Building digital operation and maintenance collaborative management system based on BIM

Abstract

The invention discloses a BIM-based building digital operation and maintenance collaborative management system, and relates to the technical field of big data analysis. Structural data of a building equipment management part is read, an associated data set is analyzed and constructed, a target sensor is selected, and historical operation data is integrated in combination with a time sequence to construct a fault feature analysis set; the operation data before the fault and the historical maintenance data are further associated, the data are subjected to feature analysis, the fault time is predicted through a neural network model, and a sensor acquisition frequency setting module connected with the output end of the neural network model calculates two types of acquisition frequency coefficients and dynamically adjusts the acquisition frequency through weighted fusion; and finally, a multi-sensor operation and maintenance monitoring module connected with the output end of the sensor acquisition frequency setting module guarantees stable operation of multiple sensors through a sliding window, threshold calculation and abnormal early warning, collaborative management of building digital operation and maintenance is realized, and data association accuracy, fault prediction reliability and operation and maintenance stability are improved.

Description

Technical Field

[0001] This invention relates to the field of big data analytics technology, specifically a BIM-based digital building operation and maintenance collaborative management system. Background Technology

[0002] The operation and maintenance (O&M) phase is the longest and most resource-intensive stage in the entire building lifecycle. Traditional O&M suffers from problems such as fragmented information across multiple systems, poor collaboration, delayed fault response, and high costs. BIM-based digital O&M collaborative management can integrate multi-dimensional O&M information, break down information barriers, enable collaborative operations among all parties and early warning of equipment failures, significantly improve O&M efficiency, extend equipment lifespan, reduce O&M costs, and ensure the safe and stable operation of buildings. It is the core support for the transformation and upgrading of building O&M in the context of smart buildings and is crucial to enhancing the value of building O&M throughout its entire lifecycle.

[0003] During the building operation and maintenance phase, the monitoring of building electromechanical systems relies on the collection of operational data from various types of sensors. However, due to differences in service environment, operating load, and maintenance cycle, the quality of monitoring data from different devices exhibits significant heterogeneity. The existing fixed-frequency static data acquisition mode is not only difficult to adapt to the actual monitoring needs of the equipment, but also easily generates redundant data that consumes BIM resources, interferes with data analysis, and may miss early warning data of high-load equipment failures, affecting the accuracy of operation and maintenance decisions. Therefore, there is an urgent need for a BIM-based digital operation and maintenance collaborative management system for buildings. Summary of the Invention

[0004] The purpose of this invention is to provide a BIM-based digital building operation and maintenance collaborative management system to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a BIM-based building digital operation and maintenance collaborative management system, which includes a target sensor determination module, a fault analysis set construction module, a fault feature analysis set construction module, a fault time prediction module, a sensor acquisition frequency setting module, and a multi-sensor operation and maintenance monitoring module.

[0006] The target sensor determination module is used to acquire structured data of the building equipment management section in the BIM model, and analyze and extract the associated data set consisting of various sensor types, monitored equipment and data acquisition parameters corresponding to the building equipment monitoring requirements.

[0007] The building equipment management section is a specialized management module in the BIM model specifically designed for various types of equipment in service within the building, such as air conditioning, water supply and drainage, and electrical systems. It integrates structured data such as inherent equipment parameters, sensor deployment information, real-time operating data, and historical maintenance records, providing data support and functional carriers for equipment operation status monitoring, fault early warning analysis, and full life cycle operation and maintenance.

[0008] The target sensor determination module includes a structured data reading unit, an associated data set construction unit, and a target sensor selection unit;

[0009] The structured data reading unit is used to read the inherent attribute data of building equipment and sensor deployment configuration data contained in the building equipment management section of the BIM model through the structured data storage path and standard data interface of the building equipment management section, and obtain structured data.

[0010] The associated data set construction unit is used to parse the fields of the read structured data, extract the type identifier of each sensor and the unique code of the monitored device, select the unique code of the monitored device as the key, and select the type identifier of the sensor as the value to construct the associated data set;

[0011] The target sensor selection unit is used to classify the system category to which the monitored device belongs in the associated data set according to the sensor type identifier, and select any sensor from the classified associated data set as the target sensor;

[0012] By relying on BIM models to read structured data, a set of associated data is constructed with the unique code of the monitored equipment as the key and the sensor type identifier as the value. Target sensors are selected by category, realizing efficient association between sensor type, monitored equipment and data acquisition parameters. This results in standardized data, clear association, strong targeting, and accurate matching of monitoring needs.

[0013] The fault analysis set construction module is used to acquire historical data and analyze the monitored equipment and its operating data that the target sensor is in service with. After being arranged in an orderly manner according to the time series, the fault analysis set of the monitored equipment is constructed.

[0014] The fault analysis set construction module includes a historical data retrieval unit, a data time-related arrangement unit, and a fault analysis set construction unit.

[0015] The historical data retrieval unit is used to read the preset data acquisition cycle parameters of the target sensor in the BIM model, retrieve the historical acquisition data generated by the target sensor during its service period from the data storage module based on the data acquisition cycle parameters, and separate the identification information of the monitored equipment and the operating data of the monitored equipment from the retrieved historical acquisition data.

[0016] The data time association and arrangement unit is used to mark the separated operating data of the monitored equipment as the operating data of the monitored equipment, extract the collection timestamp corresponding to the operating data of the monitored equipment, select the collection timestamp as the basis, associate the identification information of the monitored equipment with the corresponding operating data of the monitored equipment, and arrange them in chronological order.

[0017] The fault analysis set construction unit is used to select the identification information of the monitored equipment as a subset in the associated data set. The subset selects the collection timestamp of the monitored equipment's operating data as the key and the monitored equipment's operating data as the value to construct the monitored equipment fault analysis set.

[0018] Historical data is retrieved based on the acquisition cycle parameters of the target sensor, and the device identifier and operation data are separated. After being associated and sorted by acquisition timestamp, a fault analysis set is constructed with the device identifier as a subset, the timestamp as the key, and the operation data as the value. This achieves the time-series and structured integration of historical operation data, as well as accurate data retrieval, orderly association, and standardized set, providing a reliable data foundation for fault analysis.

[0019] The fault feature analysis set construction module is used to analyze structured data to obtain historical maintenance data of the monitored equipment, and then analyze the historical maintenance data to extract the corresponding operating data of the monitored equipment and construct the fault feature analysis set of the monitored equipment.

[0020] The fault feature analysis set construction module includes a historical maintenance data reading unit, a fault data duration threshold reading unit, and a fault feature analysis set integration unit.

[0021] The historical maintenance data reading unit is used to call the BIM model interface, read the structured data stored in the building equipment management section of the BIM model, and filter out the historical maintenance data of the monitored equipment from the structured data. The historical maintenance data includes the fault occurrence timestamp and the corresponding monitored equipment identifier. The historical maintenance data represents the data record of the monitored equipment after a fault occurred and was repaired during its historical service.

[0022] The fault data duration threshold reading unit is used to read the preset fault data acquisition duration threshold of the monitored device from the system configuration file. The fault data acquisition duration threshold is defined as a fixed acquisition duration parameter of continuous historical running data with the fault occurrence time as the endpoint when the monitored device fails. The fault data acquisition duration threshold has the same time unit as the data acquisition cycle of the target sensor.

[0023] The fault feature analysis set integration unit is used to select the fault occurrence time stamp as the benchmark for each fault occurrence time stamp and the corresponding monitored equipment identifier in the historical maintenance data, and to extract the monitored equipment operation data equal to the fault data collection duration threshold based on the fault data collection duration threshold. All the extracted monitored equipment operation data are integrated according to the time series to form the monitored equipment fault feature analysis set.

[0024] The BIM model interface is called to read historical maintenance data containing fault occurrence timestamps and equipment identifiers. The preset fault data duration threshold is read, and the corresponding equipment fault occurrence threshold duration is extracted from the fault analysis set and integrated according to the time series to construct a fault feature analysis set. This achieves accurate correlation between pre-fault operation data and historical maintenance data, as well as targeted data extraction and orderly integration, providing accurate data support for fault feature extraction.

[0025] The fault time prediction module is used to extract fluctuation characteristics from the operating data of the monitored equipment, and to build a fault prediction model to analyze the mapping relationship between fluctuation characteristics and timestamps. Based on the fault prediction model, the fault occurrence time of the monitored equipment is predicted and recorded as the predicted fault time.

[0026] The fault time prediction module includes a running data preprocessing unit, a fluctuation feature extraction unit, a fault prediction model construction unit, and a fault time prediction output unit.

[0027] The operation data preprocessing unit is used to preprocess the operation data of the monitored equipment in the fault feature analysis set of the monitored equipment. For the collection timestamps with missing data, the linear interpolation method is used to fill in the corresponding missing operation data so that each collection timestamp matches a unique operation data of the monitored equipment. The preprocessed operation data is rearranged in ascending order of collection timestamps to form an ordered preprocessed dataset.

[0028] The fluctuation feature extraction unit is used to extract fluctuation features from the ordered preprocessed dataset: N is defined as the preset number of short-term fluctuation calculation cycles; the difference between the monitored device's operating data corresponding to two adjacent acquisition timestamps is calculated and recorded as the instantaneous fluctuation value; the difference between the maximum and minimum values ​​of the monitored device's operating data within the timestamp range corresponding to N consecutive data acquisition cycles is calculated and recorded as the short-term fluctuation amplitude; the change in the monitored device's operating data within two adjacent acquisition cycles is calculated and divided by the duration of the data acquisition cycle and recorded as the fluctuation rate; the instantaneous fluctuation value, short-term fluctuation amplitude, and fluctuation rate are integrated to form a fluctuation feature set;

[0029] The fault prediction model construction unit is used to establish a mapping relationship between the instantaneous fluctuation value, short-term fluctuation amplitude and fluctuation rate in the fluctuation feature set and the corresponding fault occurrence timestamp in the fault feature analysis set of the monitored equipment, divide the fluctuation feature set into a fault training set and a fault test set according to a preset ratio, and obtain the fault prediction model by processing the fault training set and the fault test set through a neural network model.

[0030] The predicted fault time output unit is used to acquire the latest operating data of the monitored equipment transmitted by the target sensor in real time. After preprocessing, the real-time fluctuation characteristics are input into the fault prediction model, and the predicted fault time of the monitored equipment is output, which is recorded as the predicted fault time.

[0031] The formula for calculating instantaneous fluctuation values ​​is as follows: △x i =x i+1 -x i ;

[0032] In the formula, △x i The instantaneous fluctuation value x of the i-th instantaneous fluctuation is represented as follows. i+1 This represents the operating data of the monitored equipment at the timestamp corresponding to the (i+1)th instantaneous fluctuation; x i This represents the operating data of the monitored equipment, corresponding to the timestamp of the i-th instantaneous fluctuation.

[0033] The formula for calculating short-term volatility is as follows: A j =max(x j x j+1 , ..., x j+N-1 )-min(x j x j+1 , ..., x j+N-1 );

[0034] In the formula, A j This represents the fluctuation range of the monitored device's operating data within the j-th data acquisition period, specifically the difference between the maximum and minimum values ​​of the monitored device's operating data over N consecutive preset data acquisition periods; N represents the number of data acquisition periods for calculating the preset short-term fluctuation range.

[0035] The formula for calculating the rate of fluctuation is as follows: V i =△x i ÷t;

[0036] In the formula, V i Let represent the fluctuation rate of the i-th instantaneous fluctuation; t represents the duration of the data acquisition period.

[0037] The sensor acquisition frequency setting module is used to calculate the first acquisition frequency coefficient and the second acquisition frequency coefficient of the currently monitored device and perform dimensionless calculation, and perform weighted fusion based on preset weights to obtain a comprehensive dynamic adjustment coefficient and set the acquisition frequency.

[0038] The sensor acquisition frequency setting module includes a first acquisition frequency coefficient calculation unit, a second acquisition frequency coefficient calculation unit, a comprehensive dynamic adjustment coefficient calculation unit, an acquisition frequency range reading unit, an extreme value adjustment coefficient calculation unit, and an acquisition frequency calculation unit.

[0039] The first acquisition frequency coefficient calculation unit is used to extract the real-time fluctuation characteristics corresponding to the current operating data of the monitored equipment. The real-time fluctuation characteristics include instantaneous fluctuation value, short-term fluctuation amplitude, and fluctuation rate. The fluctuation characteristics corresponding to each fault sample are extracted from the fault characteristic analysis set of the monitored equipment to form a fault fluctuation feature library. The feature vector matching algorithm is used to calculate the matching degree between the real-time fluctuation characteristics and each feature vector in the fault fluctuation feature library, and the maximum matching degree value of the feature vector is selected as the first acquisition frequency coefficient of the monitored equipment.

[0040] The specific calculation process for the maximum matching degree of the feature vector is as follows:

[0041] The cosine similarity algorithm is selected as the core to perform feature vector matching. The real-time fluctuation features corresponding to the current operating data of the monitored equipment are selected to construct a real-time fluctuation feature vector, including instantaneous fluctuation value, short-term fluctuation amplitude, and fluctuation rate. Then, the fault sample fluctuation feature vector corresponding to each historical fault sample is extracted from the fault fluctuation feature library. The data dimension is consistent with the real-time vector. The cosine similarity is obtained by calculating the ratio of the dot product and the modulus product of the two types of vectors. After normalization, the similarity is mapped to the interval [0, 1] to form the normalized matching degree of each sample. Finally, the maximum normalized matching degree is selected as the first acquisition frequency coefficient. The larger the value of this coefficient, the higher the similarity between the current equipment fluctuation features and the historical fault features, and the higher the fault risk.

[0042] The second acquisition frequency coefficient calculation unit is used to obtain the current system time of the monitored device, calculate the time difference between the current system time and the predicted fault time, input the time difference into a preset inverse exponential growth function, and use the function output value as the second acquisition frequency coefficient of the monitored device.

[0043] The formula for calculating the second sampling frequency coefficient is as follows: ;

[0044] In the formula, C2 represents the second acquisition frequency coefficient; Δt represents the time interval between the current time and the predicted fault time; T represents the longest time interval between historical faults; and e is 2.718.

[0045] When Δt≤T, it is determined that the fault is within the period of concern. The second sampling frequency coefficient C2 increases exponentially as Δt decreases, and reaches the maximum value e when Δt=0.

[0046] When Δt > T, it is determined that the fault is not within the fault concern period, C2 is fixed at 1, and the sampling frequency of the target sensor does not change.

[0047] The comprehensive dynamic adjustment coefficient calculation unit is used to perform dimensionless processing on the first acquisition frequency coefficient and the second acquisition frequency coefficient respectively; call the preset weight parameters in the system configuration file, the weight parameters include the weight of the first acquisition frequency coefficient and the weight of the second acquisition frequency coefficient, and the sum of the weight of the first acquisition frequency coefficient and the weight of the second acquisition frequency coefficient is 1; multiply the dimensionless first acquisition frequency coefficient by the corresponding weight to obtain the first product, multiply the dimensionless second acquisition frequency coefficient by the corresponding weight to obtain the second product, and add the first product and the second product to obtain the comprehensive dynamic adjustment coefficient of the monitored equipment;

[0048] The sampling frequency range reading unit is used to call the system configuration file and read the minimum and maximum sampling frequencies preset by the target sensor. The minimum sampling frequency is the minimum data sampling frequency preset by the target sensor under normal operating conditions of the monitored equipment, and the maximum sampling frequency is the maximum data sampling frequency preset by the target sensor under high fault risk conditions of the monitored equipment.

[0049] The extreme value adjustment coefficient calculation unit is used to set the matching degree to 1 and the time difference to 0, and then input the weighted fusion to obtain the maximum dynamic adjustment coefficient; and to set the matching degree to 0 and the time difference to the predicted fault time, and then input the weighted fusion to obtain the minimum dynamic adjustment coefficient.

[0050] The sampling frequency calculation unit is used to calculate the difference between the preset maximum sampling frequency and the minimum sampling frequency of the target sensor, and denoted as the sampling frequency difference; calculate the difference between the maximum dynamic adjustment coefficient and the minimum dynamic adjustment coefficient of the target sensor, and denoted as the adjustment coefficient difference; divide the sampling frequency difference by the adjustment coefficient difference to obtain the proportional mapping value between the sampling frequency and the dynamic adjustment coefficient of the target sensor; and multiply the comprehensive dynamic adjustment coefficient of the target sensor by the proportional mapping value and add the minimum sampling frequency to obtain the sampling frequency of the target sensor.

[0051] The system preprocesses the operational data in the fault feature analysis set to extract fluctuation features such as instantaneous fluctuation value, short-term fluctuation amplitude, and fluctuation rate. By constructing a neural network model to map the relationship between fluctuation features and fault timestamps, it achieves accurate prediction of the fault occurrence time of the monitored equipment. The data preprocessing is complete, the feature extraction is comprehensive, and the model prediction is reliable. The sensor acquisition frequency setting module calculates the maximum matching degree between the real-time fluctuation features and the fault feature library to obtain the first acquisition frequency coefficient. The second acquisition frequency coefficient is obtained by combining the time difference and the inverse exponential growth function. After dimensionless transformation and weighted fusion, a comprehensive dynamic adjustment coefficient is formed. Then, the target acquisition frequency is calculated based on the maximum and minimum acquisition frequencies to achieve dynamic adaptation of the acquisition frequency.

[0052] The multi-sensor operation and maintenance monitoring module is used to process the various sensors corresponding to the building equipment monitoring needs and monitor the amount of data collected by each sensor.

[0053] The multi-sensor operation and maintenance monitoring module includes an operation and maintenance sliding window initialization unit, a total data monitoring threshold calculation unit, and a data increment anomaly early warning unit.

[0054] The operation and maintenance sliding window initialization unit is used to set the data operation and maintenance sliding window and call the system configuration file. It reads the preset window duration and sliding step parameters of the data operation and maintenance sliding window. The window duration represents the time span of the data operation and maintenance sliding window capturing data information from the fault analysis set of the monitored equipment in a single session. The sliding step represents the time interval of each movement of the data operation and maintenance sliding window along the time axis. The initial start time of the data operation and maintenance sliding window is set to the current system time minus the window duration, and the initial end time is set to the current system time, thus completing the initialization configuration of the data operation and maintenance sliding window.

[0055] The data total monitoring threshold calculation unit is used to obtain the historical sliding process data of the data operation and maintenance sliding window, extract the total number of newly added monitoring data records of all sensors in the building equipment management section within the time interval corresponding to each historical sliding step, and record it as the historical data increment for each step. Based on the historical data increment for each step, its arithmetic mean is calculated and recorded as the historical data increment average. At the same time, the standard deviation of the historical data increment for each step is calculated and recorded as the historical data increment standard deviation. The historical data increment average and the historical data increment standard deviation are added together to obtain the data total monitoring threshold.

[0056] The data increment anomaly early warning unit is used to, after the data operation and maintenance sliding window completes one slide along the time axis according to the preset sliding step size, count the total number of newly added monitoring data records of all sensors in the building equipment management section within the time interval corresponding to the current sliding step size, and record it as the current step size data increment. The current step size data increment is then compared with the total data monitoring threshold. If the current step size data increment exceeds the total data monitoring threshold, a data increment fluctuation anomaly early warning signal is issued. If the current step size data increment does not exceed the total data monitoring threshold, the monitoring and maintenance of the data information captured in the data operation and maintenance sliding window continues.

[0057] The operation and maintenance sliding window initialization unit sets the window duration and sliding step size and completes the window initialization. The total data monitoring threshold calculation unit calculates the arithmetic mean and standard deviation of the data increments at each historical step size and adds them together to obtain the total data monitoring threshold. The data increment anomaly early warning unit counts the data increment at the current step size after the window slides by step size and compares it with the threshold. When the limit is exceeded, an anomaly early warning is issued. When the limit is not exceeded, monitoring continues. In this way, dynamic monitoring of the data collected by various sensors related to building equipment monitoring can be achieved. Thresholds can be set scientifically, data increment anomalies can be identified in a timely manner, and the stability of multi-sensor operation and maintenance can be guaranteed.

[0058] The output of the target sensor determination module is electrically connected to the input of the fault analysis set construction module; the output of the fault analysis set construction module is electrically connected to the input of the fault feature analysis set construction module; the output of the fault feature analysis set construction module is electrically connected to the input of the fault time prediction module; the output of the fault time prediction module is electrically connected to the input of the sensor acquisition frequency setting module; and the output of the sensor acquisition frequency setting module is electrically connected to the input of the multi-sensor operation and maintenance monitoring module.

[0059] Compared with the prior art, the beneficial effects of the present invention are:

[0060] 1. This invention uses a target sensor determination module to read structured data from the building equipment management section based on the BIM model. It constructs an associated dataset with the unique code of the monitored equipment as the key and the sensor type identifier as the value, and selects the target sensor. Combined with a fault analysis set construction module, it integrates historical operating data in a time-series manner. The fault feature analysis set construction module associates pre-fault operating data and maintenance data, achieving standardized data association and accurate integration. This provides reliable data support for subsequent fault prediction and effectively matches the needs of building equipment monitoring.

[0061] 2. This invention preprocesses the data in the fault feature analysis set through the fault time prediction module, extracts fluctuation features such as instantaneous fluctuation values, and constructs a neural network model to accurately predict the fault time of the monitored equipment; the sensor acquisition frequency setting module calculates the real-time fluctuation feature matching degree to obtain the first coefficient, and combines the time difference with the inverse exponential growth function to obtain the second coefficient. The acquisition frequency is dynamically adjusted through weighted fusion to balance monitoring accuracy and resource consumption.

[0062] 3. This invention initializes the sliding window through a multi-sensor operation and maintenance monitoring module, calculates monitoring thresholds based on historical data, and provides timely warnings of anomalies by comparing current data increments, thus ensuring the stability of multi-sensor operation and maintenance. Furthermore, modules such as target sensor determination and fault analysis are electrically connected in sequence to form a complete process from sensor determination to operation and maintenance monitoring, connecting all aspects of digital building operation and maintenance, and improving collaborative management efficiency and system reliability. Attached Figure Description

[0063] Figure 1 This is a schematic diagram of the structure of a BIM-based digital building operation and maintenance collaborative management system according to the present invention. Detailed Implementation

[0064] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0065] Example 1: As Figure 1 As shown, the present invention provides a technical solution, a BIM-based building digital operation and maintenance collaborative management system, which includes: a target sensor determination module, a fault analysis set construction module, a fault feature analysis set construction module, a fault time prediction module, a sensor acquisition frequency setting module, and a multi-sensor operation and maintenance monitoring module.

[0066] The target sensor determination module is used to acquire structured data of the building equipment management section in the BIM model, and analyze and extract the associated data set consisting of various sensor types, monitored equipment and data acquisition parameters corresponding to the building equipment monitoring requirements.

[0067] The target sensor determination module includes a structured data reading unit, an associated data set construction unit, and a target sensor selection unit;

[0068] The structured data reading unit is used to read the inherent attribute data of building equipment and sensor deployment configuration data contained in the building equipment management section of the BIM model through the structured data storage path and standard data interface of the building equipment management section, and obtain structured data.

[0069] The associated data set construction unit is used to parse the fields of the read structured data, extract the type identifier of each sensor and the unique code of the monitored device, select the unique code of the monitored device as the key, and select the type identifier of the sensor as the value to construct the associated data set;

[0070] The target sensor selection unit is used to classify the system category to which the monitored device belongs in the associated data set according to the sensor type identifier, and select any sensor from the classified associated data set as the target sensor;

[0071] The fault analysis set construction module is used to acquire historical data and analyze the monitored equipment and its operating data that the target sensor is in service with. After being arranged in an orderly manner according to the time series, the fault analysis set of the monitored equipment is constructed.

[0072] The fault analysis set construction module includes a historical data retrieval unit, a data time-related arrangement unit, and a fault analysis set construction unit.

[0073] The historical data retrieval unit is used to read the preset data acquisition cycle parameters of the target sensor in the BIM model, retrieve the historical acquisition data generated by the target sensor during its service period from the data storage module based on the data acquisition cycle parameters, and separate the identification information of the monitored equipment and the operating data of the monitored equipment from the retrieved historical acquisition data.

[0074] The data time association and arrangement unit is used to mark the separated operating data of the monitored equipment as the operating data of the monitored equipment, extract the collection timestamp corresponding to the operating data of the monitored equipment, select the collection timestamp as the basis, associate the identification information of the monitored equipment with the corresponding operating data of the monitored equipment, and arrange them in chronological order.

[0075] The fault analysis set construction unit is used to select the identification information of the monitored equipment as a subset in the associated data set. The subset selects the collection timestamp of the monitored equipment's operating data as the key and the monitored equipment's operating data as the value to construct the monitored equipment fault analysis set.

[0076] The fault feature analysis set construction module is used to analyze structured data to obtain historical maintenance data of the monitored equipment, and then analyze the historical maintenance data to extract the corresponding operating data of the monitored equipment and construct the fault feature analysis set of the monitored equipment.

[0077] The fault feature analysis set construction module includes a historical maintenance data reading unit, a fault data duration threshold reading unit, and a fault feature analysis set integration unit.

[0078] The historical maintenance data reading unit is used to call the BIM model interface, read the structured data stored in the building equipment management section of the BIM model, and filter out the historical maintenance data of the monitored equipment from the structured data. The historical maintenance data includes the fault occurrence timestamp and the corresponding monitored equipment identifier. The historical maintenance data represents the data record of the monitored equipment after a fault occurred and was repaired during its historical service.

[0079] The fault data duration threshold reading unit is used to read the preset fault data acquisition duration threshold of the monitored device from the system configuration file. The fault data acquisition duration threshold is defined as a fixed acquisition duration parameter of continuous historical running data with the fault occurrence time as the endpoint when the monitored device fails. The fault data acquisition duration threshold has the same time unit as the data acquisition cycle of the target sensor.

[0080] The fault feature analysis set integration unit is used to select the fault occurrence time stamp as the benchmark for each fault occurrence time stamp and the corresponding monitored equipment identifier in the historical maintenance data, and to extract the monitored equipment operation data equal to the fault data collection duration threshold based on the fault data collection duration threshold. All the extracted monitored equipment operation data are integrated according to the time series to form the monitored equipment fault feature analysis set.

[0081] The fault time prediction module is used to extract fluctuation characteristics from the operating data of the monitored equipment, and to build a fault prediction model to analyze the mapping relationship between fluctuation characteristics and timestamps. Based on the fault prediction model, the fault occurrence time of the monitored equipment is predicted and recorded as the predicted fault time.

[0082] The fault time prediction module includes a running data preprocessing unit, a fluctuation feature extraction unit, a fault prediction model construction unit, and a fault time prediction output unit.

[0083] The operation data preprocessing unit is used to preprocess the operation data of the monitored equipment in the fault feature analysis set of the monitored equipment. For the collection timestamps with missing data, the linear interpolation method is used to fill in the corresponding missing operation data so that each collection timestamp matches a unique operation data of the monitored equipment. The preprocessed operation data is rearranged in ascending order of collection timestamps to form an ordered preprocessed dataset.

[0084] The fluctuation feature extraction unit is used to extract fluctuation features from the ordered preprocessed dataset: N is defined as the preset number of short-term fluctuation calculation cycles; the difference between the monitored device's operating data corresponding to two adjacent acquisition timestamps is calculated and recorded as the instantaneous fluctuation value; the difference between the maximum and minimum values ​​of the monitored device's operating data within the timestamp range corresponding to N consecutive data acquisition cycles is calculated and recorded as the short-term fluctuation amplitude; the change in the monitored device's operating data within two adjacent acquisition cycles is calculated and divided by the duration of the data acquisition cycle and recorded as the fluctuation rate; the instantaneous fluctuation value, short-term fluctuation amplitude, and fluctuation rate are integrated to form a fluctuation feature set;

[0085] The fault prediction model construction unit is used to establish a mapping relationship between the instantaneous fluctuation value, short-term fluctuation amplitude and fluctuation rate in the fluctuation feature set and the corresponding fault occurrence timestamp in the fault feature analysis set of the monitored equipment, divide the fluctuation feature set into a fault training set and a fault test set according to a preset ratio, and obtain the fault prediction model by processing the fault training set and the fault test set through a neural network model.

[0086] The predicted fault time output unit is used to acquire the latest operating data of the monitored equipment transmitted by the target sensor in real time. After preprocessing, the real-time fluctuation characteristics are input into the fault prediction model, and the predicted fault time of the monitored equipment is output, which is recorded as the predicted fault time.

[0087] The sensor acquisition frequency setting module is used to calculate the first acquisition frequency coefficient and the second acquisition frequency coefficient of the currently monitored device and perform dimensionless calculation, and perform weighted fusion based on preset weights to obtain a comprehensive dynamic adjustment coefficient and set the acquisition frequency.

[0088] The sensor acquisition frequency setting module includes a first acquisition frequency coefficient calculation unit, a second acquisition frequency coefficient calculation unit, a comprehensive dynamic adjustment coefficient calculation unit, an acquisition frequency range reading unit, an extreme value adjustment coefficient calculation unit, and an acquisition frequency calculation unit.

[0089] The first acquisition frequency coefficient calculation unit is used to extract the real-time fluctuation characteristics corresponding to the current operating data of the monitored equipment. The real-time fluctuation characteristics include instantaneous fluctuation value, short-term fluctuation amplitude, and fluctuation rate. The fluctuation characteristics corresponding to each fault sample are extracted from the fault characteristic analysis set of the monitored equipment to form a fault fluctuation feature library. The feature vector matching algorithm is used to calculate the matching degree between the real-time fluctuation characteristics and each feature vector in the fault fluctuation feature library, and the maximum matching degree value of the feature vector is selected as the first acquisition frequency coefficient of the monitored equipment.

[0090] The second acquisition frequency coefficient calculation unit is used to obtain the current system time of the monitored device, calculate the time difference between the current system time and the predicted fault time, input the time difference into a preset inverse exponential growth function, and use the function output value as the second acquisition frequency coefficient of the monitored device.

[0091] The comprehensive dynamic adjustment coefficient calculation unit is used to perform dimensionless processing on the first acquisition frequency coefficient and the second acquisition frequency coefficient respectively; call the preset weight parameters in the system configuration file, the weight parameters include the weight of the first acquisition frequency coefficient and the weight of the second acquisition frequency coefficient, and the sum of the weight of the first acquisition frequency coefficient and the weight of the second acquisition frequency coefficient is 1; multiply the dimensionless first acquisition frequency coefficient by the corresponding weight to obtain the first product, multiply the dimensionless second acquisition frequency coefficient by the corresponding weight to obtain the second product, and add the first product and the second product to obtain the comprehensive dynamic adjustment coefficient of the monitored equipment;

[0092] The sampling frequency range reading unit is used to call the system configuration file and read the minimum and maximum sampling frequencies preset by the target sensor. The minimum sampling frequency is the minimum data sampling frequency preset by the target sensor under normal operating conditions of the monitored equipment, and the maximum sampling frequency is the maximum data sampling frequency preset by the target sensor under high fault risk conditions of the monitored equipment.

[0093] The extreme value adjustment coefficient calculation unit is used to set the matching degree to 1 and the time difference to 0, and then input the weighted fusion to obtain the maximum dynamic adjustment coefficient; and to set the matching degree to 0 and the time difference to the predicted fault time, and then input the weighted fusion to obtain the minimum dynamic adjustment coefficient.

[0094] The sampling frequency calculation unit is used to calculate the difference between the preset maximum sampling frequency and the minimum sampling frequency of the target sensor, and denoted as the sampling frequency difference; calculate the difference between the maximum dynamic adjustment coefficient and the minimum dynamic adjustment coefficient of the target sensor, and denoted as the adjustment coefficient difference; divide the sampling frequency difference by the adjustment coefficient difference to obtain the proportional mapping value between the sampling frequency and the dynamic adjustment coefficient of the target sensor; and multiply the comprehensive dynamic adjustment coefficient of the target sensor by the proportional mapping value and add the minimum sampling frequency to obtain the sampling frequency of the target sensor.

[0095] The multi-sensor operation and maintenance monitoring module is used to process the various sensors corresponding to the building equipment monitoring needs and monitor the amount of data collected by each sensor.

[0096] The multi-sensor operation and maintenance monitoring module includes an operation and maintenance sliding window initialization unit, a total data monitoring threshold calculation unit, and a data increment anomaly early warning unit.

[0097] The operation and maintenance sliding window initialization unit is used to set the data operation and maintenance sliding window and call the system configuration file. It reads the preset window duration and sliding step parameters of the data operation and maintenance sliding window. The window duration represents the time span of the data operation and maintenance sliding window capturing data information from the fault analysis set of the monitored equipment in a single session. The sliding step represents the time interval of each movement of the data operation and maintenance sliding window along the time axis. The initial start time of the data operation and maintenance sliding window is set to the current system time minus the window duration, and the initial end time is set to the current system time, thus completing the initialization configuration of the data operation and maintenance sliding window.

[0098] The data total monitoring threshold calculation unit is used to obtain the historical sliding process data of the data operation and maintenance sliding window, extract the total number of newly added monitoring data records of all sensors in the building equipment management section within the time interval corresponding to each historical sliding step, and record it as the historical data increment for each step. Based on the historical data increment for each step, its arithmetic mean is calculated and recorded as the historical data increment average. At the same time, the standard deviation of the historical data increment for each step is calculated and recorded as the historical data increment standard deviation. The historical data increment average and the historical data increment standard deviation are added together to obtain the data total monitoring threshold.

[0099] The data increment anomaly early warning unit is used to, after the data operation and maintenance sliding window completes one slide along the time axis according to the preset sliding step size, count the total number of newly added monitoring data records of all sensors in the building equipment management section within the time interval corresponding to the current sliding step size, and record it as the current step size data increment. The current step size data increment is then compared with the total data monitoring threshold. If the current step size data increment exceeds the total data monitoring threshold, a data increment fluctuation anomaly early warning signal is issued. If the current step size data increment does not exceed the total data monitoring threshold, the monitoring and maintenance of the data information captured in the data operation and maintenance sliding window continues.

[0100] For example, the instantaneous fluctuation value is 0.6℃, the short-term fluctuation amplitude is 1.5℃, and the fluctuation rate is 0.12℃ / min; the fault sample vector V for 2024-03-15 is extracted from the fault fluctuation feature database. 故障1 =【0.5℃,1.4℃,0.11℃ / min】;

[0101] Calculate cosine similarity:

[0102] The dot product = 0.6 × 0.5 + 1.5 × 1.4 + 0.12 × 0.11 = 0.3 + 2.1 + 0.0132 = 2.4132; V 实时模长 = = = ≈1.62; V 故障1模长 = = = ≈1.491;

[0103] Similarity = 2.4132 ÷ (1.62 × 1.491) ≈ 2.4132 ÷ 2.415 ≈ 0.999;

[0104] Normalization: (0.999+1)÷2≈0.999;

[0105] The maximum matching degree C1 was obtained as 0.999;

[0106] The current system time is 2024-04-01 10:00:00, the predicted fault time is 2024-04-01 11:20:00, △t=80min; the longest historical fault interval T=60min (fault warning interval from 2024-07-20).

[0107] Since Δt > T, according to the following: when Δt ≤ T, it is determined that the fault is within the period of concern. The second sampling frequency coefficient C2 increases exponentially as Δt decreases, and reaches its maximum value e when Δt = 0.

[0108] When Δt > T, it is determined that the fault is not within the fault concern period, C2 is fixed at 1, and the sampling frequency of the target sensor does not change.

[0109] We get C2=1;

[0110] After dimensionless transformation, C1=0.999; C2=1; the weight of the first acquisition frequency coefficient is 0.8, and the weight of the second acquisition frequency coefficient is 0.2.

[0111] The comprehensive dynamic adjustment coefficient can be obtained from the weighted fusion calculation:

[0112] C 综合 =0.999×0.8+1×0.2=0.7992+0.2=0.9992;

[0113] The minimum acquisition frequency of the target sensor is 0.2 times / min, and the maximum acquisition frequency is 1 time / min.

[0114] Extreme value adjustment coefficient calculation:

[0115] 1×0.8+e×0.2≈0.8+0.5436=1.3436; (When the matching degree is 1 and Δt=0, C2 takes the natural constant e≈2.718, and the maximum dynamic adjustment coefficient is calculated.)

[0116] 0×0.8+1×0.2=0.2; (When the matching degree is 0 and Δt>T, C2=1, and the minimum dynamic adjustment coefficient is calculated.)

[0117] Sampling frequency difference: 1 - 0.2 = 0.8 times / min;

[0118] Adjustment factor difference: 1.3436 - 0.2 = 1.1436;

[0119] By dividing the difference in acquisition frequencies by the difference in adjustment coefficients, we obtain the ratio mapping value between the target sensor's acquisition frequency and the dynamic adjustment coefficient:

[0120] 0.8÷1.1436≈0.6995 times / min;

[0121] The acquisition frequency of the target sensor is obtained by multiplying the comprehensive dynamic adjustment coefficient of the target sensor by the product of the proportional mapping value and adding it to the minimum acquisition frequency.

[0122] 0.6995 × 0.9992 + 0.2 ≈ 0.6989 + 0.2 = 0.8989;

[0123] Rounding, the target sensor's acquisition frequency is 0.9 times / min.

[0124] The output of the target sensor determination module is electrically connected to the input of the fault analysis set construction module; the output of the fault analysis set construction module is electrically connected to the input of the fault feature analysis set construction module; the output of the fault feature analysis set construction module is electrically connected to the input of the fault time prediction module; the output of the fault time prediction module is electrically connected to the input of the sensor acquisition frequency setting module; and the output of the sensor acquisition frequency setting module is electrically connected to the input of the multi-sensor operation and maintenance monitoring module.

[0125] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A BIM-based digital building operation and maintenance collaborative management system, characterized in that: The building digital operation and maintenance collaborative management system includes a target sensor determination module, a fault analysis set construction module, a fault feature analysis set construction module, a fault time prediction module, a sensor acquisition frequency setting module, and a multi-sensor operation and maintenance monitoring module. The target sensor determination module is used to acquire structured data of the building equipment management section in the BIM model, and analyze and extract the associated data set consisting of various sensor types, monitored equipment and data acquisition parameters corresponding to the building equipment monitoring requirements. The fault analysis set construction module is used to acquire historical data and analyze the monitored equipment and its operating data that the target sensor is in service with. After being arranged in an orderly manner according to the time series, the fault analysis set of the monitored equipment is constructed. The fault feature analysis set construction module is used to analyze structured data to obtain historical maintenance data of the monitored equipment, and then analyze the historical maintenance data to extract the corresponding operating data of the monitored equipment and construct the fault feature analysis set of the monitored equipment. The fault time prediction module is used to extract fluctuation characteristics from the operating data of the monitored equipment, and to build a fault prediction model to analyze the mapping relationship between fluctuation characteristics and timestamps. Based on the fault prediction model, the fault occurrence time of the monitored equipment is predicted and recorded as the predicted fault time. The sensor acquisition frequency setting module is used to calculate the first acquisition frequency coefficient and the second acquisition frequency coefficient of the currently monitored device and perform dimensionless calculation, and perform weighted fusion based on preset weights to obtain a comprehensive dynamic adjustment coefficient and set the acquisition frequency. The multi-sensor operation and maintenance monitoring module is used to process the various sensors corresponding to the building equipment monitoring needs and to monitor the amount of data collected by each sensor.

2. The BIM-based digital building operation and maintenance collaborative management system according to claim 1, characterized in that: The target sensor determination module includes a structured data reading unit, an associated data set construction unit, and a target sensor selection unit; The structured data reading unit is used to read the inherent attribute data of building equipment and sensor deployment configuration data contained in the building equipment management section of the BIM model through the structured data storage path and standard data interface of the building equipment management section, and obtain structured data. The associated data set construction unit is used to parse the fields of the read structured data, extract the type identifier of each sensor and the unique code of the monitored device, select the unique code of the monitored device as the key, and select the type identifier of the sensor as the value to construct the associated data set; The target sensor selection unit is used to classify the system category to which the monitored device belongs in the associated data set according to the sensor type identifier, and select any sensor from the classified associated data set as the target sensor.

3. The BIM-based digital building operation and maintenance collaborative management system according to claim 2, characterized in that: The fault analysis set construction module includes a historical data retrieval unit, a data time-related arrangement unit, and a fault analysis set construction unit. The historical data retrieval unit is used to read the preset data acquisition cycle parameters of the target sensor in the BIM model, retrieve the historical acquisition data generated by the target sensor during its service period from the data storage module based on the data acquisition cycle parameters, and separate the identification information of the monitored equipment and the operating data of the monitored equipment from the retrieved historical acquisition data. The data time association and arrangement unit is used to mark the separated operating data of the monitored equipment as the operating data of the monitored equipment, extract the collection timestamp corresponding to the operating data of the monitored equipment, select the collection timestamp as the basis, associate the identification information of the monitored equipment with the corresponding operating data of the monitored equipment, and arrange them in chronological order. The fault analysis set construction unit is used to select the identification information of the monitored equipment as a subset in the associated data set. The subset selects the collection timestamp of the monitored equipment's operating data as the key and the monitored equipment's operating data as the value to construct the monitored equipment fault analysis set.

4. The BIM-based digital building operation and maintenance collaborative management system according to claim 3, characterized in that: The fault feature analysis set construction module includes a historical maintenance data reading unit and a fault data duration threshold reading unit; The historical maintenance data reading unit is used to call the BIM model interface, read the structured data stored in the building equipment management section of the BIM model, and filter out the historical maintenance data of the monitored equipment from the structured data. The historical maintenance data includes the fault occurrence timestamp and the corresponding monitored equipment identifier. The historical maintenance data represents the data record of the monitored equipment after a fault occurred and was repaired during its historical service. The fault data duration threshold reading unit is used to read the preset fault data acquisition duration threshold of the monitored device from the system configuration file. The fault data acquisition duration threshold is defined as a fixed acquisition duration parameter of continuous historical operating data with the fault occurrence time as the endpoint when the monitored device fails. The fault data acquisition duration threshold has the same time unit as the data acquisition cycle of the target sensor.

5. A BIM-based building digital operation and maintenance collaborative management system according to claim 4, characterized in that: The fault feature analysis set construction module also includes a fault feature analysis set integration unit; The fault feature analysis set integration unit is used to select the fault occurrence timestamp as a benchmark from the fault analysis set of the monitored equipment based on the fault data collection duration threshold and the monitored equipment identifier for each fault occurrence timestamp and the corresponding monitored equipment identifier in the historical maintenance data. It then extracts the monitored equipment operation data that is equal to the fault data collection duration threshold, and integrates all the extracted monitored equipment operation data according to the time series to form the monitored equipment fault feature analysis set.

6. A BIM-based digital building operation and maintenance collaborative management system according to claim 5, characterized in that: The fault time prediction module includes a running data preprocessing unit and a fluctuation feature extraction unit; The operation data preprocessing unit is used to preprocess the operation data of the monitored equipment in the fault feature analysis set of the monitored equipment. For the collection timestamps with missing data, the linear interpolation method is used to fill in the corresponding missing operation data so that each collection timestamp matches a unique operation data of the monitored equipment. The preprocessed operation data is rearranged in ascending order of collection timestamps to form an ordered preprocessed dataset. The fluctuation feature extraction unit is used to extract fluctuation features from the ordered preprocessed dataset: N is defined as the preset number of short-term fluctuation calculation cycles; the difference between the operating data of the monitored device corresponding to two adjacent acquisition timestamps is calculated and recorded as the instantaneous fluctuation value; the difference between the maximum and minimum values ​​of the operating data of the monitored device within the timestamp range corresponding to N consecutive data acquisition cycles is calculated and recorded as the short-term fluctuation amplitude; the change in the operating data of the monitored device within two adjacent acquisition cycles is divided by the duration of the data acquisition cycle and recorded as the fluctuation rate; the instantaneous fluctuation value, short-term fluctuation amplitude, and fluctuation rate are integrated to form a fluctuation feature set.

7. A BIM-based digital building operation and maintenance collaborative management system according to claim 6, characterized in that: The failure time prediction module also includes a failure prediction model construction unit and a failure time prediction output unit; The fault prediction model construction unit is used to establish a mapping relationship between the instantaneous fluctuation value, short-term fluctuation amplitude and fluctuation rate in the fluctuation feature set and the corresponding fault occurrence timestamp in the fault feature analysis set of the monitored equipment, divide the fluctuation feature set into a fault training set and a fault test set according to a preset ratio, and obtain the fault prediction model by processing the fault training set and the fault test set through a neural network model. The predicted fault time output unit is used to acquire the latest operating data of the monitored equipment transmitted by the target sensor in real time. After preprocessing, the real-time fluctuation characteristics are input into the fault prediction model, and the predicted fault time of the monitored equipment is output, which is recorded as the predicted fault time.

8. A BIM-based digital building operation and maintenance collaborative management system according to claim 7, characterized in that: The sensor acquisition frequency setting module includes a first acquisition frequency coefficient calculation unit, a second acquisition frequency coefficient calculation unit, and a comprehensive dynamic adjustment coefficient calculation unit; The first acquisition frequency coefficient calculation unit is used to extract the real-time fluctuation characteristics corresponding to the current operating data of the monitored equipment. The real-time fluctuation characteristics include instantaneous fluctuation value, short-term fluctuation amplitude, and fluctuation rate. The fluctuation characteristics corresponding to each fault sample are extracted from the fault characteristic analysis set of the monitored equipment to form a fault fluctuation feature library. The feature vector matching algorithm is used to calculate the matching degree between the real-time fluctuation feature and each feature vector in the fault fluctuation feature library, and the maximum matching degree value of the feature vector is selected as the first acquisition frequency coefficient of the monitored equipment. The second acquisition frequency coefficient calculation unit is used to obtain the current system time of the monitored device, calculate the time difference between the current system time and the predicted fault time, input the time difference into a preset inverse exponential growth function, and use the function output value as the second acquisition frequency coefficient of the monitored device. The integrated dynamic adjustment coefficient calculation unit is used to perform dimensionless processing on the first acquisition frequency coefficient and the second acquisition frequency coefficient respectively. The system calls the preset weight parameters in the system configuration file. The weight parameters include the weight of the first acquisition frequency coefficient and the weight of the second acquisition frequency coefficient, and the sum of the weights of the first acquisition frequency coefficient and the second acquisition frequency coefficient is 1. The dimensionless first acquisition frequency coefficient is multiplied by its corresponding weight to obtain the first product, and the dimensionless second acquisition frequency coefficient is multiplied by its corresponding weight to obtain the second product. The first product and the second product are added together to obtain the comprehensive dynamic adjustment coefficient of the monitored equipment.

9. A BIM-based digital building operation and maintenance collaborative management system according to claim 8, characterized in that: The sensor acquisition frequency setting module also includes an acquisition frequency range reading unit, an extreme value adjustment coefficient calculation unit, and an acquisition frequency calculation unit; The sampling frequency range reading unit is used to call the system configuration file and read the minimum and maximum sampling frequencies preset by the target sensor. The minimum sampling frequency is the minimum data sampling frequency preset by the target sensor under normal operating conditions of the monitored equipment, and the maximum sampling frequency is the maximum data sampling frequency preset by the target sensor under high fault risk conditions of the monitored equipment. The extreme value adjustment coefficient calculation unit is used to set the matching degree to 1, the time difference to 0, and substitute it into the weighted fusion to obtain the maximum dynamic adjustment coefficient. Set the matching degree to 0, the time difference to the predicted failure time, and substitute it into the weighted fusion to obtain the minimum dynamic adjustment coefficient; The sampling frequency calculation unit is used to calculate the difference between the preset maximum sampling frequency and the minimum sampling frequency of the target sensor, which is denoted as the sampling frequency difference. Calculate the difference between the maximum and minimum dynamic adjustment coefficients of the target sensor, and denote it as the adjustment coefficient difference. The ratio mapping value between the target sensor's acquisition frequency and the dynamic adjustment coefficient is obtained by dividing the acquisition frequency difference by the adjustment coefficient difference. The acquisition frequency of the target sensor is obtained by multiplying the comprehensive dynamic adjustment coefficient of the target sensor by the ratio mapping value and adding the minimum acquisition frequency.

10. A BIM-based digital building operation and maintenance collaborative management system according to claim 9, characterized in that: The multi-sensor operation and maintenance monitoring module includes an operation and maintenance sliding window initialization unit, a total data monitoring threshold calculation unit, and a data increment anomaly early warning unit. The operation and maintenance sliding window initialization unit is used to set the data operation and maintenance sliding window and call the system configuration file. It reads the preset window duration and sliding step parameters of the data operation and maintenance sliding window. The window duration represents the time span of the data operation and maintenance sliding window capturing data information from the fault analysis set of the monitored equipment in a single session. The sliding step represents the time interval of each movement of the data operation and maintenance sliding window along the time axis. The initial start time of the data operation and maintenance sliding window is set to the current system time minus the window duration, and the initial end time is set to the current system time, thus completing the initialization configuration of the data operation and maintenance sliding window. The data total monitoring threshold calculation unit is used to obtain the historical sliding process data of the data operation and maintenance sliding window, extract the total number of newly added monitoring data records of all sensors in the building equipment management section within the time interval corresponding to each historical sliding step, and record it as the historical data increment for each step. Based on the historical data increment for each step, its arithmetic mean is calculated and recorded as the historical data increment average. At the same time, the standard deviation of the historical data increment for each step is calculated and recorded as the historical data increment standard deviation. The historical data increment average and the historical data increment standard deviation are added together to obtain the data total monitoring threshold. The data increment anomaly early warning unit is used to, after the data operation and maintenance sliding window completes one slide along the time axis according to the preset sliding step size, count the total number of newly added monitoring data records of all sensors in the building equipment management section within the time interval corresponding to the current sliding step size, and record it as the current step size data increment. The current step size data increment is then compared with the total data monitoring threshold. If the current step size data increment exceeds the total data monitoring threshold, a data increment fluctuation anomaly early warning signal is issued. If the current step size data increment does not exceed the total data monitoring threshold, the monitoring and maintenance of the data information captured in the data operation and maintenance sliding window continues.

Images