Method and system for handling operation risks of construction power supply system

Abstract

The present invention provides a method and system for handling the operation risk of a construction power system, which relates to the field of electrical safety technology. In this solution, based on the environmental monitoring data of the construction power system operation area in a preset time period uploaded by the collection layer device and the safety monitoring data of the electrical equipment, the computer device uses the construction power prediction model to perform prediction processing, and obtains the safety prediction data for the next time period, so as to determine whether an electrical fire accident will occur. By applying this method, the prediction accuracy of electrical fire accidents in the construction power system operation area can be improved, thereby improving the overall safety level of the construction power system operation area.

Description

Technical Field

[0001] The present application relates to the field of electrical safety technology, and in particular to a method and system for handling operational risks of a construction power system. Background Art

[0002] In areas where construction power systems operate, electrical fire accidents often occur due to the use of welding and other processes.

[0003] At present, in order to prevent the occurrence of electrical fire accidents, an electrical fire warning system is usually installed in the construction power system operation area. Specifically, when the electrical fire warning system detects signs of electrical fire, it will immediately sound an alarm and initiate emergency measures to control and eliminate the fire threat as soon as possible.

[0004] However, the above electrical fire warning systems have the problem of low prediction accuracy. Summary of the invention

[0005] The present application provides a method and system for handling operation risks of a construction power system, so as to solve the problem of low accuracy of electrical fire prediction in the prior art.

[0006] In a first aspect, the present application provides a construction power system operation risk processing method, which is applied to a computer device in a construction power system operation risk processing system, and the method includes:

[0007] Based on the data uploaded by the collection layer equipment, the environmental monitoring data of the construction power system operation area in a preset time period and the safety monitoring data of the electrical equipment are obtained, wherein the preset time period is a period of time including the current moment and before the current moment, the environmental monitoring data includes the average acetylene concentration monitoring data, the average aluminum oxide powder concentration monitoring data, the average temperature monitoring data, the average humidity monitoring data and the average light intensity monitoring data, and the safety monitoring data includes the average sound intensity monitoring data and the average spark duration monitoring data of the electrical equipment;

[0008] Inputting the environmental monitoring data and the safety monitoring data into a construction power consumption prediction model for prediction processing to obtain safety prediction data for the next time period;

[0009] According to the safety prediction data, it is determined whether an electrical fire accident will occur in the next time period.

[0010] In a possible design of the first aspect, the process of acquiring the construction power consumption prediction model includes:

[0011] According to historical environmental monitoring data and historical safety monitoring data, a preset algorithm is used to determine the optimal parameter configuration scheme of the initial construction power consumption prediction model, wherein the initial construction power consumption prediction model is a neural network model;

[0012] According to the optimal parameter configuration scheme, the target parameters in the initial construction power consumption prediction model are configured to obtain the construction power consumption prediction model.

[0013] In a possible design of the first aspect, the initial construction power consumption prediction model is trained based on the historical environmental monitoring data and the historical safety monitoring data to obtain the optimal parameter configuration scheme of the initial construction power consumption prediction model, including:

[0014] Step a, obtaining a sample data set based on the historical environmental monitoring data and the historical safety monitoring data, wherein the sample data set includes a plurality of data samples, each of which includes environmental monitoring data and safety monitoring data for a time period and safety monitoring data for a time period next to the time period;

[0015] Step b, initializing N groups of parameter configuration schemes, N groups of offset configuration schemes, a first learning factor, a second learning factor, a third learning factor, and a preset maximum number of iterations, wherein N is a positive integer, and each group of parameter configuration schemes includes weight parameters and bias item parameters;

[0016] Step c, initializing N first construction power consumption prediction models with exactly the same structure to obtain N second construction power consumption prediction models;

[0017] Step d: for each second construction power consumption prediction model, select a set of parameter configuration schemes, configure the initial parameters in the second construction power consumption prediction model, and obtain N initial construction power consumption prediction models;

[0018] Step e, selecting N groups of data samples from the sample data set, and inputting the environmental monitoring data and safety monitoring data of a time period in each group of data samples into N initial construction power consumption prediction models to predict the safety monitoring data of the next time period, thereby obtaining N groups of safety prediction data;

[0019] Step f, calculating the fitness value of each group of parameter configuration schemes according to the N groups of safety prediction data and the safety monitoring data of the next time period in the N groups of data samples;

[0020] Step g, obtaining the parameter configuration scheme corresponding to the highest fitness value, and recording it as the target parameter configuration scheme;

[0021] Step h: updating each group of offset configuration schemes according to the first learning factor, the second learning factor, the third learning factor, the target parameter configuration scheme, and each group of parameter configuration schemes to obtain N groups of new offset configuration schemes;

[0022] Step i: according to each group of new offset configuration schemes, update each group of parameter configuration schemes to obtain N groups of updated parameter configuration schemes;

[0023] Step j: Repeat steps d to i until the maximum number of iterations is reached to obtain the optimal parameter configuration solution.

[0024] In a possible design of the first aspect, determining whether an electrical fire accident will occur in the next time period according to the safety prediction data includes:

[0025] If the average sound intensity data in the safety prediction data is higher than the preset sound intensity threshold data or the average spark duration data is higher than the preset spark duration threshold data, it is determined that an electrical fire accident will occur in the operation area of ​​the construction power system in the next time period.

[0026] In a possible design of the first aspect, the method further includes:

[0027] If it is determined that an electrical fire accident will occur in the operation area of ​​the construction power system in the next time period, the power supply voltage of the construction power system is automatically adjusted, and a safety warning message is pushed to the terminal device of the manager, wherein the adjustment of the power supply voltage of the construction power system includes controlling the power supply voltage of the construction power system to be reduced to a preset safety power supply voltage, or controlling the power supply of the construction power system to be turned off, and the safety warning message is used to indicate that an electrical fire accident will occur in the operation area of ​​the construction power system.

[0028] In a possible design of the first aspect, calculating the fitness value of each group of parameter configuration schemes according to the N groups of safety prediction data and the safety monitoring data of the next time period in the N groups of data samples includes:

[0029] For each set of parameter configuration schemes, the mean square error, root mean square error and mean absolute error are calculated based on the safety prediction data and the safety monitoring data of the next time period in the data sample;

[0030] The fitness value of the parameter configuration scheme is calculated according to the mean square error value, the root mean square error value and the mean absolute error value.

[0031] In a possible design of the first aspect, updating each group of offset configuration schemes according to the first learning factor, the second learning factor, the third learning factor, the target parameter configuration scheme, and each group of parameter configuration schemes to obtain N groups of new offset configuration schemes includes:

[0032] For each group of offset configuration schemes, calculating a first offset value according to the first learning factor and the offset configuration scheme;

[0033] Calculating a second offset value according to the second learning factor, the preset first random number, and the parameter configuration scheme;

[0034] Calculating a third offset value according to the third learning factor, the preset second random number, and the target parameter configuration scheme;

[0035] The new offset configuration scheme is obtained according to the first offset value, the second offset value and the third offset value.

[0036] In a possible design of the first aspect, the first learning factor is determined according to a preset maximum first learning factor, a preset minimum first learning factor, a current number of iterations, and the maximum number of iterations;

[0037] The second learning factor is determined according to a preset maximum second learning factor, a preset minimum second learning factor, the current number of iterations, the maximum number of iterations, and a preset cosine function;

[0038] The third learning factor is determined according to a preset maximum third learning factor, a preset minimum third learning factor, the current number of iterations, the maximum number of iterations, and a preset sine function.

[0039] In a second aspect, the present application provides a computer device, comprising: a receiver, a processor, and a memory connected to the processor;

[0040] The memory stores computer-executable instructions;

[0041] The processor executes the computer-executable instructions stored in the memory to implement the method for handling operation risks of a construction power system as described in any one of the first aspects.

[0042] In a third aspect, the present application provides a construction power system operation risk management system, comprising: the computer device described in the second aspect and a collection layer device connected to the computer device;

[0043] The collection layer equipment is used to monitor the environment in the operation area of ​​the construction power system and the safety of the electrical equipment in real time and upload the monitored data to the computer equipment;

[0044] The collection layer equipment includes: acetylene concentration sensor, aluminum oxide powder concentration sensor, temperature and humidity sensor, light intensity sensor, sound intensity sensor, spark detection sensor;

[0045] The acetylene concentration sensor is used to monitor acetylene concentration data;

[0046] Alumina powder concentration sensor is used to monitor the concentration data of alumina powder;

[0047] Temperature and humidity sensors are used to monitor temperature and humidity data;

[0048] The light intensity sensor is used to monitor light intensity data;

[0049] The sound intensity sensor is used to monitor the sound intensity data of electrical equipment;

[0050] Spark detection sensors are used to monitor spark duration data in electrical equipment.

[0051] In a fourth aspect, the present application provides a construction power system operation risk management device, comprising:

[0052] An acquisition module, for acquiring environmental monitoring data of the construction power system operation area in a preset time period and safety monitoring data of electrical equipment based on data uploaded by the collection layer equipment, wherein the preset time period is a period of time including the current moment and before the current moment, the environmental monitoring data includes average acetylene concentration monitoring data, average aluminum oxide powder concentration monitoring data, average temperature monitoring data, average humidity monitoring data and average light intensity monitoring data, and the safety monitoring data includes average sound intensity monitoring data and average spark duration monitoring data of the electrical equipment;

[0053] A processing module, used for inputting the environmental monitoring data and the safety monitoring data into a construction power consumption prediction model for prediction processing to obtain safety prediction data for the next time period;

[0054] The processing module is further used to determine whether an electrical fire accident will occur in the next time period based on the safety prediction data.

[0055] In a fifth aspect, the present application provides a computer-readable storage medium, in which computer execution instructions are stored. When the computer execution instructions are executed by a processor, they are used to implement the construction power system operation risk management method as described in any one of the first aspects.

[0056] In a sixth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the method for handling operation risks of a construction power system as described in any one of the first aspects.

[0057] The construction power system operation risk management method and system provided in the present application relate to the field of electrical safety technology. In the construction power system operation area, the collection layer equipment uploads the environmental monitoring data and the safety monitoring data of the electrical equipment in the preset time period to the computer equipment to ensure the real-time nature of the monitoring data. At the same time, the computer equipment uses the construction power prediction model to process the above monitoring data to obtain the safety prediction data for the next time period, so as to determine whether an electrical fire accident will occur. Through this method, not only can the high-precision prediction of electrical fire accidents in the construction power system operation area be achieved, effectively reducing the occurrence of electrical fire accidents, but also the overall safety level of the construction power system operation area can be significantly improved to ensure the smooth progress of the construction process. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0059] Figure 1 A schematic diagram of an application scenario of a construction power system operation risk management method provided in this application;

[0060] Figure 2 A flowchart of a first embodiment of a method for handling operation risks of a construction power system provided in the present application;

[0061] Figure 3 A flowchart of a second embodiment of a construction power system operation risk treatment method provided by the present application;

[0062] Figure 4 A flowchart of a third embodiment of a construction power system operation risk treatment method provided by the present application;

[0063] Figure 5 A flowchart of a fourth embodiment of a construction power system operation risk treatment method provided by the present application;

[0064] Figure 6 A flowchart of a fifth embodiment of a construction power system operation risk treatment method provided by the present application;

[0065] Figure 7 A flowchart of a sixth embodiment of a construction power system operation risk management method provided by the present application;

[0066] Figure 8 A flowchart of Embodiment 7 of a construction power system operation risk treatment method provided by the present application;

[0067] Fig. 9 A schematic diagram of a construction power system operation risk management system framework provided for this application;

[0068] Fig.10 A schematic diagram of a process for determining parameters of a construction power consumption prediction model provided in this application;

[0069] Fig.11 A schematic diagram of the structure of a first embodiment of a construction power system operation risk management device provided by the present application;

[0070] Fig.12 A schematic diagram of the structure of a computer device provided for this application;

[0071] Fig.13 A construction power system operation risk management system is provided for this application.

[0072] Reference numerals:

[0073] 901: Collection layer; 9011: Sensor; 9012: Monitoring equipment; 9013: Probe; 9014: Other sensing equipment;

[0074] 902: transport layer; 9021: mobile network; 9022: wireless network;

[0075] 903: processing layer; 9031: construction power consumption prediction model;

[0076] 904: Application layer; 9041: Smart construction site; 90411: Electrical fire warning; 90412: Power safety guarantee; 9042: Visual display; 90421: Automatic processing; 90422: Intelligent control; 90423: Construction quality; 90424: Operation and maintenance management;

[0077] 1100: construction power system operation risk processing device; 1101: acquisition module; 1102: processing module;

[0078] 1200: computer equipment; 1201: receiver; 1202: processor; 1203: memory; 1300: construction power system operation risk management system; 1301: collection layer equipment; 13011: acetylene concentration sensor; 13012: alumina powder concentration sensor; 13013: temperature and humidity sensor; 13014: light intensity sensor; 13015: sound intensity sensor; 13016: spark detection sensor.

[0079] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0080] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0081] There are usually a large number of flammable materials and flammable and explosive gases at construction sites. However, due to the widespread use of processes such as welding, electrical fire accidents occur frequently.

[0082] In order to prevent electrical fire accidents, electrical fire warning systems are usually installed in the construction power system operation area. These warning systems can monitor the operating status of electrical equipment in real time, detect abnormal current, voltage and temperature changes, and issue alarms when potential dangers occur, thereby effectively reducing the number of fires.

[0083] However, the existing electrical early warning system has deficiencies in intelligent analysis of monitoring data, which limits its prediction accuracy to a certain extent. At the same time, the existing electrical early warning system is easily disturbed by the surrounding environment information, which further reduces the accuracy of the prediction.

[0084] In view of the above problems, the inventors found in the process of studying the operation risk treatment method of the construction power system that the existing operation risk treatment method of the construction power system relies on manual monitoring of various data, and fails to make full use of a large amount of available data for pattern recognition and trend analysis, resulting in low prediction accuracy. Based on this, the inventors consider starting from the perspective of real-time monitoring data and prediction models to improve the prediction accuracy of electrical fire accidents. Specifically, the collection layer device 1301 uploads the environmental monitoring data of the construction power system operation area within a preset time period and the safety monitoring data of the electrical equipment to the computer device 1200 in real time. The computer device 1200 uses the construction power prediction model 9031 to process these monitoring data and obtain the safety prediction data for the next time period to determine whether an electrical fire accident will occur. Through this method, not only the prediction accuracy is improved, but also the probability of electrical fire accidents is reduced to a certain extent, thereby ensuring construction safety.

[0085] Figure 1 This is a schematic diagram of an application scenario of a construction power system operation risk management method provided in this application. Figure 1As shown, the construction power system operation risk processing method provided by the present application includes a collection layer device 1301 and a computer device 1200. Among them, the collection layer device 1301 is mainly used to collect environmental monitoring data of the construction power system operation area in a preset time period and safety monitoring data of electrical equipment. The construction power prediction model 9031 pre-deployed in the computer device 1200 is mainly used to process the environmental monitoring data and safety monitoring data of electrical equipment uploaded by the collection layer device 1301, and to determine whether an electrical fire accident will occur in the construction power system operation area.

[0086] The collection layer equipment 1301 includes but is not limited to an acetylene concentration sensor 13011, an aluminum oxide powder concentration sensor 13012, a temperature and humidity sensor 13013, a light intensity sensor 13014, and a spark detection sensor 13016. The computer equipment 1200 includes but is not limited to a server and a personal computer. Figure 1 Only one collection layer device 1301 and computer device 1200 is shown, but it should be understood that there may be two or more collection layer devices 1301 and computer devices 1200.

[0087] Specifically, in the operation area of ​​the construction power system, the collection layer device 1301 collects the environmental monitoring data and safety monitoring data of the electrical equipment in the area in a preset time period in real time, and transmits these monitoring data to the computer device 1200 through the mobile network 9021 and / or the wireless network 9022. The computer device 1200 uses the construction power prediction model 9031 deployed inside it to process the received monitoring data and obtain the safety prediction data for the next time period. The prediction process provides data support for determining whether an electrical fire accident will occur in the operation area of ​​the construction power system, so that corresponding preventive measures can be taken when necessary to reduce the probability of electrical fire accidents.

[0088] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0089] Figure 2 This is a flow chart of a first embodiment of a construction power system operation risk treatment method provided by this application. Figure 2 As shown, the process of the construction power system operation risk treatment method may include:

[0090] S201: Based on the data uploaded by the collection layer device 1301, obtain the environmental monitoring data of the construction power system operation area in a preset time period and the safety monitoring data of the electrical equipment, wherein the preset time period is a period of time including the current moment and before the current moment, and the environmental monitoring data include the average acetylene concentration monitoring data, the average aluminum oxide powder concentration monitoring data, the average temperature monitoring data, the average humidity monitoring data and the average light intensity monitoring data; the safety monitoring data include the average sound intensity monitoring data and the average spark duration monitoring data of the electrical equipment.

[0091] In this step, in order to reduce the probability of electrical fire accidents in the construction power system operation area, it is crucial to obtain and process the monitoring data in advance.

[0092] The construction power system operation area refers to the power supply area set up within the construction site of a building or project to ensure the normal operation of various electrical equipment and facilities during the construction process. These areas cover the distribution areas of different power demands within the construction site and are used to ensure the smooth progress and safe management of construction activities.

[0093] It should be noted that the operating area of ​​the construction power system can be a single area or multiple areas. This application does not specifically limit the number of operating areas of the construction power system.

[0094] Specifically, in order to timely obtain and process the environmental monitoring data and the safety monitoring data of the electrical equipment in the operation area of ​​the construction power system within the preset time period, the present application pre-deploys the construction power system operation risk processing system 1300 in the operation area of ​​the construction power system. In the construction power system operation risk processing system 1300, at least one collection layer device 1301 and a computer device 1200 are included. Among them, the collection layer device 1301 includes but is not limited to various sensors 9011, monitoring equipment 9012 and probes 9013. The computer device 1200 includes but is not limited to personal computers and servers.

[0095] The data collected by the collection layer device 1301 include but are not limited to environmental monitoring data and safety monitoring data of electrical equipment. The environmental monitoring data include but are not limited to average acetylene concentration monitoring data, average aluminum oxide powder concentration monitoring data, average temperature monitoring data, average humidity monitoring data, and average light intensity monitoring data, and the safety monitoring data of electrical equipment include but are not limited to average sound intensity monitoring data and average spark duration monitoring data of electrical equipment.

[0096] Among them, the average acetylene concentration monitoring data refers to the average value of the acetylene gas concentration in the construction power system operation area within the preset time period. The average aluminum oxide powder concentration monitoring data refers to the average concentration of aluminum oxide dust suspended in the air in the construction power system operation area within the preset time period. The average temperature monitoring data refers to the average value of the temperature in the construction power system operation area within the preset time period. The average humidity monitoring data refers to the average value of the humidity in the construction power system operation area within the preset time period. The average light intensity monitoring data refers to the average value of the light intensity in the construction power system operation area within the preset time period.

[0097] The average sound intensity monitoring data of electrical equipment refers to the average sound intensity of electrical equipment in operation in the construction power system operation area within a preset time period. The average spark duration monitoring data refers to the average duration of sparks in operation of electrical equipment in the construction power system operation area within a preset time period. The preset time period is a period of time including the current moment and before the current moment. It is worth noting that the length of the preset time period is determined according to actual conditions and is not limited here.

[0098] In a specific implementation, after the collection layer device 1301 acquires the monitored data within the operation area of ​​the construction power system, it promptly transmits the monitored data to the computer device 1200 via the mobile network 9021 and / or the wireless network 9022 for analysis and prediction of safety data.

[0099] S202: Input the environmental monitoring data and safety monitoring data into the construction power consumption prediction model 9031 for prediction processing to obtain safety prediction data for the next time period.

[0100] In this step, according to step S201, after obtaining the environmental monitoring data of the construction power system operation area in the preset time period and the safety monitoring data of the electrical equipment, the computer device 1200 inputs the monitoring data into the construction power prediction model 9031 for prediction processing to obtain the safety prediction data for the next time period.

[0101] Among them, the construction power consumption prediction model 9031 is used to obtain the safety prediction data for the next time period. Specifically, the construction power consumption prediction model 9031 takes the environmental monitoring data and the safety monitoring data of the electrical equipment within the preset time period as input, and outputs the safety prediction data for the next time period. The safety prediction data includes but is not limited to the average sound intensity prediction data and spark duration prediction data of the electrical equipment. It is worth noting that the acquisition of the construction power consumption prediction model 9031 can refer to steps S301 to S302, which will not be repeated here.

[0102] The average sound intensity prediction data of electrical equipment is the predicted value after averaging the sound intensity generated by electrical equipment in a specific time period. Specifically, during the operation of electrical equipment, a certain amount of sound is usually generated, which mainly comes from factors such as mechanical movement and current flow. Under normal operating conditions, the sound intensity of electrical equipment usually remains within a stable range. However, when electrical equipment encounters abnormal conditions, such as overload, short circuit or mechanical failure, the sound intensity may increase significantly. By monitoring and predicting the sound intensity of electrical equipment, abnormal conditions of the equipment can be discovered in advance, so that preventive measures can be taken.

[0103] The spark duration prediction data of electrical equipment predicts the average duration of the spark in a specific time period. Specifically, sparks may be generated during the operation of electrical equipment, especially during switch operation or poor contact. The frequency and duration of sparks are important indicators for judging the safety of electrical equipment. Frequent and long-lasting sparks may cause electrical equipment to overheat or ignite surrounding combustibles, thereby causing a fire. By predicting the spark duration, potential fire hazards can be discovered in a timely manner and corresponding preventive measures can be taken.

[0104] By predicting the average sound intensity and average spark duration of electrical equipment, comprehensive monitoring of the operating status of electrical equipment can be achieved. These data not only help to detect abnormal conditions and potential failures of electrical equipment in advance, but also provide an important basis for preventing electrical fires. By taking corresponding preventive and maintenance measures in a timely manner, the probability of electrical fires can be effectively reduced and the safety of the construction site can be guaranteed.

[0105] Optionally, in a possible implementation, before inputting the environmental monitoring data and the safety monitoring data into the construction power consumption prediction model 9031 for prediction processing, the attention mechanism can be applied to calibrate different weight values ​​for the environmental monitoring data and the safety monitoring data, so as to ensure that the model can more accurately identify and focus on the key factors that have a greater impact on construction power consumption, and improve the accuracy and reliability of the prediction results. In this way, the construction power consumption prediction model 9031 can dynamically adjust the weights of various types of data, make full use of the characteristics of the environmental and safety monitoring data, and optimize the prediction performance.

[0106] S203: Determine whether an electrical fire accident will occur in the next time period based on the safety prediction data.

[0107] In this step, according to step S202, after the computer device 1200 obtains the safety prediction data of the construction power system operation area in the next time period, it is necessary to determine whether there is a potential electrical fire accident in the construction power system operation area based on the safety prediction data.

[0108] Specifically, by comparing the safety prediction data with the preset safety prediction data threshold, it is possible to accurately determine whether an electrical fire accident will occur in the operation area of ​​the construction power system, so as to take corresponding prevention and control measures to ensure the safety and controllability of the operation area of ​​the construction power system.

[0109] The method for handling the operation risks of the construction power system provided in the embodiment of the present application is based on the construction power prediction model 9031, and processes the environmental monitoring data and the safety monitoring data of the electrical equipment collected by the collection layer device 1301 in real time to obtain the safety prediction data of the operation area of ​​the construction power system in the next time period. By analyzing the safety prediction data, it is possible to accurately determine whether an electrical fire accident will occur. This solution makes full use of the environmental monitoring data and the safety monitoring data of the electrical equipment, and can identify potential electrical fire hazards in a timely manner, thereby minimizing the probability of electrical fire accidents and improving the safety of the operation area of ​​the construction power system.

[0110] Figure 3 This is a flow chart of a second embodiment of a construction power system operation risk treatment method provided by this application. Figure 3 As shown, based on the above embodiment, the acquisition process of the construction power consumption prediction model 9031 may include:

[0111] S301: Based on historical environmental monitoring data and historical safety monitoring data, a preset algorithm is used to determine the optimal parameter configuration scheme of an initial construction power consumption prediction model, where the initial construction power consumption prediction model is a neural network model.

[0112] In this step, before applying the construction power consumption prediction model 9031 to process the environmental monitoring data and the safety monitoring data of the electrical equipment, it is necessary to obtain the construction power consumption prediction model 9031 in advance.

[0113] Specifically, the construction power consumption prediction model 9031 is determined by the initial construction power consumption prediction model and parameters. The initial construction power consumption prediction model refers to the model framework when the specific parameters are not determined, which is mainly a neural network model.

[0114] Optionally, in a possible implementation, the initial construction power consumption prediction model includes an input layer, a hidden layer, and an output layer. The number of neurons in the input layer is determined according to the number of input values, the number of neurons in the hidden layer is determined according to the number of neurons in the input layer and the number of neurons in the output layer, and the number of neurons in the output layer is determined according to the number of predicted values.

[0115] Optionally, the number of neurons in the hidden layer can be determined using the following calculation formula:

[0116]

[0117] Among them, Number yc Indicates the number of neurons in the hidden layer, Number input Indicates the number of neurons in the input layer, Number output represents the number of neurons in the output layer, a represents a random integer, and its value range is generally 10 to 14.

[0118] Through the above calculation formula, the number of hidden layer neurons is determined according to the number of neurons in the input layer and the output layer, which helps to balance the complexity and generalization ability of the initial construction power consumption prediction model, avoid overfitting and underfitting, improve computational efficiency, and support effective feature extraction and learning representation, thereby optimizing the training effect of the construction power consumption prediction model 9031.

[0119] When training the initial construction power consumption prediction model, the parameters that need to be determined through training include but are not limited to weight parameters and bias parameters. In the existing technical solutions, the parameters to be optimized of the neural network model are usually obtained by training with a large amount of historical data. However, using historical data to train the neural network model to determine the parameters usually faces the problems of high time complexity and insufficient prediction accuracy.

[0120] In order to improve the training efficiency and prediction accuracy of the initial construction power consumption prediction model, this application adopts a preset algorithm to determine the parameters to be optimized in the initial construction power consumption prediction model, so as to determine the optimal parameter configuration scheme.

[0121] Among them, the preset algorithm can find a near-optimal parameter configuration in a short time through intelligent search and optimization strategies, thereby significantly reducing the computational complexity and time cost. Optionally, the preset algorithm includes but is not limited to a particle swarm optimization algorithm, a genetic algorithm, and a simulated annealing algorithm.

[0122] Specifically, by analyzing historical environmental monitoring data and historical safety monitoring data, a preset algorithm is used to determine the optimal parameter configuration scheme for the initial construction power consumption prediction model. This process ensures that the model can accurately reflect the actual situation, thereby improving the accuracy and reliability of the prediction. By optimizing the parameter configuration, the model can more effectively process real-time data, provide more accurate construction power consumption prediction and safety monitoring results, and provide technical support for safety management of construction sites.

[0123] Optionally, in a possible implementation, considering that the current operating state of the construction power system is closely related to the operating state in the past time period, the construction power prediction model 9031 used in this application can be based on a nonlinear autoregressive neural network model with exogenous inputs (Nonlinear Auto-Regressive model with Exogenous Inputs, referred to as NARX) to process environmental monitoring data and safety monitoring data of electrical equipment, and obtain safety prediction data. Below, the specific process of obtaining safety prediction data will be introduced using the NARX model as an example.

[0124] Specifically, before applying the NARX model to obtain safety prediction data, it is necessary to predetermine the network structure of the NARX model and the target parameter configuration scheme of the parameters to be optimized.

[0125] The network structure of the NARX model usually includes an input layer, a hidden layer, and an output layer. Among them, the hidden layer is the key part of the NARX model to handle complex tasks. In traditional technical solutions, the number of neurons in the hidden layer is usually determined based on artificial experience. However, this method based on artificial experience is prone to unstable model performance and is difficult to adapt to the needs of different data sets and application scenarios.

[0126] At the same time, the parameters to be optimized in the NARX model include but are not limited to the delay order, weight term parameters, and bias term parameters. Among them, the delay order determines the number of time steps of past input and output data used by the NARX model when making predictions. By properly selecting the delay order, the NARX model can more accurately model the time dependence and historical behavior of the system, thereby improving the accuracy and reliability of the prediction. If the delay order is not selected properly, the model may not be able to fully utilize historical information, affecting its prediction performance.

[0127] To this end, this application adopts an improved beetle whisker algorithm to determine the number of neurons and the delay order of the hidden layer in the NARX model to improve the performance and adaptability of the NARX model, thereby improving the prediction accuracy of the model.

[0128] First, initialize the number of neurons and the delay order of the hidden layer to form a network structure parameter configuration scheme. Then, determine the direction of the network structure parameter configuration scheme according to the network structure parameter configuration scheme. Then, determine the first position according to the network structure parameter configuration scheme, the direction of the network structure parameter configuration scheme and the preset length factor. Determine the second position according to the network structure parameter configuration scheme, the direction of the network structure parameter configuration scheme and the preset length factor. Finally, update the network structure parameter configuration scheme according to the preset fitness function and adaptive step size factor to obtain a new network structure parameter configuration scheme. At the end of the iteration, the network structure parameter configuration scheme corresponding to the lowest fitness value will be used as the optimal number of neurons in the hidden layer and the optimal delay order in the NARX model.

[0129] Optionally, the default fitness function can be set to:

[0130]

[0131] Among them, S tnx (l) represents the fitness value corresponding to the network structure parameter configuration scheme at the lth iteration, μ1, μ2 and μ3 are the preset first network structure weight parameter, second network structure weight parameter and third network structure weight parameter respectively, Number yc (l) represents the number of neurons in the hidden layer at the lth iteration, Indicates the mean square error value corresponding to the network structure parameter configuration scheme at the lth iteration.

[0132] Optionally, in the traditional beetle whisker algorithm, the variation range of the step size factor is a fixed value, which makes it difficult to adjust the step size in time as the number of iterations increases. If the step size factor gradually decreases, the global search capability of the algorithm is stronger, but the convergence speed is slower. If the step size factor decreases quickly, the global optimal solution cannot be obtained. Therefore, when applying the beetle whisker algorithm, the present application improves the algorithm by adopting an adaptive step size factor. Specifically, the calculation formula of the adaptive step size factor θ(l) is as follows:

[0133]

[0134] Among them, θ(l) represents the value of the step factor at the lth iteration, θ(l-1) represents the value of the step factor at the (l-1)th iteration, S tnx (l) represents the fitness value corresponding to the network structure parameter configuration scheme at the lth iteration, represents the global optimal fitness value, L represents the preset maximum number of iterations, ε1, ε2 and ε3 are the preset first parameter, second parameter and third parameter, and the first parameter is greater than the second parameter.

[0135] According to the above calculation formula, if the fitness value S corresponding to the current network structure parameter configuration scheme is tnx (l) is less than or equal to the global optimal fitness value, indicating that the search performance is good at this time, so the step size factor corresponding to the previous iteration is maintained in the next iteration. If the fitness value S corresponding to the current network structure parameter configuration scheme is tnx (l) When it is greater than the global optimal fitness value, it means that the search performance is poor at this time, and the step size factor should be appropriately reduced to increase the convergence speed. At the same time, as the algorithm iterates, the step size factor should also be gradually reduced to improve the accuracy of the algorithm.

[0136] It is worth noting that in the traditional NARX model, the NARX model takes the data of the current moment and a preset number of moments continuous with the current moment as input, thereby improving the accuracy of the prediction.

[0137] Based on this, the present application can divide the preset time period into a preset number of sub-time periods to meet the input requirements of the NARX model. For example, the preset time period is divided into a first preset time period and a second preset time period, the time lengths of the first preset time period and the second preset time period are consistent, and the time length between the current moment and moment 1 is recorded as the first preset time period, and the time length between moment 1 and moment 2 is recorded as the second preset time period. The computer device 1200 will divide and process the environmental monitoring data and the safety monitoring data of the electrical equipment uploaded by the collection layer device 1301 according to these two time periods. First, the average value of the environmental monitoring data between the current moment and moment 1 is calculated to obtain the average environmental monitoring data at the current moment. Then, the average value of the environmental monitoring data between moment 1 and moment 2 is calculated to obtain the average environmental monitoring data at moment 1. At the same time, the average value of the safety monitoring data of the electrical equipment between moment 1 and moment 2 is calculated to obtain the average safety monitoring data of the electrical equipment at moment 1. Finally, the input of the NARX model includes the average environmental monitoring data at the current moment, the average environmental monitoring data at moment 1, and the average safety monitoring data of the electrical equipment at moment 1. These data may include, but are not limited to, parameters such as acetylene concentration, aluminum oxide powder concentration, temperature, humidity, light intensity, sound intensity, and spark duration.

[0138] Optionally, in a possible implementation, in order to improve the accuracy of the NARX model in predicting safety data, an attention mechanism method may be pre-applied to process environmental monitoring data and safety monitoring data of electrical equipment.

[0139] In this embodiment, the attention mechanism can dynamically determine the contribution weight of the input data at each time step to the prediction result. Specifically, the attention mechanism can adopt a neural network-based model, accept the environmental monitoring data and electrical equipment safety monitoring data of the current moment and a preset number of moments before it as input, and generate an attention weight vector with a length equal to the number of input data, and each element represents the model's attention to the data at the corresponding moment. Among them, these weights are processed by the softmax function to ensure that their sum is 1. Based on these attention weights, the data at each moment is assigned different weights and used as input to the NARX model to improve the accuracy of safety predictions. In this way, the NARX model can dynamically adjust the impact of historical data at each moment, more effectively capture the most critical information for the current prediction, and thus improve the accuracy of the prediction results.

[0140] S302: According to the optimal parameter configuration scheme, the target parameters in the initial construction power consumption prediction model are configured to obtain the construction power consumption prediction model 9031.

[0141] In this step, according to step S301, after using the preset algorithm to determine the optimal parameter configuration scheme of the initial construction power consumption prediction model, it is necessary to configure the target parameters in the initial construction power consumption prediction model according to the optimal parameter configuration scheme to obtain the final construction power consumption prediction model 9031.

[0142] Specifically, the initial construction power consumption prediction model adopts a neural network model. According to the neural network structure and optimal parameter configuration scheme of the initial construction power consumption prediction model, the parameters of each corresponding position in the initial construction power consumption prediction model are determined. Optionally, the parameters of each corresponding position in the initial construction power consumption prediction model include but are not limited to weight parameters and bias item parameters.

[0143] By accurately configuring these parameters, an efficient construction power consumption prediction model 9031 for predicting the safety data of the construction power consumption system operation area can be obtained.

[0144] The method for handling the operational risks of the construction power system provided in the embodiment of the present application mainly provides a process for obtaining the construction power prediction model 9031. Specifically, the basic framework of the construction power prediction model 9031 is a neural network model. When determining the optimal parameter configuration scheme, historical environmental monitoring data and historical safety monitoring data are analyzed, and a preset algorithm is used for optimization. Among them, the preset algorithm has a strong global search capability, and can find a near-optimal parameter configuration in a relatively short time, significantly reducing the computational complexity and time cost, thereby improving the training efficiency of the model and ensuring the stability and accuracy of the model in various complex environments. At the same time, through this method, the target parameters in the construction power prediction model 9031 can be accurately determined, thereby significantly improving the prediction accuracy and reliability of the model.

[0145] Figure 4 This is a flow chart of a third embodiment of a construction power system operation risk treatment method provided by this application. Figure 4 As shown, based on any one of the above embodiments, the process of the construction power system operation risk processing method may include:

[0146] Step a: obtaining a sample data set based on historical environmental monitoring data and historical safety monitoring data, wherein the sample data set includes multiple data samples, each of which includes environmental monitoring data and safety monitoring data for a time period and safety monitoring data for the next time period of the time period.

[0147] In this step, in order to determine the optimal parameter configuration scheme of the construction power consumption prediction model 9031, it is necessary to obtain historical environmental monitoring data and historical safety monitoring data in advance.

[0148] Among them, the historical environmental monitoring data includes but is not limited to all historical average acetylene concentration monitoring data, historical average alumina powder concentration monitoring data, historical average temperature monitoring data, historical average humidity monitoring data and historical average light intensity monitoring data.

[0149] Historical safety monitoring data includes but is not limited to historical average sound intensity monitoring data and historical average spark duration monitoring data of electrical equipment.

[0150] After the historical environmental monitoring data and the historical safety monitoring data are acquired, a sample data set is determined based on the acquired historical environmental monitoring data and the historical safety monitoring data.

[0151] The sample data set includes multiple data samples, each of which includes environmental monitoring data and safety monitoring data of a time period and safety monitoring data of the next time period of the time period.

[0152] Optionally, before determining the sample data set based on the acquired historical environmental monitoring data and historical safety monitoring data, preprocessing operations may be performed on these data. The preprocessing operations include but are not limited to filling missing data and data normalization.

[0153] Specifically, in the process of obtaining historical environmental monitoring data and historical safety monitoring data, data may be missing due to various reasons. At the same time, different historical environmental monitoring data and historical safety monitoring data have different dimensions. In order to eliminate the influence of these data dimensions, the maximum and minimum values ​​of the historical environmental monitoring data and historical safety monitoring data can be obtained in advance, and all historical environmental monitoring data and historical safety monitoring data can be normalized according to these maximum and minimum values.

[0154] Through these preprocessing operations, the consistency and integrity of the data can be ensured, thereby improving the accuracy and reliability of subsequent analysis and modeling.

[0155] Step b, initialize N groups of parameter configuration schemes, N groups of offset configuration schemes, a first learning factor, a second learning factor, a third learning factor, and a preset maximum number of iterations, wherein N is a positive integer, and each group of parameter configuration schemes includes weight parameters and bias item parameters.

[0156] In this step, before applying the preset algorithm to determine the optimal parameter configuration scheme of the construction power consumption prediction model 9031, it is necessary to pre-initialize the relevant parameters involved in the preset algorithm. The relevant parameters involved in the preset algorithm include but are not limited to N groups of parameter configuration schemes, N groups of offset configuration schemes, the first learning factor, the second learning factor, the third learning factor, and the preset maximum number of iterations.

[0157] Specifically, N groups of parameter configuration schemes are used to initialize the parameter configuration of the construction power consumption prediction model 9031. Each group of schemes includes weight parameters and bias item parameters, etc. These parameters will be used as candidates for the initial solution. The weights and bias items in each group of parameter configuration schemes will provide initial settings for initializing the construction power consumption prediction model 9031 so that they can be adjusted and improved in the subsequent optimization process.

[0158] N groups of offset configuration schemes are used to initialize the offset of the parameter configuration of the construction power consumption prediction model 9031. These offset configuration schemes determine the update method and pace of the parameter configuration scheme, thereby affecting the convergence speed and final performance of the initialized construction power consumption prediction model 9031 during the training process.

[0159] By introducing multiple sets of parameter configurations and offset configuration schemes, diverse initial conditions can be provided for initializing the construction power consumption prediction model 9031, which helps to explore a wider solution space and increase the possibility of finding the global optimal solution.

[0160] The first learning factor is used to control the influence of the current offset configuration scheme in the update process. By adaptively adjusting the first learning factor, the capabilities of global search and local search can be effectively balanced, thereby improving the optimization efficiency of the model.

[0161] The second learning factor is used to control the degree to which the current parameter configuration scheme approaches the global optimal configuration scheme. By adaptively adjusting the second learning factor, the parameter configuration scheme can enhance its use of the global optimal configuration scheme, making the model approach the global optimal solution more quickly.

[0162] The third learning factor is used to control the degree to which the current parameter configuration scheme approaches the local optimal configuration scheme. By adaptively adjusting the third learning factor, the parameter configuration scheme can enhance the utilization of the local optimal configuration scheme, thereby improving the accuracy of the model in local search.

[0163] The preset maximum number of iterations refers to the maximum number of cycles that the preset algorithm runs, which is used to control the termination condition of the preset algorithm. By setting the maximum number of iterations, the preset algorithm can be prevented from falling into an infinite loop, while ensuring that the optimal or near-optimal parameter configuration solution is found within a reasonable time frame.

[0164] By adaptively adjusting these learning factors and setting the maximum number of iterations, the efficiency and effect of the preset algorithm can be improved, thereby optimizing the performance of the construction electricity consumption prediction model 9031.

[0165] Wherein, N is a positive integer, indicating the number of parameter configuration schemes and offset configuration schemes. The parameters in each set of parameter configuration schemes are key parameters of the construction power consumption prediction model 9031, and their initial values ​​will affect the final performance of the construction power consumption prediction model 9031. By reasonably initializing these parameters, the convergence speed and optimization effect of the preset algorithm can be improved.

[0166] In a specific implementation, N is 30 and the maximum number of iterations is 100.

[0167] Step c: Initialize N first construction power consumption prediction models with exactly the same structure to obtain N second construction power consumption prediction models.

[0168] In this step, according to step b, after initializing the relevant parameters of the preset algorithm, a first construction power consumption prediction model with the same number and structure as the parameter configuration scheme is constructed, thereby obtaining N second construction power consumption prediction models. Each second construction power consumption prediction model corresponds to a set of parameter configuration schemes.

[0169] In a possible implementation, each second construction power consumption prediction model includes 1 input layer, 1 hidden layer and 1 output layer, wherein the input layer includes 9 neurons, the hidden layer includes 13 neurons, and the output layer includes 1 neuron.

[0170] It should be noted that the number of neurons and the number of layers in the construction power consumption prediction model 9031 can be adjusted according to actual application requirements. For example, the number of neurons in the hidden layer can be increased to improve the expression ability of the construction power consumption prediction model 9031.

[0171] In the subsequent optimization process, the preset algorithm will iteratively update the parameters of the N second construction power consumption prediction models to find the optimal parameter configuration scheme of the construction power consumption prediction model 9031, thereby improving the prediction performance and accuracy of the construction power consumption prediction model 9031.

[0172] Step d: for each second construction power consumption prediction model, select a set of parameter configuration schemes, configure the initial parameters in the second construction power consumption prediction model, and obtain N initial construction power consumption prediction models.

[0173] In this step, after obtaining N second construction power consumption prediction models according to step c, it is necessary to select a set of parameter configuration schemes for each second construction power consumption prediction model and configure the initial parameters in these models to obtain N initial construction power consumption prediction models.

[0174] Specifically, the initial parameters in each initial construction power consumption prediction model include but are not limited to initial weight parameters and initial bias parameters, etc. These initial parameters are consistent with the number and meaning of parameters in the selected parameter configuration scheme, ensuring the consistency and accuracy of the model initialization process.

[0175] Through this method, a reasonable initial state can be provided for each second construction power consumption prediction model, so that it can converge to the optimal solution more quickly in the subsequent training and optimization process. At the same time, selecting multiple sets of parameter configuration schemes for initialization can increase the diversity and robustness of the model and improve the stability and reliability of the overall prediction system.

[0176] Step e: select N groups of data samples from the sample data set, input the environmental monitoring data and safety monitoring data of a time period in each group of data samples into N initial construction power consumption prediction models respectively to predict the safety monitoring data of the next time period, and obtain N groups of safety prediction data.

[0177] In this step, according to step d, after obtaining N initial construction power consumption prediction models, it is necessary to select N data samples from the sample data set obtained in step a. Then, the environmental monitoring data and safety monitoring data of a time period in the N data samples are respectively input into the N initial construction power consumption prediction models to realize the prediction of the safety monitoring data of the next time period, thereby obtaining N sets of safety prediction data.

[0178] In this way, the historical environmental monitoring data, historical safety monitoring data and the prediction capability of the initial construction power consumption prediction model can be fully utilized to accurately predict the safety data for the next time period. This prediction method not only improves the prediction accuracy and reliability of the model, but also effectively prevents the occurrence of electrical fire accidents.

[0179] Step f: Calculate the fitness value of each set of parameter configuration schemes based on the N sets of safety prediction data and the safety monitoring data of the next time period in the N sets of data samples.

[0180] In this step, after obtaining N groups of safety prediction data according to step e, it is necessary to combine the actual safety monitoring data of the next time period in the data sample and the fitness function corresponding to the preset algorithm to calculate the fitness value of each group of parameter configuration schemes.

[0181] Specifically, by comparing N groups of safety prediction data with actual safety monitoring data and using a preset fitness function, the prediction performance of each initial construction power consumption prediction model can be determined. The fitness function is usually used to measure the difference between the predicted data and the actual data. Common fitness functions include but are not limited to mean square error and mean absolute error.

[0182] By calculating the fitness value of each set of parameter configuration schemes, we can identify which parameter configuration schemes perform better in the prediction task, thus providing a basis for subsequent model optimization. This process not only helps to improve the prediction accuracy and reliability of the model, but also accelerates the convergence speed of the model and ensures its effectiveness in practical applications.

[0183] Step g: Obtain the parameter configuration scheme corresponding to the highest fitness value and record it as the target parameter configuration scheme.

[0184] In this step, after obtaining the fitness value of each group of parameter configuration schemes according to step f, it is necessary to sort the N groups of parameter configuration schemes according to these fitness values.

[0185] Specifically, the fitness values ​​may be arranged in order from high to low or from low to high, so as to find the parameter configuration scheme with the highest fitness value and mark it as the target parameter configuration scheme.

[0186] Through this sorting and marking process, the optimal parameter configuration scheme for the current iteration can be clearly determined. This target parameter configuration scheme represents the parameter setting with the best model prediction performance under the current iteration conditions, providing a clear direction for subsequent model optimization and adjustment.

[0187] At the same time, marking the target parameter configuration scheme not only helps to record the optimal configuration during the model optimization process, but also provides a reference for the next parameter update and model improvement. Through continuous iteration and optimization, the construction power consumption prediction model 9031 will gradually move towards the global optimal solution, thereby significantly improving its prediction accuracy and reliability.

[0188] Step h: updating each group of offset configuration schemes according to the first learning factor, the second learning factor, the third learning factor, the target parameter configuration scheme, and each group of parameter configuration schemes to obtain N groups of new offset configuration schemes.

[0189] In this step, after obtaining the fitness value corresponding to each group of parameter configuration schemes according to step f, it is necessary to combine the first learning factor, the second learning factor, the third learning factor, the target parameter configuration scheme and each group of parameter configuration schemes to update each group of offset configuration schemes, thereby generating N groups of new offset configuration schemes.

[0190] Specifically, by using the first learning factor, the influence of the current offset configuration scheme in the update process can be controlled to ensure the balance between global search and local search. The second learning factor is used to adjust the degree of proximity of the current parameter configuration scheme to the target parameter configuration scheme, and enhance the utilization of the global optimal configuration scheme. The third learning factor is used to adjust the degree of proximity of the current parameter configuration scheme to the local optimal configuration scheme, and enhance the utilization of the local optimal configuration scheme.

[0191] During the update process, by combining the target parameter configuration scheme and each group of parameter configuration schemes, each group of new offset configuration schemes can be calculated according to the preset calculation formula. These new offset configuration schemes will provide better initial conditions for the next round of iterations, further improving the prediction performance of the model.

[0192] Through this method of dynamically updating and optimizing the offset configuration scheme, the parameter setting of the construction power consumption prediction model 9031 can be continuously improved so that it gradually approaches the optimal solution.

[0193] Step i: According to each set of new offset configuration schemes, update each set of parameter configuration schemes to obtain N sets of updated parameter configuration schemes.

[0194] In this step, after obtaining the updated N groups of new offset configuration schemes according to step h, each group of parameter configuration schemes is updated to obtain the updated N groups of parameter configuration schemes.

[0195] In a specific implementation, the update process of each group of parameter configuration schemes is calculated using the following formula:

[0196]

[0197] in, Indicates the value of the sth parameter in the nth parameter configuration scheme at the (k+1)th iteration, Indicates the value of the sth parameter in the nth parameter configuration scheme at the kth iteration, It represents the offset value corresponding to the sth parameter in the nth group of offset configuration schemes at the (k+1)th iteration, and θ and τ are preset weight parameters.

[0198] Step j: Repeat steps d to i until the maximum number of iterations is reached to obtain the optimal parameter configuration solution.

[0199] In this step, after obtaining the updated N groups of parameter configuration schemes according to step i, it is necessary to determine whether the current number of iterations has reached the preset maximum number of iterations.

[0200] If the preset maximum number of iterations is not reached, the N groups of updated parameter configuration schemes are determined as new N groups of parameter configuration schemes, and steps d to i are repeated until the preset maximum number of iterations is reached.

[0201] If the preset maximum number of iterations is reached, the target parameter configuration scheme found in the current iteration process is determined as the optimal parameter configuration scheme. The optimal parameter configuration scheme is the optimal solution obtained after multiple iterations and optimization processes, and can be used to configure the target parameters in the initial construction power consumption prediction model.

[0202] This iterative optimization method ensures that the construction power consumption prediction model 9031 is continuously improved in each iteration, and finally finds the optimal parameter configuration scheme. This optimal parameter configuration scheme will significantly improve the prediction accuracy and reliability of the model, and provide technical support for safety monitoring and electrical fire prevention at the construction site.

[0203] Optionally, in a possible implementation, after obtaining the updated N groups of parameter configuration schemes based on step i, the fitness values ​​of the N groups of parameter configuration schemes that have not been updated are compared with the fitness values ​​of the N groups of parameter configuration schemes that have been updated to ensure that there are parameter configuration schemes with better quality in the next iteration. Specifically, for each group of unupdated parameter configuration schemes and each group of updated parameter configuration schemes, the parameter configuration scheme with high fitness is used as the parameter configuration scheme after screening, thereby obtaining N groups of parameter configuration schemes after screening. These parameter configuration schemes after screening will be used as new N groups of parameter configuration schemes to participate in the next step of optimization and iteration to gradually improve the prediction performance and accuracy of the initial construction power consumption prediction model.

[0204] Optionally, in a possible implementation, after obtaining the updated N sets of parameter configuration schemes based on step i, the fitness values ​​of the updated N sets of parameter configuration schemes are calculated and arranged in order from high to low or from low to high with the fitness values ​​of the N sets of parameter configuration schemes that have not been updated. The parameter configuration scheme of each parameter is used as the first part of the parameter configuration scheme.

[0205] After the first part of the parameter configuration scheme is determined, the remaining parameter configuration schemes are determined from the original 2N groups of parameter configuration schemes by tournament selection. The specific steps are as follows: First, a predefined number of parameter configuration schemes are randomly selected from the N groups of updated parameter configuration schemes and the N groups of unupdated parameter configuration schemes. Then, the parameter configuration scheme with the highest fitness value is selected as the first parameter configuration scheme among the remaining parameter configuration schemes. Repeat the above steps until all the remaining parameter configuration schemes are determined. In the above manner, the diversity of parameter configuration schemes is ensured, and the performance of the initial construction power consumption prediction model is significantly improved. Each time, a random selection is made from the updated and unupdated schemes to ensure that the selected parameter configuration schemes contain various changes and strategies, thereby increasing the diversity of the parameter space. This approach effectively avoids over-reliance on a specific type of parameter configuration and significantly improves the effect of global search. At the same time, each time the scheme with the highest fitness value is selected as the next determined parameter configuration scheme, which helps to gradually optimize the overall performance. This step-by-step optimization strategy can find a better solution within a limited number of iterations and avoid falling into the dilemma of a local optimal solution. The initial random selection and subsequent fitness judgment effectively reduce the bias that may be caused by a fixed selection strategy. This randomness can help explore the parameter space more comprehensively during the search process, thereby increasing the chance of discovering excellent parameter configurations.

[0206] The method for handling the operation risk of the construction power system provided in the embodiment of the present application determines the target parameter configuration scheme of the construction power prediction model 9031 based on a preset algorithm. Specifically, within the preset maximum number of iterations, the parameter configuration scheme is repeatedly updated and compared until the target parameter configuration scheme is found. Through this iterative optimization process, the preset algorithm can gradually converge to the optimal solution, thereby providing a reliable foundation for constructing a high-precision construction power prediction model 9031.

[0207] Figure 5 This is a flow chart of a fourth embodiment of a construction power system operation risk treatment method provided by this application. Figure 5 As shown, based on any of the above embodiments, whether an electrical fire accident will occur in the next time period is determined according to the safety prediction data. The process of the construction power system operation risk processing method may include:

[0208] S501: If the average sound intensity data in the safety prediction data is higher than the preset sound intensity threshold data or the average spark duration data is higher than the preset spark duration threshold data, it is determined that an electrical fire accident will occur in the construction power system operation area in the next time period.

[0209] In this step, after obtaining the safety prediction data for the next time period based on the construction power consumption prediction model 9031, the average sound intensity data in the safety prediction data is compared with the preset sound intensity threshold data, and the average spark duration data in the safety prediction data is compared with the preset spark duration threshold data to determine whether an electrical fire accident will occur in the construction power consumption system operation area in the next time period.

[0210] The method for determining the preset sound intensity threshold data includes but is not limited to determining based on historical sound intensity data or laboratory tests. Specifically, based on the collection and analysis of a large amount of historical sound intensity data, the preset sound intensity threshold data is determined based on calculation methods such as mean and standard deviation. Alternatively, an experiment is conducted in a controlled environment in advance to simulate the explosion of electrical equipment, measure its sound intensity, and determine the preset sound intensity threshold data based on the experimental results.

[0211] Methods for determining the preset cremation duration threshold data include but are not limited to determining based on historical spark duration data or laboratory tests, etc. Specifically, based on the collection and analysis of a large amount of spark duration data under normal and abnormal conditions, the preset cremation duration threshold data is determined according to calculation methods such as mean and standard deviation, or experiments are conducted in a controlled environment in advance to simulate spark durations under different conditions, and the preset cremation duration threshold data is determined based on the experimental results.

[0212] Specifically, when the average sound intensity data in the safety prediction data is higher than the preset sound intensity threshold data or the average spark duration data is higher than the preset spark duration threshold data, it can be determined that an electrical fire accident will occur in the construction power system operation area in the next time period.

[0213] Based on this, managers will take further preventive and handling measures to avoid electrical fire accidents and ensure the safety of the construction power system operation area.

[0214] Optionally, when the average sound intensity data in the safety prediction data is lower than or equal to the preset sound intensity threshold data and the average spark duration data is lower than or equal to the preset spark duration threshold data, it can be determined that no electrical fire accident will occur in the construction power system operation area in the next time period.

[0215] The construction power system operation risk processing method provided in the embodiment of the present application determines whether an electrical fire accident will occur in the construction power system operation area in the next time period by comparing the high and low relationship between the preset sound intensity threshold data and the preset spark duration threshold data and the acquired average sound intensity data and the average spark duration data. The method can predict in advance the average sound intensity and average spark duration of electrical equipment in the construction power system operation area in the next time period, so as to take preventive and response measures in time when necessary to ensure the safety of the construction power system operation area.

[0216] Figure 6 This is a flow chart of a fifth embodiment of a construction power system operation risk treatment method provided by this application. Figure 6 As shown, based on any one of the above embodiments, the process of the construction power system operation risk processing method also includes:

[0217] S601: If it is determined that an electrical fire accident will occur in the construction power system operation area in the next time period, the power supply voltage of the construction power system is automatically adjusted, and a safety warning message is pushed to the terminal device of the manager, wherein adjusting the power supply voltage of the construction power system includes controlling the power supply voltage of the construction power system to be reduced to a preset safe power supply voltage, or controlling the power supply of the construction power system to be turned off, and the safety warning message is used to indicate that an electrical fire accident will occur in the construction power system operation area.

[0218] In this step, according to step S501, if it is determined that an electrical fire accident will occur in the construction power system operation area in the next time period, the computer device 1200 will automatically execute a series of emergency response measures. First, the computer device 1200 will automatically adjust the power supply voltage of the construction power system to reduce it to a preset safe power supply voltage, or directly control the power supply of the construction power system to shut down, so as to reduce the possibility and severity of the electrical fire. Adjusting the power supply voltage can reduce the load of electrical equipment, thereby reducing the risk of overheating and arcing, and further preventing the occurrence of fire.

[0219] At the same time, the computer device 1200 will also push safety warning information to the terminal device of the manager. The safety warning information is used to indicate that an electrical fire accident will occur in the operation area of ​​the construction power system, and usually includes the currently detected risk status, specific risk location, possible cause of the accident, and recommended response measures. By sending warning information to managers in a timely manner, it can be ensured that they quickly understand the current risk situation and take necessary emergency measures, such as evacuating personnel, cutting off power, or starting the fire extinguishing system.

[0220] The implementation of this dual measure, namely, automatically adjusting the power supply voltage and pushing safety warning information, can ensure the safety of the construction power system operation area to the greatest extent, reduce the probability of electrical fire accidents and potential losses. Through this method, the safety management level of the construction site has been significantly improved, ensuring the smooth progress of the construction process.

[0221] The embodiment of the present application provides a method for handling the operation risks of a construction power system. When an electrical fire accident is detected, the computer device 1200 will automatically adjust the power supply voltage of the construction power system while pushing safety warning information to the terminal device of the manager. Through this method, it is possible to ensure that the manager responds in a timely manner, minimize the possible losses and hazards, and ensure the safe operation of the construction power system. At the same time, the automated emergency response measures of the computer device 1200 not only improve the response speed, but also reduce the delays and errors that may be caused by human operation. Managers can monitor the system status in real time through terminal devices, obtain detailed accident analysis reports and response suggestions, and make more scientific and effective decisions.

[0222] Figure 7 This is a flow chart of a sixth embodiment of a construction power system operation risk treatment method provided by this application. Figure 7 As shown, based on any of the above embodiments, according to N groups of safety prediction data and the safety monitoring data of the next time period in the N groups of data samples, the fitness value of each group of parameter configuration schemes is calculated, and the process of the construction power system operation risk processing method includes:

[0223] S701: For each set of parameter configuration schemes, the mean square error value, the root mean square error value and the mean absolute error value are calculated based on the safety prediction data and the safety monitoring data of the next time period in the data sample.

[0224] In this step, for each set of parameter configuration schemes, after the corresponding safety prediction data is obtained based on the initial construction power consumption prediction model, the relevant error index is calculated based on the safety prediction data and the safety monitoring data of the next time period in the data sample. The error index reflects the degree of difference between the safety prediction data and the safety monitoring data of the next time period in the data sample, including but not limited to the mean square error value, the root mean square error value, and the mean absolute error value.

[0225] Specifically, the mean square error value is measured by calculating the average of the sum of squares of the differences between the safety prediction data and the safety monitoring data of the next time period in the data sample. The mean square error value can reflect the overall level of prediction error, but due to the square, it is more sensitive to larger errors. Therefore, the mean square error value can highlight larger prediction deviations when judging the prediction model.

[0226] Optionally, in a possible implementation manner, the mean square error value of each group of parameter configuration solutions can be calculated using the following formula:

[0227]

[0228] Among them, G cs (n,k) represents the mean square error value of the nth group of parameter configuration schemes at the kth iteration, represents the preset first weight parameter, D represents the dimension of the security prediction data, and W n,k,d It means the value of the d-th dimension of the security prediction data under the n-th parameter configuration scheme at the k-th iteration. It represents the value of the d-th dimension data in the safety monitoring data of the next time period in the data sample under the n-th parameter configuration scheme at the k-th iteration.

[0229] The RMS error value is the square root of the mean square error value, which provides a more intuitive error metric to facilitate understanding and comparison of the prediction performance of the initial construction power consumption prediction model with different parameter configuration schemes.

[0230] Optionally, in a possible implementation manner, the root mean square error value of each group of parameter configuration solutions can be calculated using the following formula:

[0231]

[0232] Among them, H cs(n,k) represents the root mean square error value of the nth group of parameter configuration schemes at the kth iteration, represents the preset second weight parameter, D represents the dimension of the security prediction data, and W n,k,d It means the value of the d-th dimension of the security prediction data under the n-th parameter configuration scheme at the k-th iteration. It represents the value of the d-th dimension data in the safety monitoring data of the next time period in the data sample under the n-th parameter configuration scheme at the k-th iteration.

[0233] The mean absolute error is the average of the absolute differences between the safety prediction data and the safety monitoring data for the next time period in the data sample. The mean absolute error gives equal weight to all errors, providing a more robust error measure.

[0234] Optionally, in a possible implementation manner, the mean absolute error value of each group of parameter configuration schemes may be calculated using the following formula:

[0235]

[0236] Among them, Q cs (n,k) represents the mean absolute error of the nth group of parameter configuration schemes at the kth iteration. represents the preset third weight parameter, D represents the dimension of the security prediction data, and W n,k,d It means the value of the d-th dimension of the security prediction data under the n-th parameter configuration scheme at the k-th iteration. It represents the value of the d-th dimension data in the safety monitoring data of the next time period in the data sample under the n-th parameter configuration scheme at the k-th iteration.

[0237] By calculating these error indicators, the prediction performance of each parameter configuration scheme can be comprehensively judged. In practical applications, the calculation and analysis of these error indicators can identify the optimal parameter configuration scheme, thereby improving the accuracy and stability of the construction power consumption prediction model 9031.

[0238] S702: Calculate the fitness value of the parameter configuration scheme according to the mean square error value, the root mean square error value and the mean absolute error value.

[0239] In this step, according to step S701, the fitness value of the parameter configuration scheme is calculated based on the mean square error value, the root mean square error value and the mean absolute error value to determine the performance of each group of parameter configuration schemes in the initial construction power consumption prediction model.

[0240] The fitness value is an important indicator to measure the quality of parameter configuration schemes. It can lay the foundation for determining the optimal parameter configuration scheme and improve the accuracy and stability of the construction power consumption prediction model 9031.

[0241] Specifically, the fitness value of each parameter configuration scheme is calculated by comprehensively considering the mean square error value, the root mean square error value and the mean absolute error value. The fitness value calculation method includes but is not limited to the weighted average method, normalization processing, etc.

[0242] Optionally, in a possible implementation manner, the fitness value of each group of parameter configuration schemes may be calculated using the following formula:

[0243]

[0244] in, Indicates the fitness value of the nth group of parameter configuration schemes at the kth iteration, G cs (n,k) represents the mean square error value of the nth group of parameter configuration schemes at the kth iteration, H cs (n,k) represents the root mean square error value of the nth group of parameter configuration schemes at the kth iteration, Q cs (n,k) represents the mean absolute error value of the nth group of parameter configuration schemes at the kth iteration, ∈ represents the preset first weight parameter, δ represents the preset second weight parameter, and ρ represents the preset third weight parameter.

[0245] Through this calculation formula, the comprehensive application of mean square error, root mean square error and mean absolute error as the calculation basis of the fitness function helps to comprehensively judge the prediction accuracy and stability of the model in the process of optimizing parameter configuration. By considering different error measurement methods, the degree of fit of the model to the training data can be more comprehensively measured, thereby improving the accuracy and reliability of selecting the optimal parameter configuration solution.

[0246] The embodiment of the present application provides a method for handling the operation risk of a construction power system, and calculates the fitness value corresponding to each group of parameter configuration schemes based on the mean square error value, root mean square error value and mean absolute error value of each group of parameter configuration schemes. Among them, the size of the fitness value directly reflects the quality of each group of parameter configuration schemes. By calculating the fitness value, the advantages and disadvantages of different parameter configuration schemes can be quantified and compared, thereby laying the foundation for determining the target parameter configuration scheme.

[0247] Figure 8 This is a flow chart of Embodiment 7 of a construction power system operation risk treatment method provided by this application. Figure 8 As shown, based on any of the above embodiments, each group of offset configuration schemes is updated according to the first learning factor, the second learning factor, the third learning factor, the target parameter configuration scheme, and each group of parameter configuration schemes to obtain N groups of new offset configuration schemes. The process of the construction power system operation risk processing method includes:

[0248] S801: For each group of offset configuration schemes, calculate a first offset value according to a first learning factor and the offset configuration scheme.

[0249] In this step, when the parameter configuration scheme needs to be updated, each group of offset configuration schemes needs to be updated in advance.

[0250] Specifically, for each group of offset configuration schemes, a first offset value of each group of offset configuration schemes is calculated according to the first learning factor and the corresponding offset configuration scheme.

[0251] Optionally, in a possible implementation manner, the first offset value may be calculated using the following formula:

[0252]

[0253] in, Indicates the first offset value corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the first learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, R py (k,n) represents the nth group of offset configuration schemes at the kth iteration.

[0254] In the process of calculating the first offset value, the offset configuration scheme and the first learning factor are taken into account simultaneously, so that the update of the offset configuration scheme can be more in line with the actual situation, ensuring that the search direction and step size can be adjusted according to the setting of the first learning factor during the algorithm iteration, thereby more effectively guiding the preset algorithm to converge quickly towards the optimal solution, while avoiding unnecessary oscillations and improving the stability and accuracy of the preset algorithm.

[0255] It is worth noting that the first learning factor is a preset first weight parameter, which is used to adjust the influence of the offset configuration scheme in the calculation process. The first learning factor can be set according to the specific application scenario and historical data to ensure the rationality and reliability of the calculation results.

[0256] Optionally, in a possible implementation manner, the first learning factor may be determined according to a preset maximum first learning factor, a preset minimum first learning factor, a current number of iterations, and a maximum number of iterations, and may be calculated using the following formula:

[0257]

[0258] in, Indicates the first learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset maximum first learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset minimum first learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, k represents the current number of iterations, k max Indicates the preset maximum number of iterations.

[0259] Through the above calculation formula, it can be ensured that the first learning factor is larger in the early stage of the preset algorithm, and smaller in the later stage of the preset algorithm. This adjustment mechanism can ensure that the preset algorithm has a strong global search capability in the early stage of iteration, while avoiding convergence to the local optimal solution in the later stage of iteration.

[0260] S802: Calculate a second offset value according to a second learning factor, a preset first random number, and a parameter configuration scheme.

[0261] In this step, based on step S801, after the first offset value is calculated, for each group of offset configuration schemes, a second offset value is calculated according to the second learning factor, the preset first random number, and the parameter configuration scheme.

[0262] Optionally, in a possible implementation manner, the second offset value may be calculated using the following formula:

[0263]

[0264] in, Indicates the second offset value corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset first random number at the kth iteration, Indicates the second learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, It represents the global optimal parameter configuration scheme of the nth group of parameter configuration schemes at the kth iteration, P py (k,n) represents the nth group of parameter configuration schemes at the kth iteration.

[0265] According to the second offset value, the trade-off between local search and global exploration can be balanced. By controlling the impact of the global optimal parameter configuration scheme on the current search space, it helps to optimize the model to find the optimal parameter configuration scheme among the parameter configuration schemes, improve the global search capability of the model, and ensure the quality of the optimal parameter configuration scheme.

[0266] It is worth noting that the second learning factor is a preset second weight parameter, which is used to adjust the influence of the global optimal parameter configuration scheme in the process of updating the offset configuration scheme.

[0267] Optionally, in a possible implementation manner, the second learning factor is determined according to a preset maximum second learning factor, a preset minimum second learning factor, a current number of iterations, a maximum number of iterations, and a preset cosine function, and can be calculated using the following formula:

[0268]

[0269] in, Indicates the second learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset maximum second learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset minimum second learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, k represents the current number of iterations, k max Indicates the preset maximum number of iterations.

[0270] Through the above calculation formula, in the early stage of the preset algorithm, the second learning factor is set larger, so that the offset configuration scheme relies more on the global optimal parameter configuration scheme of the group for extensive exploration, so as to quickly find potential excellent areas. At the same time, by emphasizing the experience of the group in the early stage, it is possible to avoid the offset configuration scheme from concentrating on a local optimal solution too early, thereby improving the robustness and global search ability of the preset algorithm.

[0271] S803: Calculate a third offset value according to a third learning factor, a preset second random number, and a target parameter configuration scheme.

[0272] In this step, based on step S801, after the second offset value is calculated, for each group of offset configuration schemes, a third offset value is calculated according to the third learning factor, the preset second random number, and the target parameter configuration scheme.

[0273] Optionally, in a possible implementation manner, the third offset value may be calculated using the following formula:

[0274]

[0275] in, Indicates the third offset value corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset second random number at the kth iteration, Indicates the third learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, It represents the local optimal parameter configuration scheme of the nth group of parameter configuration schemes at the kth iteration, P py (k,n) represents the nth group of parameter configuration schemes at the kth iteration.

[0276] According to the third offset value, the depth and breadth of the local search can be effectively balanced, helping the preset algorithm to more accurately adjust the offset configuration scheme in the local area, thereby improving the local search capability and convergence speed of the model.

[0277] It is worth noting that the third learning factor is a preset third weight parameter, which is used to adjust the influence of the local optimal parameter configuration scheme in the process of updating the offset configuration scheme.

[0278] Optionally, in a possible implementation manner, the third learning factor is determined according to a preset maximum third learning factor, a preset minimum third learning factor, a current number of iterations, a maximum number of iterations, and a preset sine function, and can be calculated using the following formula:

[0279]

[0280] in, Indicates the third learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset maximum third learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset minimum third learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, k represents the current number of iterations, k max Indicates the preset maximum number of iterations.

[0281] Through the above calculation formula, the size of the third learning factor is dynamically adjusted at different stages of the preset algorithm. Specifically, in the early stage of the algorithm, the third learning factor is small, so that the offset configuration scheme relies more on the global optimal parameter configuration scheme of the group for extensive exploration, avoiding falling into the local optimum too early. In the later stage of the algorithm, the third learning factor gradually increases, so that the offset configuration scheme relies more on the local optimal parameter configuration scheme of the group for fine optimization, improving the convergence speed and accuracy, thereby improving the efficiency and effect of the preset algorithm, ensuring the optimality of the offset configuration scheme and the improvement of system performance.

[0282] S804: Obtain a new offset configuration scheme according to the first offset value, the second offset value, and the third offset value.

[0283] In this step, after the first offset value, the second offset value and the third offset value are calculated based on step S801, step S802 and step S803, each group of offset configuration schemes is updated.

[0284] Optionally, in a possible implementation manner, for each group of offset configuration schemes, the following formula may be used for calculation:

[0285]

[0286] Among them, R py (k+1,n) indicates the nth group of offset configuration schemes at the (k+1)th iteration. Indicates the first offset value corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the second offset value corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the third offset value corresponding to the nth group of offset configuration schemes at the kth iteration, ∈ 1 Represents the preset first offset weight parameter, ∈ 2 Represents the preset second offset weight parameter, ∈ 3 Indicates the preset third offset weight parameter.

[0287] In this formula, by dynamically adjusting the first offset value, the second offset value, and the third offset value, each group of offset configuration schemes is updated, so that each group of schemes adopts different search strategies in different iteration stages, thereby enhancing the adaptability and flexibility of the preset algorithm. This method not only improves the global search capability of the preset algorithm and avoids falling into the local optimal solution, but also enhances the robustness and adaptability of the preset algorithm and optimizes the convergence speed and solution quality.

[0288] The embodiment of the present application provides a method for handling the operation risks of a construction power system and provides a method for updating the offset configuration scheme. Specifically, through a multi-stage, multi-strategy dynamic adjustment method, it is ensured that the preset algorithm can perform optimally under various conditions, thereby significantly improving the overall efficiency and accuracy, and ultimately achieving a more robust and efficient parameter optimization of the offset configuration scheme.

[0289] Fig. 9 This is a schematic diagram of a construction power system operation risk management system framework provided in this application. Fig. 9 As shown, the construction power system operation risk management system 1300 includes a collection layer 901 , a transmission layer 902 , a processing layer 903 and an application layer 904 .

[0290] The collection layer 901 includes sensors 9011, monitoring equipment 9012, probes 9013 and other sensing equipment 9014, which are mainly used to collect construction power monitoring data in the construction power system operation area in real time. These devices work together to form a comprehensive monitoring network to ensure that all aspects of the construction power system are effectively monitored. The construction power system operation area includes but is not limited to hospitals, schools, chemical industry areas, large shopping malls and residential areas. These areas each have independent electricity demand and safety requirements, so they need to be monitored and managed during the operation of the construction power system.

[0291] Specifically, the sensor 9011 includes but is not limited to an acetylene concentration sensor 13011, an aluminum oxide powder concentration sensor 13012, a temperature and humidity sensor 13013, a light intensity sensor 13014 and a spark detection sensor 13016, etc., which are used to monitor electrical parameters and ambient environmental conditions in real time.

[0292] Monitoring equipment 9012 includes but is not limited to cameras, video monitoring systems, etc., which can observe the conditions of the construction site in real time and provide visual data support. Probes 9013 can penetrate into the electrical equipment to obtain more detailed and accurate operating data. Other sensor equipment 9014 includes but is not limited to smoke detectors, flame detectors, etc., which can detect potential fire hazards in a timely manner.

[0293] Through the coordinated work of these devices, the collection layer 901 can collect various monitoring data of the construction power system operation area in real time and comprehensively. These data will be transmitted to the processing layer 903 through the mobile network 9021 and / or wireless network 9022 of the transmission layer 902 for analysis and processing, so as to further determine whether an electrical fire accident will occur in the construction power system operation area in the next time period.

[0294] The transport layer 902 includes a mobile network 9021 and a wireless network 9022, which are mainly used to transmit the data collected by the collection layer 901 to the processing layer 903. Specifically, the mobile network 9021 may include cellular networks such as 4G and 5G, which have the characteristics of wide coverage, high speed and low latency, and can achieve reliable data transmission in various complex environments. The wireless network 9022 includes technologies such as Wi-Fi and LoRa, which are suitable for efficient data transmission in local areas, especially inside buildings or dense areas, and can provide high-bandwidth and low-power connections.

[0295] Through the combination of mobile network 9021 and wireless network 9022, transport layer 902 can flexibly respond to data transmission requirements in different scenarios. For example, in a large construction site, mobile network 9021 can cover the entire area to ensure that monitoring data from every corner can be transmitted in a timely manner. Indoors or in local areas, wireless network 9022 can provide higher transmission rates and stability.

[0296] The processing layer 903 includes a construction power consumption prediction model 9031 deployed on the cloud platform, which is used to receive and process various monitoring data of the construction power system operation area transmitted by the transmission layer 902 to determine whether an electrical fire accident will occur in the construction power system operation area in the next time period.

[0297] The application layer 904 includes a smart construction site 9041 and a visual display 9042. Among them, the smart construction site 9041 can realize electrical fire early warning 90411 and power safety guarantee 90412. The visual display 9042 can realize automatic processing 90421, intelligent control 90422, construction quality 90423 and operation and maintenance management 90424. Specifically, the smart construction site 9041 can predict electrical fires, issue early warnings, and ensure power safety at the construction site. If a danger is predicted, the visual display 9042 will automatically adjust and intelligently control the voltage or cut off the power to ensure the quality of the construction.

[0298] Fig.10 A schematic diagram of a process for determining parameters of a construction power consumption prediction model provided in this application. Fig.10 As shown, the process of determining the parameters may include:

[0299] S1001: Start;

[0300] S1002: Establishing an initial construction power consumption prediction model;

[0301] S1003: Initialize the parameters of the preset algorithm;

[0302] S1004: Calculate the fitness values ​​of all parameter configuration schemes;

[0303] S1005: Update the local optimum and global optimum of the parameter configuration scheme to see whether they meet the preset maximum number of iterations;

[0304] Specifically, if the current number of iterations meets the preset maximum number of iterations, step S1006 is executed. If the current number of iterations does not meet the preset maximum number of iterations, steps S1004 to S1005 are repeatedly executed.

[0305] S1006: Establish construction power consumption prediction model 9031;

[0306] S1007: End.

[0307] Fig.11 This is a structural schematic diagram of a first embodiment of a construction power system operation risk management device provided in the present application.

[0308] like Fig.11 As shown, the construction power system operation risk processing device 1100 includes:

[0309] The acquisition module 1101 is used to acquire the environmental monitoring data of the construction power system operation area in a preset time period and the safety monitoring data of the electrical equipment based on the data uploaded by the collection layer device 1301, wherein the preset time period is a period of time including the current moment and before the current moment, the environmental monitoring data includes the average acetylene concentration monitoring data, the average aluminum oxide powder concentration monitoring data, the average temperature monitoring data, the average humidity monitoring data and the average light intensity monitoring data, and the safety monitoring data includes the average sound intensity monitoring data and the average spark duration monitoring data of the electrical equipment;

[0310] Processing module 1102, used to input environmental monitoring data and safety monitoring data into construction power consumption prediction model 9031 for prediction processing to obtain safety prediction data for the next time period;

[0311] The processing module 1102 is also used to determine whether an electrical fire accident will occur in the next time period based on the safety prediction data.

[0312] Optionally, the processing module 1102 is further used to: determine the optimal parameter configuration scheme of the initial construction power consumption prediction model using a preset algorithm according to the historical environmental monitoring data and the historical safety monitoring data, and the initial construction power consumption prediction model is a neural network model;

[0313] According to the optimal parameter configuration scheme, the target parameters in the initial construction power consumption prediction model are configured to obtain the construction power consumption prediction model 9031.

[0314] Optionally, the acquisition module 1101 is also used for: step a, acquiring a sample data set based on historical environmental monitoring data and historical safety monitoring data, the sample data set including multiple data samples, each data sample including environmental monitoring data and safety monitoring data for a time period and safety monitoring data for the next time period of the time period.

[0315] Optionally, the processing module 1102 is further used for: step b, initializing N groups of parameter configuration schemes, N groups of offset configuration schemes, a first learning factor, a second learning factor, a third learning factor, and a preset maximum number of iterations, wherein N is a positive integer, and each group of parameter configuration schemes includes a weight parameter and a bias item parameter;

[0316] Step c, initializing N first construction power consumption prediction models with exactly the same structure to obtain N second construction power consumption prediction models;

[0317] Step d: for each second construction power consumption prediction model, select a set of parameter configuration schemes, configure the initial parameters in the second construction power consumption prediction model, and obtain N initial construction power consumption prediction models;

[0318] Step e, selecting N groups of data samples from the sample data set, and inputting the environmental monitoring data and safety monitoring data of a time period in each group of data samples into N initial construction power consumption prediction models to predict the safety monitoring data of the next time period, thereby obtaining N groups of safety prediction data;

[0319] Step f, calculating the fitness value of each set of parameter configuration schemes according to the N sets of safety prediction data and the safety monitoring data of the next time period in the N sets of data samples;

[0320] Step g, obtaining the parameter configuration scheme corresponding to the highest fitness value, and recording it as the target parameter configuration scheme;

[0321] Step h, updating each group of offset configuration schemes according to the first learning factor, the second learning factor, the third learning factor, the target parameter configuration scheme, and each group of parameter configuration schemes to obtain N groups of new offset configuration schemes;

[0322] Step i: according to each group of new offset configuration schemes, update each group of parameter configuration schemes to obtain N groups of updated parameter configuration schemes;

[0323] Step j: Repeat steps d to i until the maximum number of iterations is reached to obtain the optimal parameter configuration solution.

[0324] Optionally, the processing module 1102 is also used to: if the average sound intensity data in the safety prediction data is higher than the preset sound intensity threshold data or the average spark duration data is higher than the preset spark duration threshold data, it is determined that an electrical fire accident will occur in the construction power system operation area in the next time period.

[0325] Optionally, the processing module 1102 is also used to: if it is determined that an electrical fire accident will occur in the construction power system operation area in the next time period, automatically adjust the power supply voltage of the construction power system, and push safety warning information to the terminal device of the management personnel, wherein adjusting the power supply voltage of the construction power system includes controlling the power supply voltage of the construction power system to be reduced to a preset safe power supply voltage, or controlling the power of the construction power system to be turned off, and the safety warning information is used to indicate that an electrical fire accident will occur in the construction power system operation area.

[0326] Optionally, the processing module 1102 is also used to: for each set of parameter configuration schemes, calculate the mean square error value, the root mean square error value and the mean absolute error value based on the safety prediction data and the safety monitoring data of the next time period in the data sample; calculate the fitness value of the parameter configuration scheme based on the mean square error value, the root mean square error value and the mean absolute error value.

[0327] Optionally, the processing module 1102 is also used to: for each group of offset configuration schemes, calculate a first offset value according to a first learning factor and the offset configuration scheme; calculate a second offset value according to a second learning factor, a preset first random number, and a parameter configuration scheme; calculate a third offset value according to a third learning factor, a preset second random number, and a target parameter configuration scheme; and obtain a new offset configuration scheme according to the first offset value, the second offset value, and the third offset value.

[0328] Optionally, the first learning factor is determined based on a preset maximum first learning factor, a preset minimum first learning factor, the current number of iterations, and the maximum number of iterations; the second learning factor is determined based on a preset maximum second learning factor, a preset minimum second learning factor, the current number of iterations, the maximum number of iterations, and a preset cosine function; and the third learning factor is determined based on a preset maximum third learning factor, a preset minimum third learning factor, the current number of iterations, the maximum number of iterations, and a preset sine function.

[0329] The construction power system operation risk processing device 1100 provided in this embodiment can be used to execute the construction power system operation risk processing method in any of the aforementioned method embodiments. Its implementation principle and technical effects are similar and will not be repeated here.

[0330] Fig.12 This is a schematic diagram of the structure of a computer device provided in this application. Fig.12 As shown, the computer device 1200 may specifically include a receiver 1201, a processor 1202, and a memory 1203. The receiver 1201 is used to realize data transmission between the computer device 1200 and the collection layer device 1301, and the memory 1203 stores computer execution instructions; the processor 1202 executes the computer execution instructions stored in the memory 1203 to realize the construction power system operation risk processing method in the above embodiment.

[0331] Fig.13 A construction power system operation risk management system is provided for this application. Fig.13 As shown, the construction power system operation risk processing system 1300 includes Fig.12The computer device 1200 and the collection layer device 1301 connected to the computer device 1200 are shown. The collection layer device 1301 is used to monitor the environment in the operation area of ​​the construction power system and the safety of the electrical equipment in real time and upload the monitored data to the computer device 1200. The collection layer device 1301 includes: an acetylene concentration sensor 13011, an aluminum oxide powder concentration sensor 13012, a temperature and humidity sensor 13013, a light intensity sensor 13014, a sound intensity sensor 13015, and a spark detection sensor 13016. The acetylene concentration sensor 13011 is used to monitor acetylene concentration data; the aluminum oxide powder concentration sensor 13012 is used to monitor aluminum oxide powder concentration data; the temperature and humidity sensor 13013 is used to monitor temperature data and humidity data; the light intensity sensor 13014 is used to monitor light intensity data; the sound intensity sensor 13015 is used to monitor sound intensity data of electrical equipment; and the spark detection sensor 13016 is used to monitor spark duration data of electrical equipment.

[0332] This embodiment provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by the processor 1202, they are used to implement the construction power system operation risk management method of any one of the above embodiments.

[0333] This embodiment also provides a computer program product, including a computer program, which, when executed by the processor 1202, implements the construction power system operation risk management method provided in any of the above embodiments.

[0334] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0335] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for handling operation risks of a construction power system, characterized in that: A computer device used in a construction power system operation risk processing system, the method comprising: Based on the data uploaded by the collection layer equipment, the environmental monitoring data of the construction power system operation area in a preset time period and the safety monitoring data of the electrical equipment are obtained, wherein the preset time period is a period of time including the current moment and before the current moment, the environmental monitoring data includes the average acetylene concentration monitoring data, the average aluminum oxide powder concentration monitoring data, the average temperature monitoring data, the average humidity monitoring data and the average light intensity monitoring data, and the safety monitoring data includes the average sound intensity monitoring data and the average spark duration monitoring data of the electrical equipment; Inputting the environmental monitoring data and the safety monitoring data into a construction power consumption prediction model for prediction processing to obtain safety prediction data for the next time period; Determining whether an electrical fire accident will occur in the next time period based on the safety prediction data; The process of obtaining the construction power consumption prediction model includes: Based on historical environmental monitoring data and historical safety monitoring data, a preset algorithm is used to determine the optimal parameter configuration scheme of the initial construction power consumption prediction model, wherein the initial construction power consumption prediction model is a neural network model; According to the optimal parameter configuration scheme, the target parameters in the initial construction power consumption prediction model are configured to obtain the construction power consumption prediction model; The method of determining the optimal parameter configuration scheme of the initial construction power consumption prediction model using a preset algorithm based on historical environmental monitoring data and historical safety monitoring data includes: Step a, obtaining a sample data set based on the historical environmental monitoring data and the historical safety monitoring data, wherein the sample data set includes a plurality of data samples, each of which includes environmental monitoring data and safety monitoring data for a time period and safety monitoring data for a time period next to the time period; Step b, initialize N groups of parameter configuration schemes, N groups of offset configuration schemes, a first learning factor, a second learning factor, a third learning factor and a preset maximum number of iterations, wherein N is a positive integer, and each group of parameter configuration schemes includes a weight parameter and a bias parameter; the first learning factor is used to indicate the degree of influence of each group of offset configuration schemes to be updated on its own update, the second learning factor is used to indicate the degree of influence of the optimal parameter configuration scheme on each group of parameter configuration schemes when they are updated, and the third learning factor is used to indicate the degree of influence of the local optimal parameter configuration scheme on each group of parameter configuration schemes when they are updated. The first learning factor is determined according to the preset maximum first learning factor, the preset minimum first learning factor, the current number of iterations and the maximum number of iterations, and the calculation formula of the first learning factor is: in, Indicates the first learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset maximum first learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset minimum first learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, k represents the current number of iterations, k max Indicates the preset maximum number of iterations; The second learning factor is determined according to a preset maximum second learning factor, a preset minimum second learning factor, the current number of iterations, the maximum number of iterations, and a preset cosine function. The calculation formula of the second learning factor is: in, Indicates the second learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset maximum second learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset minimum second learning factor corresponding to the nth group of offset configuration schemes at the kth iteration; The third learning factor is determined according to a preset maximum third learning factor, a preset minimum third learning factor, the current number of iterations, the maximum number of iterations, and a preset sine function. The calculation formula of the third learning factor is: in, Indicates the third learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset maximum third learning factor corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset minimum third learning factor corresponding to the nth group of offset configuration schemes at the kth iteration; Step c, initializing N first construction power consumption prediction models with exactly the same structure to obtain N second construction power consumption prediction models; Step d: for each second construction power consumption prediction model, select a set of parameter configuration schemes, configure the initial parameters in the second construction power consumption prediction model, and obtain N initial construction power consumption prediction models; Step e, selecting N groups of data samples from the sample data set, and inputting the environmental monitoring data and safety monitoring data of a time period in each group of data samples into N initial construction power consumption prediction models to predict the safety monitoring data of the next time period, thereby obtaining N groups of safety prediction data; Step f, calculating the fitness value of each group of parameter configuration schemes according to the N groups of safety prediction data and the safety monitoring data of the next time period in the N groups of data samples; Step g, obtaining the parameter configuration scheme corresponding to the highest fitness value, and recording it as the target parameter configuration scheme; Step h: updating each group of offset configuration schemes according to the first learning factor, the second learning factor, the third learning factor, the target parameter configuration scheme, and each group of parameter configuration schemes to obtain N groups of new offset configuration schemes; Step i: according to each group of new offset configuration schemes, update each group of parameter configuration schemes to obtain N groups of updated parameter configuration schemes; Step j, repeating step d to step i until the maximum number of iterations is reached, and obtaining the optimal parameter configuration solution; The updating of each group of offset configuration schemes according to the first learning factor, the second learning factor, the third learning factor, the target parameter configuration scheme, and each group of parameter configuration schemes to obtain N groups of new offset configuration schemes includes: For each group of offset configuration schemes, a first offset value is calculated according to the first learning factor and the offset configuration scheme, and the calculation formula of the first offset value is: in, Indicates the first offset value corresponding to the nth group of offset configuration schemes at the kth iteration, R py (k,n) represents the nth group of offset configuration schemes at the kth iteration; A second offset value is calculated according to the second learning factor, the preset first random number, and the parameter configuration scheme. The calculation formula of the second offset value is: in, Indicates the second offset value corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset first random number at the kth iteration, It represents the global optimal parameter configuration scheme of the nth group of parameter configuration schemes at the kth iteration, P py (k,n) represents the nth group of parameter configuration schemes at the kth iteration; A third offset value is calculated according to the third learning factor, the preset second random number, and the target parameter configuration scheme. The calculation formula of the third offset value is: in, Indicates the third offset value corresponding to the nth group of offset configuration schemes at the kth iteration, Indicates the preset second random number at the kth iteration, It represents the local optimal parameter configuration scheme of the nth group of parameter configuration schemes at the kth iteration, P py (k,n) represents the nth group of parameter configuration schemes at the kth iteration; A new offset configuration scheme is obtained according to the first offset value, the second offset value and the third offset value. The calculation formula of the new offset configuration scheme is: Among them, R py (k+1,n) indicates the nth group of offset configuration schemes at the (k+1)th iteration, ∈ 1 Represents the preset first offset weight parameter, ∈ 2 Represents the preset second offset weight parameter, ∈ 3 Indicates the preset third offset weight parameter.

2. The method according to claim 1, characterized in that Determining whether an electrical fire accident will occur in the next time period according to the safety prediction data includes: If the average sound intensity data in the safety prediction data is higher than the preset sound intensity threshold data or the average spark duration data is higher than the preset spark duration threshold data, it is determined that an electrical fire accident will occur in the operation area of ​​the construction power system in the next time period.

3. The method according to claim 2, characterized in that The method further comprises: If it is determined that an electrical fire accident will occur in the operation area of ​​the construction power system in the next time period, the power supply voltage of the construction power system is automatically adjusted, and a safety warning message is pushed to the terminal device of the manager, wherein the adjustment of the power supply voltage of the construction power system includes controlling the power supply voltage of the construction power system to be reduced to a preset safety power supply voltage, or controlling the power supply of the construction power system to be turned off, and the safety warning message is used to indicate that an electrical fire accident will occur in the operation area of ​​the construction power system.

4. The method according to claim 1, characterized in that The calculating the fitness value of each group of parameter configuration schemes according to the N groups of safety prediction data and the safety monitoring data of the next time period in the N groups of data samples includes: For each set of parameter configuration schemes, the mean square error value, root mean square error value and mean absolute error value are calculated based on the safety prediction data and the safety monitoring data of the next time period in the data sample; The fitness value of the parameter configuration scheme is calculated according to the mean square error value, the root mean square error value and the mean absolute error value.

5. A computer device, characterized in that: include: A receiver, a processor and a memory connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the construction power system operation risk management method as described in any one of claims 1 to 4.

6. A construction power system operation risk management system, characterized in that: include: The computer device of claim 5 and a collection layer device connected to the computer device; The collection layer equipment is used to monitor the environment in the operation area of ​​the construction power system and the safety of the electrical equipment in real time and upload the monitored data to the computer equipment; The collection layer equipment includes: acetylene concentration sensor, aluminum oxide powder concentration sensor, temperature and humidity sensor, light intensity sensor, sound intensity sensor, spark detection sensor; The acetylene concentration sensor is used to monitor acetylene concentration data; Alumina powder concentration sensor is used to monitor the concentration data of alumina powder; Temperature and humidity sensors are used to monitor temperature and humidity data; The light intensity sensor is used to monitor light intensity data; The sound intensity sensor is used to monitor the sound intensity data of electrical equipment; Spark detection sensors are used to monitor spark duration data in electrical equipment.

Images