An intelligent early warning reasoning system based on multi-scenario analysis

Abstract

The present invention discloses an intelligent early warning reasoning system based on multi-scenario analysis, and specifically relates to the technical field of power supply early warning. The present invention constructs a power supply multi-scenario recognition model based on acquired power supply scenario historical data and performs performance evaluation on it, which not only realizes power supply multi-scenario recognition, but also improves the accuracy of power supply scenario recognition. The present invention sets an early warning reasoning module to receive the number of scenarios recognized in real time, the number of charging abnormality scenarios, the number of circuit failure scenarios, and the number of power supply abnormality scenarios, and infers the power supply out-of-control probability based on the received scenario recognition information. When the inferred power supply out-of-control probability reaches a preset alarm threshold, an alarm is issued, and when the inferred power supply out-of-control probability does not reach the preset alarm threshold, no processing is performed, which can effectively reduce the possibility of power supply system out-of-control or the risks and losses caused by power supply system out-of-control, and to a certain extent, ensure the power supply quality.

Description

Technical Field

[0001] The present invention relates to the field of power supply early warning technology, and more specifically, to an intelligent early warning reasoning system based on multi-scenario analysis. Background Art

[0002] As an essential energy source for human activities, electricity plays an important role in people's lives and work. With the increase in population and the sharp increase in electricity consumption in the new era, the conventional electricity analysis system can no longer meet the needs of the rapidly developing economy and society. The improvement and innovation of the intelligent power supply early warning system is urgently needed.

[0003] The existing power supply management system collects current, voltage and power data in real time, compares the collected data with the preset data in the power supply plan to determine whether there is any abnormality. If there is an abnormality, the possibility of the source of the abnormality is analyzed, and abnormal maintenance is performed based on the results of the analysis of the possibility of the source of the abnormality. Power supply abnormality diagnosis and maintenance can be carried out in time to avoid subsequent losses.

[0004] However, the above system still has some problems: the above system analyzes the possible sources of abnormalities after detecting abnormal data. In actual applications, there are multiple possibilities of abnormalities, and it is impossible to directly and accurately obtain the cause of the abnormality or it takes a lot of time to determine the real cause of the abnormality. The accuracy of abnormality identification still needs to be improved. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides an intelligent early warning reasoning system based on multi-scenario analysis to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: an intelligent early warning reasoning system based on multi-scenario analysis, comprising:

[0007] Power supply scenario data acquisition module: classifies power supply scenarios and obtains historical power supply cases of different power supply scenarios based on the Internet, and then obtains historical data of power supply scenarios;

[0008] Power supply multi-scenario recognition model construction module: Build a power supply multi-scenario recognition model based on the acquired power supply scenario historical data and evaluate its performance. If it passes the evaluation, it will be put into use. If it fails the evaluation, the model will be tuned.

[0009] Power data online monitoring module: uses sensors to monitor the power data of the power supply system online, and transmits the monitored power data to the scene recognition module in real time;

[0010] Scene recognition module: inputs the received power data into the power supply multi-scene recognition model for scene recognition and records the scene recognition results;

[0011] Early warning reasoning module: infers the probability of power outage based on the scene recognition results, and issues an alarm when the probability of power outage reaches a preset value;

[0012] Power supply maintenance module: arranges maintenance personnel to maintain the power supply system when an alarm is received, records maintenance data and calculates the maintenance efficiency coefficient based on the recorded maintenance data;

[0013] Power supply data regular collection module: used to regularly collect the expected energy consumption, actual energy consumption, maintenance times, scene recognition times, normal charging and power supply time, out-of-control time, and maintenance efficiency of each maintenance of the power supply system within a preset fixed period and send them to the power supply data processing module;

[0014] Power supply data processing module: processes the collected power supply data within a preset fixed period, and calculates the energy consumption excess coefficient, power supply stability coefficient, effective power supply coefficient, and average maintenance efficiency;

[0015] Power supply quality control capability index calculation module: calculates the power supply quality control capability index based on energy consumption excess coefficient, power supply stability coefficient, effective power supply coefficient, and average maintenance efficiency;

[0016] Comprehensive evaluation module: Compare the calculated power supply quality control capability index with the preset value to determine whether the calculated value meets the demand. If it does, continue to use the current system; if it does not, optimize the system;

[0017] Database: used to store data information of all modules in the system.

[0018] Preferably, the power supply scenario recognition model construction module includes a data receiving unit, a data annotation unit, a data preprocessing unit, a feature extraction unit, a data partitioning unit, a model construction unit, a model training unit, a model evaluation unit, and a model tuning unit, wherein the data receiving unit is used to receive the acquired power supply scenario historical data; the data annotation unit annotates the corresponding historical scenario data with the scenario category and data occurrence time of different power supply scenarios; the data preprocessing unit is used to perform timeline integration on the annotated historical scenario data, and remove outliers after the integration is completed; the feature extraction unit compares different historical scenario data of the same power supply scheme, and respectively screens the parameters that are inconsistent with the normal charging and power supply scenario data for charging abnormality, circuit failure, and discharge abnormality scenario data, and screens all power supply schemes. The selected parameters are integrated, and the inconsistent parameters common to all circuit fault scenarios are obtained as circuit fault characteristic parameters, the inconsistent parameters common to all charging abnormality scenarios are charging abnormality characteristic parameters, and the inconsistent parameters common to all power supply abnormality scenarios are power supply abnormality characteristic parameters; the data division unit divides the sample data into a training data set and a test data set by a cross-validation method; the model construction unit selects a network architecture and designs a network structure and the number of layers based on the characteristics of the historical scenario characteristic parameters used and the model construction requirements, and then constructs a power supply multi-scenario recognition convolutional neural network model; the model training unit uses the training data set to train the power supply multi-scenario recognition convolutional neural network model; the model evaluation unit uses the test data set to evaluate the performance of the trained model and records the number of samples in the test data set and the number of samples correctly identified And calculate the power supply scene recognition accuracy , the specific formula is: The model tuning unit adjusts the weight and bias of the model through the back propagation algorithm and the optimizer when the calculated power supply scenario recognition accuracy is lower than the preset value, and confirms the use of the model when the calculated power supply scenario recognition accuracy is greater than or equal to the preset value.

[0019] Preferably, the scene recognition module includes a monitoring data receiving unit, a power supply plan entry unit, a monitoring data entry unit, a scene recognition unit, a scene recognition result recording unit, and a recognition result output unit, wherein the monitoring data receiving unit is used to receive monitoring data; the power supply plan entry unit is used to enter the formulated normal charging plan and normal power supply plan; the monitoring data entry unit is used to enter the real-time monitored power data; the scene recognition unit performs scene recognition based on the entered power supply plan and monitoring data; the scene recognition result recording unit is used to record real-time scene recognition results, including the number of recognized scenes, the number of charging abnormality scenes, the number of circuit failure scenes, and the number of power supply abnormality scenes; the recognition result output unit is used to display the recognized scenes to the early warning reasoning module.

[0020] Preferably, the early warning reasoning module includes a data receiving unit, a power outage probability reasoning unit, and an intelligent early warning unit. The data receiving unit is used to receive the number of scenes identified in real time, the number of charging abnormalities, the number of circuit failure scenes, and the number of power supply abnormalities; the power outage probability reasoning unit is based on the number of scenes identified in real time. , Number of abnormal charging scenarios , Number of circuit failure scenarios , and the number of power supply abnormality scenarios Inference of the probability of power supply failure , the specific formula is: , , is the proportionality coefficient, The intelligent early warning unit issues an alarm when the inferred probability of power outage reaches a preset alarm threshold, and does not perform any processing when the inferred probability of power outage does not reach the preset alarm threshold.

[0021] Preferably, the power supply maintenance module includes an alarm receiving unit, a maintenance task allocation unit, a maintenance progress real-time update unit, a maintenance data recording unit, a maintenance efficiency coefficient calculation unit, and a data output unit. The alarm receiving unit is used to receive alarm information; the maintenance task allocation unit sends the power supply maintenance task to all maintenance personnel terminals on duty after receiving the alarm information, and the maintenance personnel who first confirms to receive the maintenance task is the target maintenance personnel; the maintenance progress real-time update unit updates the maintenance progress in real time after the target maintenance personnel confirms to receive the maintenance task; the maintenance data recording unit is used to record the power supply outage probability before and after maintenance and the maintenance time required to reduce the power supply outage probability to below the alarm threshold; the maintenance efficiency coefficient calculation unit is based on the power supply outage probability before maintenance , Probability of power supply outage after maintenance is completed and the maintenance time required to reduce the probability of power outage below the alarm threshold Calculating the maintenance efficiency factor The specific formula is: ; The data output unit sends the calculated maintenance efficiency coefficient to the power supply data processing module.

[0022] Preferably, the power supply data processing module includes a data receiving unit, an energy consumption excess coefficient calculation unit, a power supply stability coefficient calculation unit, an effective power supply coefficient calculation unit, an average maintenance efficiency calculation unit, and a data output unit. The data receiving unit is used to receive the data collected by the power supply data periodic collection module; the energy consumption excess coefficient calculation unit is based on the expected energy consumption within a preset fixed period. and actual energy consumption Calculate the excess energy consumption factor The specific formula is: The power supply stability coefficient calculation unit is based on the number of maintenance times within a preset fixed period. and scene recognition times Calculate the power supply stability factor The specific formula is: , e is a natural constant; the effective power supply coefficient calculation unit is based on the normal charging and power supply time within a preset fixed period and duration of loss of control Calculate the effective power supply factor The specific formula is: The average maintenance efficiency calculation unit is based on the number of maintenance times within a preset fixed period. and the maintenance efficiency of the ith maintenance Calculate the average maintenance efficiency factor The specific formula is: ; The data output unit is used to send the calculated indicators to the power supply quality control capability index calculation module.

[0023] Preferably, the power quality control capability index calculation module is based on the energy consumption excess coefficient , Power supply stability coefficient , Effective power supply coefficient , and the average maintenance efficiency Calculate the power supply quality control capability index The specific formula is: , e is a natural constant.

[0024] Technical effects and advantages of the present invention:

[0025] 1. The present invention sets a power supply multi-scenario recognition model construction module to construct a power supply multi-scenario recognition model based on the acquired power supply scenario historical data and conducts a performance evaluation on it. If the evaluation is passed, the model is put into use. If the evaluation is not passed, the model is tuned. This not only realizes the recognition of power supply multiple scenarios, but also improves the accuracy of power supply scenario recognition. The present invention sets an early warning reasoning module to receive the number of scenarios recognized in real time, the number of charging abnormality scenarios, the number of circuit failure scenarios, and the number of power supply abnormality scenarios, and infers the power supply out-of-control probability based on the received scenario recognition information. When the inferred power supply out-of-control probability reaches a preset alarm threshold, an alarm is issued. When the inferred power supply out-of-control probability does not reach the preset alarm threshold, no processing is performed. This can effectively reduce the possibility of power supply system out-of-control or the risks and losses caused by power supply system out-of-control, and to a certain extent, ensure the power supply quality.

[0026] 2. The present invention sets a power supply data periodic collection module to periodically collect the expected energy consumption, actual energy consumption, maintenance times, scene recognition times, normal charging and power supply time, out-of-control time, and maintenance efficiency of each maintenance of the power supply system within a preset fixed period, and sends the collected power supply data to the power supply data processing module; the power supply data processing module is set to process the collected power supply data within the preset fixed period, and calculate the energy consumption excess coefficient, power supply stability coefficient, effective power supply coefficient, and average maintenance efficiency respectively; a power supply quality control capability index calculation module is set to calculate the power supply quality control capability index based on the energy consumption excess coefficient, power supply stability coefficient, effective power supply coefficient, and average maintenance efficiency; a comprehensive evaluation module is set to compare the calculated power supply quality control capability index with a preset value to determine whether the calculated value meets the requirements; if it meets the requirements, the current system continues to be used; if it does not meet the requirements, the system is optimized, which not only improves the accuracy of the power supply quality control capability evaluation, but also helps to achieve iterative optimization of the system and improve the applicability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a system structure block diagram of the present invention.

[0028] Figure 2 The present invention is a system operation flow chart. DETAILED DESCRIPTION

[0029] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0030] like Figure 1 The embodiment shown provides an intelligent early warning reasoning system based on multi-scenario analysis, including a power supply scenario data acquisition module, a power supply multi-scenario recognition model construction module, a power data online monitoring module, a scenario recognition module, an early warning reasoning module, a power supply maintenance module, a power supply data periodic collection module, a power supply data processing module, a power supply quality control capability index calculation module, a comprehensive evaluation module, and a database. The power supply scenario data acquisition module, the power supply multi-scenario recognition model construction module, the scenario recognition module, the early warning reasoning module, the power supply maintenance module, the power supply data periodic collection module, the power supply data processing module, the power supply quality control capability index calculation module, and the comprehensive evaluation module are connected in sequence, the power data online monitoring module is connected to the scenario recognition module, the scenario recognition module is connected to the power data periodic collection module, and all modules in the system are connected to the database.

[0031] The power supply scenario data acquisition module classifies the power supply scenarios and acquires historical power supply cases of different power supply scenarios based on the Internet, thereby acquiring historical power supply scenario data.

[0032] What needs to be specifically explained in this embodiment is that the power supply scenario data acquisition module includes a power supply scenario division unit, a power supply scenario keyword retrieval unit, a power supply scenario historical data acquisition unit, and a power supply scenario historical data transmission unit. The power supply scenario division unit divides the existing power supply scenarios into five types: normal charging, normal power supply, circuit failure, charging abnormality, and power supply abnormality, wherein the circuit failure includes seven causes: open circuit, short circuit, grounding, leakage, component failure, overload, and wiring error; the power supply scenario keyword retrieval unit retrieves historical power supply cases online according to the keywords of the power supply scenario type; the power supply scenario historical data acquisition unit acquires battery charging data, battery discharge data, and circuit data from the retrieved historical power supply cases, the battery charging data includes battery charging current, voltage, speed, efficiency, and temperature, the battery discharge data includes battery discharge current, voltage, power, battery discharge current frequency, voltage frequency, and battery discharge temperature, and the circuit data includes input and output voltage, input and output power, and temperature of each component or device of the circuit; the power supply scenario historical data transmission unit sends the acquired power supply scenario historical data to the power supply scenario recognition model construction module.

[0033] The power supply multi-scenario recognition model construction module constructs a power supply multi-scenario recognition model based on the acquired power supply scenario historical data and performs performance evaluation on the model. If the model passes the evaluation, it is put into use; if the model fails the evaluation, the model is tuned.

[0034] Furthermore, the power supply scenario recognition model construction module includes a data receiving unit, a data labeling unit, a data preprocessing unit, a feature extraction unit, a data partitioning unit, a model construction unit, a model training unit, a model evaluation unit, and a model tuning unit. The data receiving unit is used to receive the acquired power supply scenario historical data; the data labeling unit labels the corresponding historical scenario data with the scenario category and data occurrence time of different power supply scenarios; the data preprocessing unit is used to perform timeline integration on the labeled historical scenario data, and remove outliers after the integration is completed; the feature extraction unit compares different historical scenario data of the same power supply scheme, and respectively screens the parameters that are inconsistent with the normal charging and power supply scenario data for charging abnormalities, circuit failures, and discharge abnormalities, and removes all power supply schemes. The screened parameters are integrated, and the inconsistent parameters common to all circuit fault scenarios are obtained as circuit fault characteristic parameters, the inconsistent parameters common to all charging abnormality scenarios are charging abnormality characteristic parameters, and the inconsistent parameters common to all power supply abnormality scenarios are power supply abnormality characteristic parameters; the data division unit divides the sample data into a training data set and a test data set by a cross-validation method; the model construction unit selects a network architecture and designs a network structure and the number of layers based on the characteristics of the historical scenario characteristic parameters used and the model construction requirements, and then constructs a power supply multi-scenario recognition convolutional neural network model; the model training unit uses the training data set to train the power supply multi-scenario recognition convolutional neural network model; the model evaluation unit uses the test data set to evaluate the performance of the trained model and records the number of samples in the test data set and the number of samples correctly identified And calculate the power supply scene recognition accuracy , the specific formula is: The model tuning unit adjusts the weight and bias of the model through the back propagation algorithm and the optimizer when the calculated power supply scenario recognition accuracy is lower than the preset value, and confirms the use of the model when the calculated power supply scenario recognition accuracy is greater than or equal to the preset value.

[0035] Specifically, in this embodiment, in order to facilitate understanding of the timeline integration principle and the outlier removal standard, a set of current data and voltage data of batteries during charging in the same power supply scenario are provided for explanation. The specific steps are as follows:

[0036] A1. List the charging current timeline data: 10:00-10:05 charging current is 50mA, 10:05-10:15 charging current is 100mA, 10:15-10:30 charging current is 75mA;

[0037] A2. List the charging voltage timeline data: 10:00-10:10 charging voltage is 1.5V, 10:10-10:20 charging voltage is 3.2V, 10:20-10:30 charging voltage is 2.3V;

[0038] A3, time period information integration: 10:00-10:05 charging current is 50mA, charging voltage is 1.5V, 10:05-10:10 charging current is 100mA, charging voltage is 1.5V, 10:10-10:15 charging current is 100mA, charging voltage is 3.2V, 10:15-10:20 charging current is 75mA, charging voltage is 3.2V, 10:20-10:30 charging current is 75mA, charging voltage is 2.3V;

[0039] A4. Remove abnormal values: 10:00-10:05 charging current is 50mA, charging voltage is 1.5V, 10:05-10:10 charging current is 100mA, charging voltage is 1.5V, charging current increases, charging voltage has no corresponding increase trend, it is determined to be abnormal data, remove the charging current and charging voltage data in the time period of 10:05-10:10, 10:00-10:05 charging current is 50mA, charging voltage is 1.5V, 10:10-10:15 charging current is 100mA, charging voltage is 3.2V, charging current increases by 1 times, and the increase multiple of charging voltage is similar to the increase multiple of charging current, it is determined to be normal data, retain 10:10-10: The charging current and charging voltage data in the 15 time period, 10:10-10:15 charging current is 100mA, charging voltage is 3.2V, 10:15-10:20 charging current is 75mA, charging voltage is 3.2V, the charging current decreases, and the charging voltage has no corresponding decreasing trend, which is determined to be abnormal data. The charging current and charging voltage data in the 10:15-10:20 time period are removed. The charging current and charging voltage data in the 10:10-10:15 time period are 100mA, charging voltage is 3.2V, 10:20-10:30 charging current is 75mA, charging voltage is 2.3V, the charging current is reduced by 0.25 times, and the corresponding charging voltage reduction multiple is similar to the charging current reduction multiple, which is determined to be normal data.

[0040] Specifically, in order to facilitate understanding of the principle of feature extraction, a hypothetical case is given for illustration: Assume that three power supply schemes are retrieved, which are respectively recorded as , the charging current, charging voltage, charging speed, charging efficiency, and charging temperature are respectively ;for , the corresponding charging current, charging voltage, charging speed, charging efficiency, and charging temperature during normal charging are recorded as , the corresponding charging current, charging voltage, charging speed, charging efficiency, and charging temperature when charging is abnormal are recorded as ,for , the corresponding charging current, charging voltage, charging speed, charging efficiency, and charging temperature during normal charging are recorded as , the corresponding charging current, charging voltage, charging speed, charging efficiency, and charging temperature when charging is abnormal are recorded as ,for , the corresponding charging current, charging voltage, charging speed, charging efficiency, and charging temperature during normal charging are recorded as , the corresponding charging current, charging voltage, charging speed, charging efficiency, and charging temperature when charging is abnormal are recorded as ; Assume that for , the inconsistent parameters are , assuming that , the inconsistent parameters are Assume that , the inconsistent parameters are By integrating the assumed inconsistent parameters, it is found that when charging is abnormal, the If both are abnormal, It is the characteristic parameter of abnormal charging.

[0041] The power data online monitoring module uses sensors to monitor the power data of the power supply system online, and transmits the monitored power data to the scene recognition module in real time.

[0042] The scene recognition module inputs the received power data into the power supply multi-scene recognition model to perform scene recognition and record the scene recognition results.

[0043] Furthermore, the scene recognition module includes a monitoring data receiving unit, a power supply plan entry unit, a monitoring data entry unit, a scene recognition unit, a scene recognition result recording unit, and a recognition result output unit. The monitoring data receiving unit is used to receive monitoring data; the power supply plan entry unit is used to enter the formulated normal charging plan and normal power supply plan; the monitoring data entry unit is used to enter the real-time monitored power data; the scene recognition unit performs scene recognition based on the entered power supply plan and monitoring data; the scene recognition result recording unit is used to record real-time scene recognition results, including the number of recognized scenes, the number of charging abnormality scenes, the number of circuit failure scenes, and the number of power supply abnormality scenes; the recognition result output unit is used to display the recognized scenes to the early warning reasoning module.

[0044] The early warning reasoning module infers the probability of power outage based on the scene recognition result, and issues an alarm when the probability of power outage reaches a preset value.

[0045] Furthermore, the early warning reasoning module includes a data receiving unit, a power outage probability reasoning unit, and an intelligent early warning unit. The data receiving unit is used to receive the number of scenes identified in real time, the number of charging abnormalities, the number of circuit failure scenes, and the number of power supply abnormalities; the power outage probability reasoning unit is based on the number of scenes identified in real time. , Number of abnormal charging scenarios , Number of circuit failure scenarios , and the number of power supply abnormality scenarios Inference of the probability of power supply failure , the specific formula is: , is the proportionality coefficient, The intelligent early warning unit issues an alarm when the inferred probability of power outage reaches a preset alarm threshold, and does not perform any processing when the inferred probability of power outage does not reach the preset alarm threshold.

[0046] The power supply maintenance module arranges maintenance personnel to perform maintenance on the power supply system upon receiving an alarm, records maintenance data, and calculates a maintenance efficiency coefficient based on the recorded maintenance data.

[0047] Furthermore, the power supply maintenance module includes an alarm receiving unit, a maintenance task allocation unit, a maintenance progress real-time update unit, a maintenance data recording unit, a maintenance efficiency coefficient calculation unit, and a data output unit. The alarm receiving unit is used to receive alarm information; the maintenance task allocation unit sends the power supply maintenance task to all maintenance personnel terminals on duty after receiving the alarm information, and the maintenance personnel who first confirms to receive the maintenance task is the target maintenance personnel; the maintenance progress real-time update unit updates the maintenance progress in real time after the target maintenance personnel confirms to receive the maintenance task; the maintenance data recording unit is used to record the power supply outage probability before and after maintenance and the maintenance time required to reduce the power supply outage probability to below the alarm threshold; the maintenance efficiency coefficient calculation unit is based on the power supply outage probability before maintenance , Probability of power supply outage after maintenance is completed and the maintenance time required to reduce the probability of power outage below the alarm threshold Calculating the maintenance efficiency factor The specific formula is: ; The data output unit sends the calculated maintenance efficiency coefficient to the power supply data processing module.

[0048] The power supply data periodic collection module is used to periodically collect the expected energy consumption, actual energy consumption, maintenance times, scene recognition times, normal charging and power supply time, out-of-control time, and maintenance efficiency of each maintenance of the power supply system within a preset fixed period and send them to the power supply data processing module.

[0049] The power supply data processing module processes the collected power supply data within a preset fixed period, and calculates the energy consumption excess coefficient, the power supply stability coefficient, the effective power supply coefficient, and the average maintenance efficiency.

[0050] Furthermore, the power supply data processing module includes a data receiving unit, an energy consumption excess coefficient calculation unit, a power supply stability coefficient calculation unit, an effective power supply coefficient calculation unit, an average maintenance efficiency calculation unit, and a data output unit. The data receiving unit is used to receive the data collected by the power supply data periodic collection module; the energy consumption excess coefficient calculation unit is based on the expected energy consumption within a preset fixed period. and actual energy consumption Calculate the excess energy consumption factor The specific formula is: The power supply stability coefficient calculation unit is based on the number of maintenance times within a preset fixed period. and scene recognition times Calculate the power supply stability factor The specific formula is: , e is a natural constant; the effective power supply coefficient calculation unit is based on the normal charging and power supply time within a preset fixed period and duration of loss of control Calculate the effective power supply factor The specific formula is: The average maintenance efficiency calculation unit is based on the number of maintenance times within a preset fixed period. and the maintenance efficiency of the ith maintenance Calculate the average maintenance efficiency factor The specific formula is: ; The data output unit is used to send the calculated indicators to the power supply quality control capability index calculation module.

[0051] The power supply quality control capability index calculation module calculates the power supply quality control capability index based on the energy consumption excess coefficient, the power supply stability coefficient, the effective power supply coefficient, and the average maintenance efficiency.

[0052] Furthermore, the power supply quality control capability index calculation module is based on the energy consumption excess coefficient XE, the power supply stability coefficient , Power supply stability coefficient , Effective power supply coefficient , and the average maintenance efficiency Calculate the power supply quality control capability index The specific formula is: , e is a natural constant.

[0053] The comprehensive evaluation module compares the calculated power supply quality control capability index with a preset value to determine whether the calculated value meets the requirements. If it does, the current system continues to be used; if it does not, the system is optimized.

[0054] The database is used to store data information of all modules in the system.

[0055] It should be specifically noted in this embodiment that the preset values ​​and coefficients used are selected based on actual needs and are not limited to specific values ​​here.

[0056] like Figure 2 The embodiment shown provides an operation process of an intelligent early warning reasoning system based on multi-scenario analysis, including the following steps:

[0057] S1: Classify power supply scenarios and obtain historical power supply cases of different power supply scenarios based on the Internet, and then obtain historical data of power supply scenarios;

[0058] S2: Build a power supply multi-scenario recognition model based on the acquired power supply scenario historical data and evaluate its performance. If it passes the evaluation, it will be put into use. If it fails the evaluation, the model will be tuned.

[0059] S3: Use sensors to monitor the power data of the power supply system online;

[0060] S4: inputting the power data into the power supply multi-scenario recognition model for scene recognition, and recording the number of recognized scenes, the number of charging abnormality scenes, the number of circuit failure scenes, and the number of power supply abnormality scenes;

[0061] S5: infer the probability of power outage based on the number of scenarios identified in real time, the number of charging abnormalities, the number of circuit failure scenarios, and the number of power supply abnormalities. When the inferred probability of power outage reaches a preset alarm threshold, an alarm is issued. When the inferred probability of power outage does not reach the preset alarm threshold, no processing is performed.

[0062] S6: Arrange maintenance personnel to perform maintenance on the power supply system upon receiving the alarm, record maintenance data, and calculate a maintenance efficiency coefficient based on the recorded maintenance data;

[0063] S7: regularly collect the expected energy consumption, actual energy consumption, maintenance times, scene recognition times, normal charging and power supply time, out-of-control time, and maintenance efficiency of each maintenance of the power supply system within a preset fixed period;

[0064] S8: Processing the collected power supply data within a preset fixed period, and calculating the energy consumption excess coefficient, the power supply stability coefficient, the effective power supply coefficient, and the average maintenance efficiency;

[0065] S9: Calculate the power supply quality control capability index based on the energy consumption excess coefficient, power supply stability coefficient, effective power supply coefficient, and average maintenance efficiency;

[0066] S10: Compare the calculated power supply quality control capability index with the preset value to determine whether the calculated value meets the demand. If it does, continue to use the current system; if it does not, optimize the system.

[0067] Finally: 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, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent early warning reasoning system based on multi-scenario analysis, characterized by: include: Power supply scenario data acquisition module: classifies power supply scenarios and obtains historical power supply cases of different power supply scenarios based on the Internet, and then obtains historical data of power supply scenarios; Power supply multi-scenario recognition model construction module: Build a power supply multi-scenario recognition model based on the acquired power supply scenario historical data and evaluate its performance. If it passes the evaluation, it will be put into use. If it fails the evaluation, the model will be tuned. Power data online monitoring module: uses sensors to monitor the power data of the power supply system online, and transmits the monitored power data to the scene recognition module in real time; Scene recognition module: inputs the received power data into the power supply multi-scene recognition model for scene recognition and records the scene recognition results; Early warning reasoning module: infers the probability of power outage based on the scene recognition results, and issues an alarm when the probability of power outage reaches a preset value; Power supply maintenance module: arranges maintenance personnel to maintain the power supply system when an alarm is received, records maintenance data and calculates the maintenance efficiency coefficient based on the recorded maintenance data; Power supply data regular collection module: used to regularly collect the expected energy consumption, actual energy consumption, maintenance times, scene recognition times, normal charging and power supply time, out-of-control time, and maintenance efficiency of each maintenance of the power supply system within a preset fixed period and send them to the power supply data processing module; Power supply data processing module: processes the collected power supply data within a preset fixed period, and calculates the energy consumption excess coefficient, power supply stability coefficient, effective power supply coefficient, and average maintenance efficiency; Power supply quality control capability index calculation module: calculates the power supply quality control capability index based on energy consumption excess coefficient, power supply stability coefficient, effective power supply coefficient, and average maintenance efficiency; The power supply data processing module includes a data receiving unit, an energy consumption excess coefficient calculation unit, a power supply stability coefficient calculation unit, an effective power supply coefficient calculation unit, an average maintenance efficiency calculation unit, and a data output unit. The data receiving unit is used to receive the data collected by the power supply data periodic collection module; the energy consumption excess coefficient calculation unit calculates the expected energy consumption within a preset fixed period based on the expected energy consumption within a preset fixed period. and actual energy consumption Calculate the excess energy consumption factor The specific formula is: The power supply stability coefficient calculation unit is based on the number of maintenance times within a preset fixed period. and scene recognition times Calculate the power supply stability factor The specific formula is: , e is a natural constant; the effective power supply coefficient calculation unit is based on the normal charging and power supply time within a preset fixed period and duration of loss of control Calculate the effective power supply factor The specific formula is: The average maintenance efficiency calculation unit is based on the number of maintenance times within a preset fixed period. and the maintenance efficiency of the ith maintenance Calculate the average maintenance efficiency factor The specific formula is: ; The data output unit is used to send the calculated index to the power supply quality control capability index calculation module; The power supply quality control capability index calculation module is based on the energy consumption excess coefficient , Power supply stability coefficient , Effective power supply coefficient , and the average maintenance efficiency Calculate the power supply quality control capability index The specific formula is: , e is a natural constant.

2. The intelligent early warning reasoning system based on multi-scenario analysis according to claim 1 is characterized in that: The power supply multi-scenario recognition model construction module includes a data receiving unit, a data labeling unit, a data preprocessing unit, a feature extraction unit, a data partitioning unit, a model construction unit, a model training unit, a model evaluation unit, and a model tuning unit. The data receiving unit is used to receive the acquired power supply scenario historical data; the data labeling unit labels the corresponding historical scenario data with the scenario category and data appearance time of different power supply scenarios; the data preprocessing unit is used to perform timeline integration on the labeled historical scenario data, and remove outliers after the integration is completed; The feature extraction unit compares different historical scenario data of the same power supply scheme, and respectively screens out parameters that are inconsistent with the normal charging and power supply scenario data, such as charging abnormality, circuit fault and discharge abnormality scenario data, and integrates the parameters screened out by all power supply schemes, and obtains the inconsistent parameters common to all circuit fault scenarios as circuit fault characteristic parameters, the inconsistent parameters common to all charging abnormality scenarios as charging abnormality characteristic parameters, and the inconsistent parameters common to all power supply abnormality scenarios as power supply abnormality characteristic parameters; the data division unit divides the sample data into a training data set and a test data set by a cross-validation method; the model construction unit selects a network architecture and designs a network structure and the number of layers based on the characteristics of the historical scenario characteristic parameters used and the model construction requirements, and then constructs a power supply multi-scenario recognition convolutional neural network model; the model training unit uses the training data set to train the power supply multi-scenario recognition convolutional neural network model; the model evaluation unit uses the test data set to perform performance evaluation on the trained model and records the number of samples in the test data set and the number of samples correctly identified And calculate the power supply scene recognition accuracy , the specific formula is: The model tuning unit adjusts the weight and bias of the model through the back propagation algorithm and the optimizer when the calculated power supply scenario recognition accuracy is lower than the preset value, and confirms the use of the model when the calculated power supply scenario recognition accuracy is greater than or equal to the preset value.

3. The intelligent early warning reasoning system based on multi-scenario analysis according to claim 1 is characterized in that: The scene recognition module includes a monitoring data receiving unit, a power supply scheme input unit, a monitoring data input unit, a scene recognition unit, a scene recognition result recording unit, and a recognition result output unit, and the monitoring data receiving unit is used to receive monitoring data; The power supply plan input unit is used to input the normal charging plan and the normal power supply plan that have been formulated; The monitoring data input unit is used to input the power data monitored in real time; The scene recognition unit performs scene recognition based on the input power supply scheme and monitoring data; The scene recognition result recording unit is used to record real-time scene recognition results, including the number of recognized scenes, the number of scenes with abnormal charging, the number of scenes with circuit failure, and the number of scenes with abnormal power supply; the recognition result output unit is used to display the recognized scenes to the early warning reasoning module.

4. The intelligent early warning reasoning system based on multi-scenario analysis according to claim 1 is characterized in that: The early warning reasoning module includes a data receiving unit, a power outage probability reasoning unit, and an intelligent early warning unit. The data receiving unit is used to receive the number of scenes identified in real time, the number of charging abnormalities, the number of circuit failure scenes, and the number of power supply abnormalities; the power outage probability reasoning unit is based on the number of scenes identified in real time. , Number of abnormal charging scenarios , Number of circuit failure scenarios , and the number of power supply abnormality scenarios Inference of the probability of power supply failure , the specific formula is: , , is the proportionality coefficient, The intelligent early warning unit issues an alarm when the inferred probability of power outage reaches a preset alarm threshold, and does not perform any processing when the inferred probability of power outage does not reach the preset alarm threshold.

5. The intelligent early warning reasoning system based on multi-scenario analysis according to claim 1 is characterized in that: The power supply maintenance module includes an alarm receiving unit, a maintenance task allocation unit, a maintenance progress real-time updating unit, a maintenance data recording unit, a maintenance efficiency coefficient calculation unit, and a data output unit, and the alarm receiving unit is used to receive alarm information; After receiving the alarm information, the maintenance task allocation unit sends the power supply maintenance task to all maintenance personnel terminals on duty. The maintenance personnel who first confirms to receive the maintenance task is the target maintenance personnel; the maintenance progress real-time update unit updates the maintenance progress in real time after the target maintenance personnel confirms to receive the maintenance task; the maintenance data recording unit is used to record the power supply outage probability before and after maintenance and the maintenance time required to reduce the power supply outage probability to below the alarm threshold; the maintenance efficiency coefficient calculation unit is based on the power supply outage probability before maintenance , Probability of power supply outage after maintenance is completed and the maintenance time required to reduce the probability of power outage below the alarm threshold Calculating the maintenance efficiency factor The specific formula is: ; The data output unit sends the calculated maintenance efficiency coefficient to the power supply data processing module.

Images