and their corresponding correction measures, and stores a set of abnormal energy consumption features. This is used to trace the source of the abnormality during the abnormal energy consumption identification process, and to generate corresponding processing measures based on the deviation between the obtained peak value and the prediction range. These processing measures and correction measures are preset by relevant personnel and can be modified in actual use to output corresponding suggested correction instructions, rather than direct instructions. The update of the reference map set is based on the time recurrent neural network of the long short-term memory network, which integrates historical energy consumption data, abnormal features and environmental variables to construct multivariate prediction. Through a feedback-driven algorithm, the abnormal energy consumption data and suggested correction instructions involved in the energy consumption anomaly monitoring process are used as sample data and input into the long short-term memory network for updating.

[0052] The abnormal duration score deviation represents the difference between the acquired abnormal duration score and the set abnormal duration score. The abnormal duration is acquired through the system's built-in timer, and the set abnormal duration score is represented by the average of historical abnormal duration scores obtained during the historical feature comparison process. During the abnormality source classification process, different suggested handling methods are sent based on the level of energy consumption anomaly. For example, in the first-level energy consumption anomaly, which is currently judged to be caused by equipment failure, the following output can be provided: "Refer to maintenance records, elevator main motor power exceeds limit, use a clamp meter to test the driver." In the second-level energy consumption anomaly, where the system itself is abnormal, the following output can be provided: "Air conditioning group is abnormal, chilled water pump energy consumption and cooling tower fan energy consumption are both increased, and the operation log shows abnormal valve opening during the same period." In this case, a reset of the main unit outlet water temperature setpoint should be generated, and the cooling tower fan frequency should be reduced simultaneously. In the third-level energy consumption anomaly, if there is extreme hot and humid weather, the load increases, and the energy consumption of multiple dehumidifiers periodically increases. Combined with weather station warning information, the following output can be provided: "Start the dehumidifier one hour in advance to reduce the fresh air load."

[0053] It should be noted that the examples in these abnormal energy consumption classifications are only to provide one application scenario. In actual application scenarios, there are more than just these examples, and personnel can adjust the specific output suggestions according to the actual scenario.

[0054] The system dynamically expands processing thread resources based on anomaly duration scores, accelerating the feature comparison process and ensuring efficient differentiated handling of sudden short-term anomalies and persistent anomalies. A three-level anomaly source classification system enables efficient localization: single-device faults are identified through device operation logs, system-level coupled faults are identified based on cross-device state synergy, and environmental factor anomalies are judged by periodic fluctuation characteristics, constructing a hierarchical diagnostic network from local to global. This design simultaneously triggers three deep optimizations: anomaly duration improves system resource allocation efficiency, the three-level classification mechanism improves fault location accuracy, and in the closed-loop feedback loop, continuous adaptive evolution is driven by real-time updates to the reference mapping set and energy consumption prediction interval.

[0055] like Figure 5The diagram shows the structure of a property energy consumption anomaly monitoring system based on big data analysis provided in this embodiment of the invention. This system includes the following modules: a property energy consumption monitoring and prediction module, an energy consumption anomaly data processing module, and an anomaly energy consumption association verification and update module. The property energy consumption monitoring and prediction module acquires energy consumption power groups for a specified property service area within each specified energy consumption monitoring period and performs energy consumption prediction interval comparisons. The energy consumption anomaly data processing module performs repeated checks on the energy consumption anomaly data corresponding to the anomaly energy consumption periods in each energy consumption monitoring period based on the results of the energy consumption prediction interval comparisons. This eliminates transient normal fluctuations caused by equipment switching or sudden changes in the external environment, reducing the false judgment rate in the initial screening. The anomaly energy consumption association verification and update module performs association verification by combining the anomaly energy consumption feature set after completing the repeated checks.

[0056] In this embodiment, the property energy consumption monitoring and prediction module constitutes the system perception layer. It acquires energy consumption power groups for each time period in real time through a high-frequency acquisition terminal. The comparison results of the energy consumption prediction intervals output by the module directly drive the operation of the energy consumption anomaly data processing module. Based on the comparison results, this module initiates a differentiated processing procedure: it performs peak anomaly detection and verification for the period to be detected to filter transient interference, performs interpolation supplementation and energy consumption area comparison for the period to be verified to verify authenticity, and pushes the confirmed abnormal data and features to the abnormal energy consumption association verification and update module. The latter, as the decision-making center, integrates the abnormal energy consumption feature set to perform three-level source classification and cause diagnosis, and simultaneously triggers disposal instructions and generates a comprehensive report. The three modules form a closed loop of "monitoring screening → deep cleaning → intelligent diagnosis → dynamic optimization" through data flow, instruction flow, and feedback flow, realizing a synergistic leap in anomaly identification accuracy, disposal response speed, and system adaptive capability, and ultimately building a continuously evolving property energy consumption governance ecosystem.

[0057] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the flow or function according to the embodiments of the present invention is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. A computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

[0058] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0059] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0060] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0061] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0062] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0063] The technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0064] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for monitoring abnormal energy consumption in properties based on big data analysis, characterized in that, Includes the following steps: S1. Within each energy consumption monitoring period specified by the smart property integrated service management platform, acquire the energy consumption power group of the designated property service area and compare it with the energy consumption prediction interval of the smart property integrated service management platform. The energy consumption power group is used to quantify the real-time energy consumption and dynamic change trend of the designated property service area within the corresponding energy consumption monitoring period. The comparison with the energy consumption prediction interval is specifically analyzed as follows: acquire the instantaneous energy consumption peak value corresponding to each energy consumption power group, and determine the matching relationship between the instantaneous energy consumption peak value and the maximum and minimum values ​​of the energy consumption prediction interval. Q1, if the instantaneous energy consumption peak value is greater than the maximum value of the energy consumption prediction interval, then the time period of the energy consumption power group corresponding to the instantaneous energy consumption peak value is recorded as the abnormal energy consumption period to be detected. Q2, if the instantaneous energy consumption peak value is greater than the minimum value of the energy consumption prediction interval but not greater than the maximum value of the energy consumption prediction interval, then the time period of the energy consumption power group corresponding to the instantaneous energy consumption peak value is recorded as the period with no abnormal energy consumption. Q3, if the instantaneous energy consumption peak value is not greater than the minimum value of the energy consumption prediction interval, then the time period of the energy consumption power group corresponding to the instantaneous energy consumption peak value is recorded as the abnormal energy consumption period to be verified. S2. Based on the comparison results of energy consumption prediction intervals, peak anomaly detection is performed on the energy consumption anomaly data corresponding to the abnormal energy consumption periods to be detected by the smart property integrated service management platform in each energy consumption monitoring period. The energy consumption anomaly data corresponding to the abnormal energy consumption periods to be verified is reviewed and verified to eliminate transient normal fluctuations caused by the switching of normal operating conditions of equipment within the monitoring range of the smart property integrated service management platform or sudden changes in the external environment. S3. After completing the review and verification, the abnormal energy consumption data of the period to be verified is correlated and verified by combining the abnormal energy consumption feature set, and the results of the correlation verification are uploaded to the smart property integrated service management platform. The abnormal energy consumption feature set is used to visualize the multi-dimensional features of the abnormal energy consumption of the property corresponding to the specified property service area. The correlation verification is used to input the abnormal energy consumption data obtained after the review and verification into the smart property integrated service management platform for matching analysis. The specific steps for peak anomaly detection are as follows: The instantaneous energy consumption peak value is obtained within the period of abnormal energy consumption to be detected, and the average energy consumption fluctuation value is obtained according to the set neighborhood time period. The instantaneous energy consumption peak value represents the highest energy consumption power value that appears in the energy consumption power group within the period of abnormal energy consumption to be detected. The set neighborhood time period represents a continuous and equal-length neighborhood time interval defined before and after the occurrence of the peak value, with the time point corresponding to the instantaneous energy consumption peak value as the reference. The neighborhood time interval includes a left neighborhood and a right neighborhood. The average energy consumption fluctuation value represents the average value between the difference between the average energy consumption power of the left neighborhood and the instantaneous energy consumption peak value, and the difference between the average energy consumption power of the right neighborhood and the instantaneous energy consumption peak value. The sum of the instantaneous peak energy consumption and the average energy consumption fluctuation value is recorded as the energy consumption fluctuation intensity index. The energy consumption fluctuation intensity index is used to quantify the degree of instantaneous power fluctuation in a specified property service area during abnormal energy consumption periods. The peak anomaly detection also includes: Based on the energy consumption fluctuation intensity index deviation and the energy consumption fluctuation-window step size mapping relationship between the historical energy consumption fluctuation-window step size adjustment value, the actual time window step size adjustment value is obtained. If the energy consumption fluctuation intensity index is greater than the maximum value of the historical energy consumption fluctuation intensity index, the actual time window step size adjustment value is recorded as the time window step size increase value. The time window step size is gradually increased to improve the response rate to instantaneous energy consumption peaks. If the energy consumption fluctuation intensity index is greater than the reference energy consumption fluctuation intensity index but not greater than the maximum value of the historical energy consumption fluctuation intensity index, then the actual time window step size adjustment value is recorded as the time window step size reduction value. The time window step size is gradually reduced to reduce the amount of energy consumption fluctuation calculation and suppress false alarms. The energy consumption fluctuation intensity index deviation represents the difference between the obtained energy consumption fluctuation intensity index and the set energy consumption fluctuation intensity index. The set energy consumption fluctuation intensity is represented by the result of summing and averaging the historical energy consumption fluctuation intensities during the historical peak anomaly detection process. During the time window step adjustment process, retrieve the associated operation log for the time period corresponding to the instantaneous energy consumption peak. If the operating power of the specified operating device recorded in the associated operation log is greater than the corresponding set rated power, send a power over-limit warning; otherwise, complete the window step adjustment. The verification includes interpolation adjustment and energy consumption area comparison; The energy consumption area comparison process is as follows: Obtain the energy consumption-power curve corresponding to the abnormal energy consumption period to be verified, and perform integration processing on the abnormal energy consumption period to be verified to obtain the energy consumption area to be verified. The obtained energy consumption area to be verified is compared with the reference energy consumption area to be verified. If the energy consumption area to be verified is greater than the reference energy consumption area to be verified, the operating status of the abnormal energy consumption period to be verified is determined to be qualified. Otherwise, the feature extraction frequency in the denoising process is adjusted based on the area deviation-feature extraction frequency mapping relationship through the energy consumption area deviation to be verified, so as to improve the temporal resolution of the signal and eliminate the misjudgment of abnormal identification. The energy consumption area deviation to be verified represents the difference between the obtained energy consumption area to be verified and the reference energy consumption area to be verified. The reference energy consumption area to be verified is represented by the result of the summation and averaging of the historical energy consumption areas to be verified in the historical review and verification process.

2. The method for monitoring abnormal energy consumption in properties based on big data analysis as described in claim 1, characterized in that, The specific steps for obtaining the energy consumption power groups of designated property service areas are as follows: Within each defined energy consumption monitoring period, there exists an energy consumption-power curve, specifically as follows: Through multiple types of high-frequency acquisition terminals, energy consumption data in the designated property service area is collected in real time, forming a continuous energy consumption-power curve with second-level time as the horizontal axis and corresponding energy consumption power value as the vertical axis. Each data point on the continuous energy consumption-power curve contains structured information. The complete curve within a single energy consumption monitoring period of equal duration on the continuous energy consumption-power curve is defined as an energy consumption-power group, and the structured information of each data point on the continuous energy consumption-power curve is called energy consumption data.

3. The method for monitoring abnormal energy consumption in properties based on big data analysis as described in claim 1, characterized in that, The comparison of the energy consumption prediction intervals also includes: The number of verification cycles is determined based on the degree of deviation between the instantaneous energy consumption peak and the energy consumption prediction range, specifically as follows: Based on the degree of deviation, the value is input into the reference mapping set to obtain the adjustment value of the maximum number of verifications. The adjustment value is used to quantify the degree of increase of the maximum number of verifications. If the instantaneous energy consumption peak status is not confirmed when the number of verifications exceeds the maximum number of verifications, an energy consumption data anomaly warning message will be generated to trigger a further manual verification process.

4. The method for monitoring abnormal energy consumption in properties based on big data analysis as described in claim 1, characterized in that, The interpolation process is used to process missing data in the energy consumption log, and the energy consumption area comparison is used to verify whether the operating status during abnormal energy consumption periods is qualified. The interpolation processing adjustment specifically involves: reading the energy consumption log of the specified property service area; if there is missing data, interpolation is performed to supplement it; the interpolation supplementation means interpolating based on the energy consumption power groups corresponding to the left and right neighboring domains to obtain the reading time length of the corresponding energy consumption log after interpolation supplementation; if there is no missing data, energy consumption area comparison is performed. After completing the interpolation adjustment, the energy consumption logs of the specified property service area are reread based on the obtained reading time length, and the energy consumption area is compared with the supplementary data obtained after the interpolation adjustment.

5. The method for monitoring abnormal energy consumption in properties based on big data analysis as described in claim 1, characterized in that, The specific steps for correlation verification by combining the abnormal energy consumption feature set are as follows: The abnormal energy consumption period to be verified is overlapped with the abnormal energy consumption feature set to identify change events in the abnormal energy consumption feature set, and feature comparison is performed based on the abnormal energy consumption feature set in the reference mapping set, specifically including: The abnormal duration of the abnormal energy consumption period to be verified is obtained and compared with the length of the time interval corresponding to the abnormal energy consumption period to be verified to obtain the abnormal duration score. The duration of the anomaly is as follows: if, within the time interval corresponding to the period of the anomaly energy consumption to be verified, the energy consumption power is in the energy consumption prediction range, then the time interval corresponding to the energy consumption prediction range is changed from not being in the energy consumption prediction range to being in the energy consumption prediction range, and this is taken as the duration of the anomaly. If there is no energy consumption power within the energy consumption prediction area within this time interval, then the duration corresponding to the abnormal energy consumption period shall be taken as the abnormal duration. The obtained abnormal duration score deviation is input into the duration-processing thread mapping relationship to obtain the increase in the processing thread of the change event, which accelerates the feature comparison speed and performs abnormal source classification. The abnormal source classification is used to classify the energy consumption anomaly level. The abnormal duration score deviation represents the difference between the obtained abnormal duration score and the set abnormal duration score. The set abnormal duration score is represented by the result of summing and averaging the historical abnormal duration scores in the historical feature comparison process.

6. The method for monitoring abnormal energy consumption in properties based on big data analysis as described in claim 5, characterized in that, The specific process for classifying the sources of anomalies is as follows: The first level locates the abnormal location by specifying the operating status of the running equipment and the abnormal energy consumption data. If a single device or a group of similar devices has an abnormal energy consumption, the corresponding operating log is read. If there is no relevant maintenance record, it is determined to be a device failure, and a first-level energy consumption abnormality command is sent. The second level indicates that if the operating status and energy consumption of two or more types of equipment rise or fall at the same time during the same period, and the operating logs of these equipment all show the first level of energy consumption anomaly, it means that the coupling effect of the equipment's coordinated operation has caused the overall anomaly, i.e., the system failure, and a second level of energy consumption anomaly command is sent. At the third level, if the operating status and energy consumption of multiple devices and systems exhibit periodic fluctuations, and the energy consumption recovers after the periodic fluctuations end, it is determined to be an environmental anomaly caused by changes in the external environment, and a third-level energy consumption anomaly command is sent. After the abnormal sources are classified, the abnormal energy consumption data, corresponding time periods, abnormal energy consumption feature set data, data sources and related feature classification results are input into the smart property integrated service management platform to update the reference mapping set and energy consumption prediction interval for subsequent property energy consumption anomaly monitoring.

7. A property energy consumption anomaly monitoring system based on big data analysis, employing the property energy consumption anomaly monitoring method based on big data analysis as described in any one of claims 1-6, comprising: Property energy consumption monitoring and prediction module, energy consumption anomaly data processing module, and anomaly energy consumption correlation verification and update module; The property energy consumption monitoring and prediction module is used to obtain the energy consumption power group of a designated property service area during each energy consumption monitoring period specified by the smart property integrated service management platform, and compare it with the energy consumption prediction interval of the smart property integrated service management platform. The energy consumption anomaly data processing module is used to perform peak anomaly detection on the energy consumption anomaly data corresponding to the energy consumption anomaly period selected by the smart property integrated service management platform in each energy consumption monitoring period based on the results of the energy consumption prediction interval comparison. It also performs verification and validation on the energy consumption anomaly data corresponding to the anomaly energy consumption period to be verified, so as to exclude transient normal fluctuations caused by the normal operation switching of equipment within the monitoring range of the smart property integrated service management platform or sudden changes in the external environment. The abnormal energy consumption association verification and update module is used to perform association verification on the abnormal energy consumption data of the period to be verified after completing the review and verification, in combination with the abnormal energy consumption feature set, and upload the association verification results to the smart property integrated service management platform.

Images