Building electrical system fault diagnosis system and method

Abstract

The invention discloses a building electrical system fault diagnosis system and method. The system comprises a data acquisition module, a load analysis module, an equipment monitoring module, a model adjustment module, a fault diagnosis module and an alarm and display module. Through cooperative work of the data acquisition module, the load analysis module and the model adjustment module, dynamic changes of an electrical system can be automatically adapted, the working intensity of operation and maintenance personnel is reduced, the operation and maintenance efficiency is remarkably improved, the model adjustment module is utilized, a diagnosis model can be automatically optimized according to load changes and equipment updating, and the diagnosis efficiency is improved. The accuracy and reliability of fault diagnosis are remarkably improved, real faults are accurately recognized, misjudgment and missed judgment are avoided, and through accurate judgment of the fault diagnosis module and timely notification of the alarm and display module, the operation and maintenance cost is effectively reduced, the damage risk of equipment is reduced, and safety accidents are effectively avoided.

Description

Technical Field

[0001] The present invention belongs to the technical field of electrical system fault diagnosis, and more specifically, relates to a building electrical system fault diagnosis system and method. Background Art

[0002] Building electrical system is a general term for electrical systems in different stages of the building life cycle from construction to commissioning, which is based on the building as a platform. It covers multiple components to ensure the normal operation and power safety of various electrical equipment at all stages. In modern buildings, the stable operation of electrical systems is crucial. However, due to the complexity and diversity of building electrical systems, their failures are often difficult to predict.

[0003] Traditional fault diagnosis methods usually rely on fixed diagnostic models and preset parameters. These models and parameters can play their role to a certain extent when the load of the electrical system is relatively stable and the equipment status is relatively constant. However, with the continuous expansion of existing building functions and the trend of intelligent development, the load of the electrical system changes more frequently, and the connection of new equipment and the updating of old equipment have also become the norm. In this dynamically changing environment, the traditional fixed diagnostic mode is gradually unable to meet the existing usage needs.

[0004] The traditional fixed diagnosis mode is often difficult to adapt to the dynamic changes of electrical system load and the replacement of equipment when in use. When the load suddenly increases significantly, it may be misjudged as a fault according to the original diagnostic threshold, resulting in unnecessary alarms and maintenance operations; and when the equipment is updated, its performance and characteristics change. If the diagnostic model is not adjusted accordingly, the real fault will be missed, which will bring potential risks to the safe operation of the building electrical system. Therefore, it is necessary to propose a building electrical system fault diagnosis system and method to solve the problems existing in the prior art. Summary of the invention

[0005] In response to the above defects or improvement needs of the prior art, the present invention provides a building electrical system fault diagnosis system and method, which enables the building electrical system to automatically adjust the parameters and algorithms of the diagnosis model according to its load changes, equipment updates and other real-time conditions, so as to maintain efficient and accurate diagnosis capabilities; and the fault diagnosis system realizes the transmission and alarm functions of fault information, so that the fault can be repaired as soon as possible.

[0006] In order to achieve the above object, the present invention adopts the following technical solution to provide a building electrical system fault diagnosis system, which system includes:

[0007] Data acquisition module, used to collect building electrical system parameter data in real time and comprehensively;

[0008] Load analysis module, used to conduct in-depth analysis of collected data using data analysis algorithms and mathematical models;

[0009] Equipment monitoring module, which is seamlessly integrated with the control system and management system of electrical equipment to monitor the operating status of electrical equipment in real time;

[0010] The model adjustment module, based on the gradient descent algorithm, automatically and quickly adjusts the parameters and algorithms of the diagnostic model according to the information provided by the load analysis module and the equipment monitoring module;

[0011] Fault diagnosis module, which integrates multiple advanced fault diagnosis methods, including rule-based diagnosis, model-based diagnosis and data-driven diagnosis, to conduct all-round and multi-level fault diagnosis of electrical systems;

[0012] The alarm and display module is used to send out an alarm signal in time when a fault is diagnosed, and to display the fault information on the display interface in a clear and intuitive manner.

[0013] Furthermore, the data acquisition module collects real-time electrical parameter data of current, voltage, and power, collects power quality-related data of frequency and phase angle, and records equipment operation time by installing high-precision sensors at key nodes of the building electrical system;

[0014] The data acquisition module has a collection frequency of no more than once every 5 seconds, and the data accuracy is better than ±0.5% of the sensor.

[0015] Furthermore, the load analysis module can calculate the real-time load, average load, load change rate, load periodicity, volatility and trend characteristics of the electrical system based on the data collected and transmitted by the data acquisition module, and generate a load analysis report, with the report update cycle not exceeding 5 minutes;

[0016] The equipment monitoring module monitors the operating status of electrical equipment in real time, including key information such as the equipment's service life, maintenance records, model changes, operating temperature, vibration frequency, etc., to build a health file for the equipment, with the file being updated no less than once a day.

[0017] Furthermore, the model adjustment module, based on the gradient descent algorithm, automatically adjusts the parameters and algorithms of the diagnosis model according to the real-time information provided by the load analysis module and the equipment monitoring module, and the adjustment response time does not exceed 1 minute.

[0018] Furthermore, the various fault diagnosis methods in the fault diagnosis module include, but are not limited to, rule-based, model-based, and data-driven diagnosis;

[0019] When the fault diagnosis module determines the fault type, it locates the fault location, sends an alarm signal through the alarm and display module according to the fault judgment result, and displays the fault information on the display interface;

[0020] The sending delay of the alarm signal shall not exceed 5 seconds, and the updating frequency of the display interface shall not be less than once every 5 seconds.

[0021] Furthermore, the load analysis module calculates the instantaneous power, average load and load change rate after cleaning and screening the data collected by the data collection module, wherein the instantaneous power P(t):

[0022] P(t)=U(t)×I(t)×cosθ(t)

[0023] Among them, P(t) represents instantaneous power, U(t) represents instantaneous voltage, I(t) represents instantaneous current, and cosθ(t) represents power factor;

[0024] The real-time load power at each moment is calculated based on the instantaneous power, and the average load over a period of time is calculated by the time integration method:

[0025]

[0026] in, represents the average load, T is the time interval for calculating the average load,

[0027] Based on the average load calculated for a period of time, the load change rate is calculated:

[0028]

[0029] Among them, P current Indicates the current load, P previous Indicates the load at the previous moment.

[0030] 8. A building electrical system fault diagnosis system according to claim 7, characterized in that the load fluctuation of the load analysis module is determined by calculating the load standard deviation, and the load standard deviation is:

[0031]

[0032] Where N is the number of data points, P i is the load value for each data point, To average the load, evaluate the load volatility;

[0033] Use linear regression method to fit load data and get the load trend line to determine whether the load is increasing, decreasing or remaining stable.

[0034] Furthermore, the load analysis module determines the load stability by establishing a linear regression model, wherein the linear regression model includes:

[0035] P=α+βt

[0036] Where P is the predicted load value, t is the time variable, α is the intercept, and β is the slope;

[0037] The parameters α and β are estimated by the least squares method to minimize the residual sum of squares. The calculation formula for the residual sum of squares is: Calculate the partial derivatives of S with respect to α and β respectively, and set the partial derivatives to 0 to obtain the estimated values ​​of the parameters α and β;

[0038] Substitute the obtained α and β into the linear regression model to determine the load trend. If β>0, the load shows an upward trend; if β<0, the load shows a downward trend; and if β=0, the load remains stable.

[0039] Furthermore, the load change rate can determine the device update status, and the model algorithm is adjusted according to the load change and device update status through the model adjustment module. When the load increases significantly, an adaptive threshold adjustment algorithm is used. For device updates, a model reconstruction algorithm based on device characteristics is used;

[0040] Adaptive threshold adjustment, the original overcurrent protection threshold is I th , the new threshold I new Adjust according to load change rate: I new =I th ×(1+k×ΔP%), where k is the adjustment coefficient;

[0041] Model reconstruction based on device characteristics: for newly connected devices and changes in key device parameters, obtain the characteristic parameters of the new devices and rebuild the relevant parts of the fault diagnosis model based on the new characteristic parameters;

[0042] The adjusted model parameters are optimized using the gradient descent optimization algorithm, and the objective function is to minimize the diagnostic error, that is, Where yi is the actual fault status, is the fault state predicted by the model;

[0043] Use a portion of new data to verify the adjusted model, calculate the accuracy and recall rate indicators, verify that the results meet the requirements, and update the parameters and algorithms of the diagnostic model.

[0044] Furthermore, the fault diagnosis module performs fault judgment based on rules, models, and data drive, and includes:

[0045] Rule-based diagnosis uses pre-set rules to make preliminary judgments;

[0046] Model-based diagnosis, substitute the input data into the diagnostic model, calculate the fault index, and for short-circuit fault diagnosis, use the short-circuit current calculation formula Where U is the voltage, Z is the short-circuit impedance, and the calculated fault index is compared with the preset threshold to determine whether there is a fault;

[0047] Based on data-driven diagnosis, cluster analysis methods are used to classify historical fault data and current data. If the current data point is similar to the known fault cluster characteristics, it is judged that the corresponding fault exists;

[0048] According to the diagnosis results of the fault diagnosis model, a weighted fusion method is used to determine the final fault diagnosis conclusion, and the weights of the three methods are defined as w1, w2, and w3, respectively. The final diagnosis score S = w1S1+w2S2+w3S3, where S1, S2, and S3 are the diagnosis scores of the three methods respectively;

[0049] According to the diagnosis results, the type of fault is determined, and the specific location of the fault is determined by using the circuit analysis methods of node voltage method and branch current method combined with the distribution location of sensors;

[0050] By analyzing the fault current size, voltage change amplitude, and equipment temperature rise parameters and rates, the severity of the fault is evaluated, and the development trend of the fault is predicted based on the ARIMA model.

[0051] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0052] The present invention uses the synergistic effect of the data acquisition module and the load analysis module, the equipment monitoring module, and the model adjustment module to analyze the load data in real time to monitor load changes while monitoring the electrical system parameters, and uses the model adjustment module to automatically optimize the diagnosis model according to load changes and equipment updates, thereby significantly improving the accuracy and reliability of fault diagnosis, accurately identifying real faults, avoiding misjudgments and missed judgments, and effectively reducing operation and maintenance costs, reducing the risk of equipment damage, and effectively avoiding safety accidents through the accurate judgment of the fault diagnosis module and the timely notification of the alarm and display module. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A schematic diagram of the structure of a building electrical system fault diagnosis system according to an embodiment of the present invention;

[0054] Figure 2 A diagnostic flow chart of a building electrical system fault diagnosis method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0056] like Figure 1 As shown, an embodiment of the present invention provides a building electrical system fault diagnosis system, the system comprising:

[0057] The data acquisition module is used to collect the electrical system parameter data of the building construction stage in real time and comprehensively, including but not limited to current, voltage, power, frequency, phase angle, temperature, humidity and peak power consumption time. The acquisition frequency is not less than once every 5 seconds, and the data accuracy is better than ±0.5%. It adopts advanced sensor technology and high-speed data acquisition equipment to comprehensively and real-timely collect various parameters of the building electrical system. These parameters include not only basic electrical quantities such as current, voltage, power, but also environmental and operating characteristics such as temperature, humidity, and operating time of the equipment and cables nearby;

[0058] For example, high-precision current transformers and smart meters are installed in the dewatering distribution cabinet of the construction pit to collect current and voltage data once a minute. At the same time, temperature and humidity sensors and water flow sensors are installed on the main drainage pipe to collect temperature, humidity and water flow data once a minute.

[0059] The load analysis module is used to use data analysis algorithms and mathematical models to conduct in-depth analysis of the collected data, calculate the real-time load, average load, load change rate, load periodicity, volatility and trend characteristics of the electrical system, and generate a load analysis report with a report update cycle of no more than 5 minutes;

[0060] For example, for the electrical system of a construction site, by analyzing the load variation patterns in different stages of foundation pit, main body and decoration, it is found that the loads in different stages, work areas and processes show obvious periodicity, which is less deviated from the electricity construction organization design. The feasibility of the plan is reversely verified, and lean construction is continuously corrected and improved to achieve the energy-saving and material-saving requirements of green construction from the source.

[0061] The equipment monitoring module is seamlessly integrated with the control system and management system of electrical equipment to monitor the operating status of electrical equipment in real time, including key information such as the service life of the equipment, maintenance records, model changes, operating temperature, vibration frequency, etc., and to build a health file for the equipment. The file is updated at least once a day. By closely integrating with the control system and management system of electrical equipment, the operating status and key parameters of the equipment can be monitored in real time.

[0062] Taking the tower crane distribution box as an example, the system monitors its operation start and end time, maintenance time and active power changes in real time, and records the information of parts replaced during each maintenance.

[0063] The model adjustment module, based on the gradient descent algorithm, automatically and quickly adjusts the parameters and algorithms of the diagnostic model according to the information provided by the load analysis module and the equipment monitoring module. The adjustment response time does not exceed 1 minute, ensuring that the diagnostic model always matches the real-time situation of the electrical system. Based on the gradient descent algorithm, it automatically adjusts the parameters and algorithms of the diagnostic model according to the real-time information provided by the load analysis module and the equipment monitoring module. For example, when summer comes, the load of the air-conditioning system increases significantly, and the model adjustment module will accordingly increase the threshold of the overcurrent protection to avoid false alarms. When a device is updated, the module will re-optimize the relevant parameters of the diagnostic model according to the characteristics of the new device.

[0064] The fault diagnosis module is used to integrate a variety of advanced fault diagnosis methods, including rule-based diagnosis, model-based diagnosis and data-driven diagnosis, to conduct all-round and multi-level fault diagnosis of the electrical system, determine the type, location, severity and development trend of the fault, and the accuracy of the diagnosis result is not less than 95%. It can accurately determine the type of fault (such as short circuit, open circuit, overload, leakage, etc.), accurately locate the fault location, and evaluate the severity and development trend of the fault. For example, when the system detects a sudden increase in current and lasts for a long time, combined with the load characteristics and historical data of the equipment, it is diagnosed as a possible overload fault and quickly locates the specific line or equipment.

[0065] The alarm and display module is used to send out alarm signals in time through various channels when a fault is diagnosed, such as sound and light alarms, text message notifications, emails, etc., and display the fault information on the display interface in a clear and intuitive manner, including detailed information such as the type, location, severity, and occurrence time of the fault. At the same time, the historical data and trend analysis of the fault are displayed in the form of charts and reports. The sending delay of the alarm signal does not exceed 5 seconds, and the display interface is updated at least once every 5 seconds. When the fault diagnosis module confirms that a fault has occurred, it can send out alarm signals in time through various channels and display the fault information to the user in an intuitive and clear manner. For example, the fault location and type are displayed in eye-catching colors on the large screen of the monitoring center, and text message notifications are sent to the mobile phones of relevant maintenance personnel. A detailed fault report pops up on the system interface, including the time of the fault, cause analysis, and recommended treatment measures.

[0066] Specifically, in order to ensure the accuracy of the data collected by the data acquisition module, high-precision current sensors, voltage sensors and power sensors are installed at key nodes of the building electrical system, including incoming line cabinets, distribution boxes, and front-end locations of important electrical equipment. At the same time, temperature sensors and humidity sensors are arranged around the electrical equipment and in the environment to obtain environmental parameters. The data acquisition frequency is set to once per second to capture the rapid changes in electrical system parameters in a timely manner, collect electrical parameter data of current, voltage, and power in real time, collect power quality-related data such as frequency and phase angle, and record equipment operating time to evaluate equipment fatigue and potential failures. In order to reduce the risk of failure, sensors with an accuracy better than ±0.5% are used to ensure that the collected data are accurate and reliable. Sensors are calibrated and maintained regularly to maintain data accuracy. Preliminary filtering and noise reduction processing is performed on the data at the collection end to remove data interference and outliers. The collected data is quickly and stably transmitted to the subsequent load analysis module and central database through wired and wireless communication methods. Through the above specific execution steps, the data acquisition module can provide comprehensive, accurate and real-time basic data support for the entire building electrical system fault diagnosis system, laying a solid foundation for subsequent load analysis, equipment monitoring, fault diagnosis and other links.

[0067] The load analysis module receives the parameter data of current, voltage, power, frequency, phase angle, temperature, humidity and equipment operation time collected in real time from the data acquisition module, cleans and filters the received data, removes invalid and abnormal data points, uses the moving average filtering method to smooth the data, reduces noise interference, and calculates the instantaneous power P(t):

[0068] P(t)=U(t)×I(t)×cosθ(t)

[0069] Among them, P(t) represents instantaneous power, U(t) represents instantaneous voltage, I(t) represents instantaneous current, and cosθ(t) represents power factor;

[0070] Furthermore, the real-time load power at each moment is calculated based on the instantaneous power, and the average load over a period of time is calculated by the time integration method:

[0071]

[0072] in, represents the average load, T is the time interval for calculating the average load,

[0073] Furthermore, the load change rate can be calculated based on the average load in a certain period of time:

[0074]

[0075] Among them, Pcurrent Indicates the current load, P previous Indicates the load at the previous moment.

[0076] Furthermore, the fast Fourier transform (FFT) algorithm is used to convert the load time series into the frequency domain, analyze its spectrum characteristics, determine whether there is a periodic component, and calculate the standard deviation of the load:

[0077]

[0078] Where N is the number of data points, P i is the load value for each data point, is the average load;

[0079] Evaluate the volatility of the load, use the linear regression method to fit the load data, get the trend line of the load, and determine whether the load is rising, falling, or remaining stable. Collect the load data over a period of time, recorded as {P1, P2, ..., P n}, where P i represents the load value at the i-th moment;

[0080] A linear regression model is established, and the form of the linear regression model is:

[0081] P=α+βt

[0082] Where P is the predicted load value, t is the time variable, α is the intercept, and β is the slope;

[0083] Specifically, the least squares method is used to estimate the parameters α and β so that the residual sum of squares is minimized. The calculation formula for the residual sum of squares is:

[0084]

[0085] By taking partial derivatives of S with respect to α and β respectively and setting the partial derivatives to 0, we get the following system of equations:

[0086]

[0087] By solving the equation group, the estimated values ​​of parameters α and β are obtained, and the obtained α and β are substituted into the linear regression model to obtain the load trend line equation P = α + βt to determine the load trend;

[0088] If β>0, the load is on the rise, if β<0, the load is on the decline, and if β=0, the load remains stable;

[0089] Through the above specific process, the linear regression algorithm can be used to effectively fit the load data, obtain the load trend line, and accurately judge the load change trend. The above calculation and analysis results are organized into a load analysis report, including real-time load value, average load value, load change rate, periodicity, volatility and trend characteristics. Detailed description and chart display, the report update cycle does not exceed 5 minutes.

[0090] The equipment monitoring module establishes a communication connection with the control system and management system of the electrical equipment to obtain basic information of the equipment, such as model, specification, rated parameters, etc., and continuously receives real-time data during the operation of the equipment, including but not limited to current, voltage, power, temperature, vibration, etc., detects and removes abnormal values ​​of the received data, and calculates the service life of the equipment according to the time when the equipment was put into use and the current time:

[0091] Service life = current time - time put into use,

[0092] Establish a maintenance event database to record the time, reason, maintenance content and equipment status after each maintenance. Compare the initial model of the equipment with the current model. If there is any inconsistency, mark it as a model change.

[0093] The device monitoring module continuously monitors the operating temperature of the device as T. When T> the set temperature threshold T th When an abnormal temperature alarm is issued, the collected vibration frequency data is subjected to fast Fourier transform (FFT), the time domain signal is converted into a frequency domain signal, and the main frequency components of the vibration are analyzed. If the main frequency components deviate from the normal range, it indicates that the equipment has a fault. Combining the above monitoring and analysis results, a detailed health file is built for each device, including basic equipment information, operation history, maintenance records, and status evaluation.

[0094] The model adjustment module receives load characteristic data, such as real-time load, average load, load change rate, etc., through the load analysis module, and obtains equipment status information from the equipment monitoring module, including service life, maintenance records, model changes, etc., to check whether the load change exceeds a preset threshold.

[0095] When the load change rate ΔP% is greater than the set threshold ΔP th %, it is considered that the load has changed significantly, and it is determined whether the equipment has been updated, such as the connection of new equipment or the change of key parameters of old equipment. According to the load change and equipment update, the adjustment algorithm is selected. The load increases significantly, and the adaptive threshold adjustment algorithm is adopted. For equipment updates, the model reconstruction algorithm based on equipment characteristics is used, and the adaptive threshold is adjusted (taking the overcurrent protection threshold adjustment as an example). The new threshold is adjusted according to the load change rate:

[0096] I new =Ith ×(1+k×ΔP%)

[0097] Among them, I new is the new threshold, I th is the original over-current protection threshold, k is the adjustment coefficient,

[0098] Model reconstruction based on device characteristics, for newly connected devices and changes in key device parameters, obtain the characteristic parameters of the new device, such as rated current I rated , power factor cosφ, etc., and reconstruct the relevant parts of the fault diagnosis model according to the new characteristic parameters. Use the gradient descent optimization algorithm to optimize the adjusted model parameters. The objective function is to minimize the diagnosis error, that is:

[0099]

[0100] Among them, yi is the actual fault state, For the fault state predicted by the model, a part of the new data is used to verify the adjusted model, and the accuracy and recall rate indicators are calculated. The verification results meet the requirements, and the parameters and algorithms of the diagnosis model are updated to put it into practical use.

[0101] The fault diagnosis module receives data from the data acquisition module, the load analysis module, the equipment monitoring module and the model adjustment module, including electrical parameters, load characteristics, equipment status and adjusted diagnostic model parameters, and makes a preliminary judgment based on rule-based diagnosis by applying pre-set rules;

[0102] If the current I exceeds the overcurrent protection threshold I th , it is marked as a possible overcurrent fault. Based on the model diagnosis, the input data is substituted into the diagnosis model to calculate the fault index. For short-circuit fault diagnosis, the short-circuit current is calculated:

[0103]

[0104] Among them, U is voltage, Z is short-circuit impedance, and the calculated fault index is compared with the preset threshold to determine whether there is a fault. Based on data-driven diagnosis, the cluster analysis method is used to classify historical fault data and current data. Assuming that the K-Means clustering algorithm is used, the goal is to divide the data into K clusters, and update the cluster center through continuous iteration so that the sum of the distances from the data point to the center of the cluster to which it belongs is minimized. The current data point is similar to the known fault cluster characteristics, and it is determined that the corresponding fault exists.

[0105] Combining the diagnosis results based on rules, models and data-driven, the weighted fusion method is used to determine the final fault diagnosis conclusion. The weights of the three methods are defined as w1, w2, and w3 respectively. The final diagnosis score S = w1S1+w2S2+w3S3.

[0106] Among them, S1, S2, and S3 are the diagnosis scores of the three methods respectively. According to the diagnosis results, the type of fault is clarified, such as short circuit, open circuit, overload, leakage, etc. The circuit analysis methods of node voltage method and branch current method are used, combined with the distribution position of sensors, to determine the specific location of the fault. By analyzing the fault current size, voltage change amplitude and equipment temperature rise parameters, the severity of the fault is evaluated. Based on the ARIMA model, the development trend of the fault is predicted.

[0107] The alarm and display module is used to receive the fault information output by the fault diagnosis module, including the fault type, location, severity, occurrence time, etc., and make a judgment according to the preset alarm rules. When the severity of the fault exceeds the set severity level threshold S th , then an alarm is triggered. According to the urgency of the fault and user settings, the alarm mode is selected, including sound and light alarm, SMS notification and email. For sound and light alarm, different sound frequencies and light flashing frequencies can be set to indicate different fault types and severity. Detailed alarm information is constructed, including fault description and recommended treatment measures. The fault information is displayed in real time on the display interface in the form of intuitive charts and reports; for example, a bar chart is used to show the frequency of occurrence of different types of faults, and a line chart is used to show the changing trend of the fault severity. The alarm and fault information are stored in the database for subsequent query and analysis.

[0108] like Figure 2 As shown, another embodiment of the present invention provides a building electrical system fault diagnosis method, which includes:

[0109] S100: Complete parameter setting and initialization of each module, and connect it to the equipment control system and management system in the electrical system;

[0110] S200: Setting a data acquisition module and a load analysis module, processing the data parameters of the real-time collected electrical system through a load analysis model, and forming a load analysis report;

[0111] S300: Build equipment health files and synchronously obtain the operating status and key parameter information of electrical equipment through the equipment monitoring module;

[0112] S400: Setting a model adjustment module and automatically adjusting the diagnostic model and parameter algorithm according to the load analysis report and equipment monitoring results;

[0113] S500: Use a variety of diagnostic methods to determine fault type, location, severity and development trend. Receive diagnostic results through the alarm and display module, trigger an alarm and display it on the interface after judging according to preset rules.

[0114] Specifically, the data acquisition module collects electrical parameter data of current, voltage, and power in real time at a preset frequency (not less than once every 5 seconds), collects power quality-related data of frequency and phase angle, and records equipment operation time to evaluate equipment fatigue and potential failure risks; for example, in the office distribution box (room), the current transformer collects current data once every 5 seconds, and the temperature and humidity sensor collects ambient temperature and humidity data once a minute.

[0115] The load analysis module calculates the instantaneous power P(t) of the equipment according to the equipment parameter data collected from the data acquisition module:

[0116] P(t)=U(t)×I(t)×cosθ(t)

[0117] Among them, P(t) represents instantaneous power, U(t) represents instantaneous voltage, I(t) represents instantaneous current, and cosθ(t) represents power factor;

[0118] Furthermore, the real-time load power at each moment is calculated based on the instantaneous power, and the average load over a period of time is calculated by the time integration method:

[0119]

[0120] in, represents the average load, T is the time interval for calculating the average load,

[0121] Furthermore, the load change rate can be calculated based on the average load in a certain period of time:

[0122]

[0123] Among them, P current Indicates the current load, P previous Indicates the load at the previous moment.

[0124] Furthermore, the fast Fourier transform (FFT) algorithm is used to convert the load time series into the frequency domain, analyze its spectrum characteristics, determine whether there is a periodic component, and calculate the standard deviation of the load:

[0125]

[0126] Where N is the number of data points, P i is the load value for each data point, is the average load;

[0127] Evaluate the volatility of the load, use the linear regression method to fit the load data, get the trend line of the load, and determine whether the load is rising, falling, or remaining stable. Collect the load data over a period of time, recorded as {P1, P2, ..., Pn}, where P i represents the load value at the i-th moment;

[0128] A linear regression model is established, and the form of the linear regression model is:

[0129] P=α+βt

[0130] Where P is the predicted load value, t is the time variable, α is the intercept, and β is the slope;

[0131] Specifically, the least squares method is used to estimate the parameters α and β so that the residual sum of squares is minimized. The calculation formula for the residual sum of squares is:

[0132]

[0133] By taking partial derivatives of S with respect to α and β respectively and setting the partial derivatives to 0, we get the following system of equations:

[0134]

[0135] By solving the equation group, the estimated values ​​of parameters α and β are obtained, and the obtained α and β are substituted into the linear regression model to obtain the load trend line equation P = α + βt to determine the load trend;

[0136] If β>0, the load is on the rise, if β<0, the load is on the decline, and if β=0, the load remains stable;

[0137] Through the above specific process, the linear regression algorithm can be used to effectively fit the load data, obtain the load trend line, and accurately judge the load change trend. The above calculation and analysis results are organized into a load analysis report, including real-time load value, average load value, load change rate, periodicity, volatility and trend characteristics. Detailed description and chart display, the report update cycle does not exceed 5 minutes.

[0138] The equipment monitoring module synchronously obtains the operating status and key parameter information of the electrical equipment, including service life, maintenance records, model changes, operating temperature, etc., and builds and updates the health record of each device. The model adjustment module determines whether it is necessary to adjust the parameters and algorithms of the diagnostic model based on the results of load analysis and equipment monitoring. If there is a significant change in the load (such as a sharp increase in load due to a deadline) or there is a large-scale equipment update, corresponding adjustments are automatically made, for example, the overcurrent protection threshold is increased to adapt to the increase in load.

[0139] The fault diagnosis module uses a variety of diagnostic methods to determine the fault type (such as short circuit, open circuit, overload, leakage, etc.), locate the fault location, and evaluate the severity and development trend of the fault based on the adjusted diagnostic model and input data. Once the fault is diagnosed, the alarm and display module responds immediately and sends an alarm signal to relevant personnel through various methods such as sound and light alarms, text message notifications, and e-mails. The fault information is displayed on the display interface in the form of intuitive charts, reports, etc., including detailed content such as fault type, location, severity, and occurrence time.

[0140] The system runs continuously, repeating the above process to monitor the building electrical system in real time. Based on new data and situations, it further optimizes the diagnostic models and parameters to improve the accuracy and reliability of the diagnosis. The entire workflow forms a dynamic loop, adapting to changes in the building electrical system in real time, ensuring timely and accurate diagnosis of faults, and ensuring the safe and stable operation of the system.

[0141] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A building electrical system fault diagnosis system, characterized in that: include: Data acquisition module, used to collect building electrical system parameter data in real time and comprehensively; Load analysis module, used to conduct in-depth analysis of collected data using data analysis algorithms and mathematical models; Equipment monitoring module, which is seamlessly integrated with the control system and management system of electrical equipment to monitor the operating status of electrical equipment in real time; The model adjustment module, based on the gradient descent algorithm, automatically and quickly adjusts the parameters and algorithms of the diagnostic model according to the information provided by the load analysis module and the equipment monitoring module; Fault diagnosis module, which integrates multiple advanced fault diagnosis methods, including rule-based diagnosis, model-based diagnosis and data-driven diagnosis, to conduct all-round and multi-level fault diagnosis of electrical systems; The alarm and display module is used to send out an alarm signal in time when a fault is diagnosed, and to display the fault information on the display interface in a clear and intuitive manner.

2. A building electrical system fault diagnosis system according to claim 1, characterized in that: The data acquisition module collects real-time electrical parameter data of current, voltage, and power, collects power quality data related to frequency and phase angle, and records equipment operation time by installing high-precision sensors at key nodes of the building electrical system; The data acquisition module has a collection frequency of no more than once every 5 seconds, and the data accuracy is better than ±0.5% of the sensor.

3. A building electrical system fault diagnosis system according to claim 1, characterized in that: The load analysis module can calculate the real-time load, average load, load change rate, load periodicity, volatility and trend characteristics of the electrical system based on the data collected and transmitted by the data acquisition module, and generate a load analysis report, with a report update cycle of no more than 5 minutes; The equipment monitoring module monitors the operating status of electrical equipment in real time, including key information such as the equipment's service life, maintenance records, model changes, operating temperature, vibration frequency, etc., to build a health file for the equipment, with the file being updated no less than once a day.

4. A building electrical system fault diagnosis system according to any one of claims 1 to 3, characterized in that: The model adjustment module, based on the gradient descent algorithm, automatically adjusts the parameters and algorithms of the diagnosis model according to the real-time information provided by the load analysis module and the equipment monitoring module, and the adjustment response time does not exceed 1 minute.

5. A building electrical system fault diagnosis system according to claim 4, characterized in that: The various fault diagnosis methods in the fault diagnosis module include but are not limited to rule-based, model-based, and data-driven diagnosis; When the fault diagnosis module determines the fault type, it locates the fault location, sends an alarm signal through the alarm and display module according to the fault judgment result, and displays the fault information on the display interface; The sending delay of the alarm signal shall not exceed 5 seconds, and the updating frequency of the display interface shall not be less than once every 5 seconds.

6. A building electrical system fault diagnosis system according to claim 5, characterized in that: The load analysis module calculates the instantaneous power, average load and load change rate after cleaning and screening the data collected by the data acquisition module, where the instantaneous power P(t) is: P(t)=U(t)×I(t)×cosθ(t) Among them, P(t) represents instantaneous power, U(t) represents instantaneous voltage, I(t) represents instantaneous current, and cosθ(t) represents power factor; The real-time load power at each moment is calculated based on the instantaneous power, and the average load over a period of time is calculated by the time integration method: in, represents the average load, T is the time interval for calculating the average load, Based on the average load calculated for a period of time, the load change rate is calculated: Among them, P current Indicates the current load, P previous Indicates the load at the previous moment.

7. A building electrical system fault diagnosis system according to claim 6, characterized in that: The load fluctuation of the load analysis module is determined by calculating the load standard deviation, which is: Where N is the number of data points, P i is the load value for each data point, To average the load, evaluate the load volatility; Use linear regression method to fit load data and get the load trend line to determine whether the load is increasing, decreasing or remaining stable.

8. A building electrical system fault diagnosis system according to claim 7, characterized in that: The load analysis module determines load stability by establishing a linear regression model, and the linear regression model includes: P=α+βt Where P is the predicted load value, t is the time variable, α is the intercept, and β is the slope; The parameters α and β are estimated by the least squares method to minimize the residual sum of squares. The calculation formula for the residual sum of squares is: Calculate the partial derivatives of S with respect to α and β respectively, and set the partial derivatives to 0 to obtain the estimated values ​​of the parameters α and β; Substitute the obtained α and β into the linear regression model to determine the load trend. If β>0, the load shows an upward trend; if β<0, the load shows a downward trend; and if β=0, the load remains stable.

9. A building electrical system fault diagnosis system according to any one of claims 6 to 8, characterized in that: The load change rate can determine the device update status. The model algorithm is adjusted according to the load change and device update status through the model adjustment module. If the load increases significantly, an adaptive threshold adjustment algorithm is used. For device updates, a model reconstruction algorithm based on device characteristics is used; Adaptive threshold adjustment, the original overcurrent protection threshold is I th , the new threshold I new Adjust according to load change rate: I new =I th ×(1+k×ΔP%), where k is the adjustment coefficient; Model reconstruction based on device characteristics: for newly connected devices and changes in key device parameters, obtain the characteristic parameters of the new devices and rebuild the relevant parts of the fault diagnosis model based on the new characteristic parameters; The adjusted model parameters are optimized using the gradient descent optimization algorithm, and the objective function is to minimize the diagnostic error, that is, where y i is the actual fault state, is the fault state predicted by the model; Use a portion of new data to verify the adjusted model, calculate the accuracy and recall indicators, verify that the results meet the requirements, and update the parameters and algorithms of the diagnostic model.

10. A building electrical system fault diagnosis system according to claim 9, characterized in that: The fault diagnosis module performs fault diagnosis based on rules, models, and data drive, and includes: Rule-based diagnosis uses pre-set rules to make preliminary judgments; Model-based diagnosis, substitute the input data into the diagnostic model, calculate the fault index, and for short-circuit fault diagnosis, use the short-circuit current calculation formula Where U is the voltage, Z is the short-circuit impedance, and the calculated fault index is compared with the preset threshold to determine whether there is a fault; Based on data-driven diagnosis, cluster analysis methods are used to classify historical fault data and current data. If the current data point is similar to the known fault cluster characteristics, it is judged that the corresponding fault exists; According to the diagnosis results of the fault diagnosis model, a weighted fusion method is used to determine the final fault diagnosis conclusion, and the weights of the three methods are defined as w1, w2, and w3, respectively. The final diagnosis score S = w1S1+w2S2+w3S3, where S1, S2, and S3 are the diagnosis scores of the three methods respectively; According to the diagnosis results, the type of fault is determined, and the specific location of the fault is determined by using the circuit analysis methods of node voltage method and branch current method combined with the distribution location of sensors; By analyzing the fault current size, voltage change amplitude, and equipment temperature rise parameters and rates, the severity of the fault is evaluated, and the development trend of the fault is predicted based on the ARIMA model.

Images