Emergency resource modular allocation method with rapid and accurate early warning response

Abstract

The invention belongs to the technical field of emergency resource allocation, and particularly relates to an emergency resource modular allocation method capable of achieving rapid and accurate early warning response. The method comprises the following steps: firstly, collecting emergency event related information by using a multi-source data acquisition channel, and carrying out data cleaning preprocessing; secondly, evaluating a risk level by using an enhanced random forest algorithm, and triggering an early warning mechanism; then, based on historical data and current information, using a multilayer perceptron (MLP) to predict emergency resource types and quantity, and dividing the emergency resources into different modules for deployment; and finally, planning an optimal resource transportation path by using a genetic algorithm. The method can effectively improve the data quality and prediction accuracy, accurately judge the risk level, reasonably deploy resources, optimize the transportation path, improve the emergency response efficiency, and guarantee the safe and stable operation of a power grid.

Description

Technical Field

[0001] The present invention belongs to the technical field of emergency resource allocation, and in particular relates to a modular emergency resource allocation method for rapid and accurate early warning response. Background Art

[0002] Nowadays, the safe and stable operation of the power grid is crucial to the normal operation of society. During the operation of the power grid, it may face various emergency events, such as natural disasters, equipment failures, etc. These events may have a serious impact on the power grid and even cause large-scale power outages. Therefore, rapid and accurate early warning response and reasonable allocation of emergency resources are of great significance to ensure the safe and stable operation of the power grid.

[0003] However, existing technologies have some shortcomings in the allocation of emergency resources for power grids. In terms of data collection, the data source is relatively single, and it is difficult to fully obtain information related to emergency events, such as event type, location, scope of impact, casualties, etc. This leads to an incomplete understanding of the event, which affects subsequent risk assessment and resource allocation. In terms of risk assessment, it is impossible to quickly assess the risk level of emergency events, making it difficult to trigger an effective early warning mechanism. In terms of resource forecasting, the lack of an accurate forecasting model based on historical data and current event information makes it impossible to accurately predict the type and quantity of emergency resources required, affecting the rational allocation of resources. In terms of transportation route planning, traditional route planning methods may not take into account multiple factors such as transportation time, cost, and reliability, making it difficult to plan the optimal resource transportation route, which affects the rapid and accurate arrival of materials at the site.

[0004] In summary, the existing technology has many problems in the allocation of emergency resources for power grids and cannot meet the needs of safe and stable operation of power grids. Therefore, an innovative modular allocation method of emergency resources with fast and accurate early warning response is needed to improve the ability of power grids to respond to sudden emergencies. Summary of the invention

[0005] In view of the above-mentioned technical problems in emergency resource allocation, the present invention proposes a modular emergency resource allocation method which is simple in method, easy to operate and can effectively solve the problem of rapid and accurate early warning response.

[0006] In order to achieve the above object, the technical solution adopted by the present invention comprises the following steps:

[0007] S1. First, use sensors, IoT devices, and social media multi-source data collection channels to collect information related to emergency events, including event type, location, impact range, and casualties, and clean and pre-process the collected data to ensure data quality and availability.

[0008] S2. Secondly, the enhanced random forest algorithm is used to analyze the pre-processed data to assess the risk level of the emergency event. According to the risk assessment results, the corresponding early warning mechanism is triggered and early warning information is sent to relevant personnel through SMS and APP push;

[0009] S3, then based on the historical emergency event data and the information and risk assessment results of the current event, the multi-layer perceptron MLP is used to predict the type and quantity of emergency resources required;

[0010] S4. Finally, the genetic algorithm is used to plan the optimal resource transportation route to ensure that the materials arrive at the site quickly and accurately.

[0011] Preferably, in step S1, the collected data is cleaned using the modified Laida criterion, which is specifically implemented as follows:

[0012] S11. First, the collected data set is {X 1 ,X 2 ,...X n}, calculate the mean and standard deviation of the data;

[0013] S12, then determine the interval of abnormal judgment. The traditional Laida criterion is that if X satisfies in is the mean, σ is the standard deviation, and the modified Laida criterion adjusts the multiple according to the actual situation, then the interval of abnormal judgment becomes

[0014] S13. Finally, for each data point X in the data set i , determine whether it is within the outlier judgment interval. If it is not within the interval, it is considered an outlier, and the outlier is replaced by the mean value.

[0015] Preferably, the implementation process of step S2 using the enhanced random forest algorithm to evaluate the risk level of the emergency event is as follows:

[0016] S21. First, define the enhanced random forest model. The random forest is Where B is the number of trees, f b is the prediction result of the b-th tree, and the adaptive boosting algorithm AdaBoost is where h m (x) is the mth weak classifier, α m is the weight of the classifier;

[0017] S22, then combine the random forest and adaptive boosting algorithm AdaBoost for training, initialize the weights of all training samples to equal values, train a decision tree as a weak classifier, use the current sample weights, and calculate the classification error rate of the weak classifier where ω i is the weight of the i-th sample, I is the indicator function, if the classification error is 1, the classification is correct is 0; then the weight α of the weak classifier is calculated according to the classification error rate m , Then, the sample weights are updated so that the weights of samples with incorrect classification are increased; then, the sample weights are normalized so that their sum is 1; finally, the weak classifier is trained repeatedly to the normalized weight step until the preset number of iterations is reached or the error rate reaches a threshold;

[0018] S23. Use the trained enhanced random forest model to conduct risk assessment on new data to obtain the risk level. Combine the outputs of all weak classifiers to calculate the final risk score. Then, set the threshold according to the risk score and divide the risk level: risk score > 0.7: high risk, 0.4 < risk score ≤ 0.7: medium risk, risk score ≤ 0.4: low risk.

[0019] Preferably, the specific implementation operation of step S3 using a multi-layer perceptron MLP to predict the required emergency resource types and quantities is:

[0020] S31. First, determine the model input type, encode the event type, quantify the risk level, and express the impact range and casualties with actual values;

[0021] S32, secondly, dividing the data set into a training set, a validation set, and a test set in a ratio of 7:2:1, using the training set to train the selected deep learning model, and adjusting the parameters of the model by minimizing the mean square error;

[0022] S33, then use the validation set to evaluate the trained model to check the accuracy and generalization ability of the model;

[0023] S34. Finally, when the model evaluation is passed, the characteristic data of the current event is input into the trained model, and the model will output the predicted type and quantity of emergency resources.

[0024] Preferably, after predicting the type and quantity of emergency resources in step S3, the emergency resources are divided into a medical resource module, a rescue equipment module, and a daily necessities module. According to the resource demand prediction results, the corresponding resources are quickly allocated from each module to complete the emergency allocation task.

[0025] Preferably, the steps for implementing the planning of the optimal resource transportation path using the genetic algorithm in step S4 are:

[0026] S41. First, all the transport paths that need to be taken are expressed as genetic codes;

[0027] S42. Secondly, a fitness function is constructed to evaluate the pros and cons of each transportation route. This function comprehensively considers multiple factors such as transportation time, transportation cost, and transportation reliability, namely Where T is the transportation time, C is the transportation cost, and R is the transportation reliability, ranging from 0 to 1. The shorter the transportation time, the lower the cost, and the higher the reliability, the higher the value of the fitness function, and the better the corresponding path;

[0028] S43, then randomly generate a certain number of initial transport path individuals to form an initial population, the encoding of the individuals is a randomly arranged node sequence; set the parameters of the genetic algorithm, the crossover probability is 0.75, and the mutation probability is 0.04;

[0029] S44, then performing population evolution operations according to the set algorithm parameters, including selection, crossover and mutation;

[0030] S45, then use the newly generated offspring individuals to replace some individuals in the current population to form a new population; repeat the population evolution operation, after 80 rounds of iterations, the individual fitness of the population will gradually increase;

[0031] S46. Finally, after the iteration, the transportation path represented by the individual with the highest fitness in the population is the planned optimal transportation path.

[0032] Compared with the prior art, the advantages and positive effects of the present invention are that the modified Laida criterion is used for data cleaning, which can effectively identify and replace outliers in the data, improve data quality, and thus ensure the accuracy of subsequent predictions. The enhanced random forest algorithm is used to analyze the preprocessed data, comprehensively consider multiple factors, accurately judge the risk level of emergency events, and provide a reliable basis for early warning and resource allocation. Based on historical emergency event data and the information and risk assessment results of current events, the multi-layer perceptron MLP is used to predict the type and quantity of emergency resources required, which can deeply explore the potential patterns and relationships in the data, make more accurate predictions, and improve the rationality of resource allocation. The genetic algorithm is used to plan the optimal resource transportation route, comprehensively consider multiple factors such as transportation time, transportation cost, and transportation reliability, and can find the optimal solution among many possible routes, ensuring that materials can arrive at the scene quickly and accurately, and improving the efficiency of emergency response. DETAILED DESCRIPTION

[0033] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described below in conjunction with embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0034] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways than those described herein. Therefore, the present invention is not limited to the specific embodiments of the following disclosure.

[0035] Embodiment, in modern society, the stable operation of the power grid is crucial to the normal production of all walks of life and people's daily life. However, the power grid may be affected by various natural disasters, such as lightning strikes, wind disasters, floods, etc., resulting in power grid equipment failures, power outages and other problems. These disasters will not only cause economic losses to power companies, but may also affect the normal operation of society and bring inconvenience to people's lives. In order to solve these problems, this patent proposes a modular deployment method for emergency resources with fast and accurate early warning response.

[0036] First, we use sensors, IoT devices, social media and other multi-source data collection channels to collect information related to emergency events, including event type, location, impact range, and casualties. In order to improve the quality of collected data and make accurate predictions later, the present invention uses the modified Laida criterion for data cleaning. First, the collected data set is {X 1 ,X 2 ,...X n}, calculate the mean and standard deviation of the data; then determine the interval of abnormal judgment. The traditional Laida criterion is that if X satisfies in is the mean, σ is the standard deviation, and the modified Laida criterion adjusts the multiple according to the actual situation, then the interval of abnormal judgment becomes Finally, for each data point X in the data set i , determine whether it is within the outlier judgment interval. If it is not within the interval, it is considered an outlier, and the outlier is replaced by the mean. The data quality after cleaning has been significantly improved, and the outliers in the data have been effectively identified and replaced, making the data more accurate and reliable.

[0037] Next, considering that we should understand the risk level of emergency events, the present invention uses the enhanced random forest algorithm to analyze the pre-processed data and comprehensively consider multiple factors to accurately determine the risk level of emergency events. First, the enhanced random forest model is defined. The random forest is Where B is the number of trees, f b is the prediction result of the b-th tree, and the adaptive boosting algorithm AdaBoost is where h m (x) is the mth weak classifier, α mis the weight of the classifier; then the random forest and adaptive boosting algorithm AdaBoost are combined for training, the weights of all training samples are initialized to equal values, a decision tree is trained as a weak classifier, and the classification error rate of the weak classifier is calculated using the current sample weight. where ω i is the weight of the i-th sample, I is the indicator function, if the classification error is 1, the classification is correct is 0; then the weight α of the weak classifier is calculated according to the classification error rate m , Then update the sample weights so that the weights of samples with classification errors increase; then normalize the sample weights so that their sum is 1; finally repeat the training of the weak classifier to the normalized weight step until the preset number of iterations is reached or the error rate reaches the threshold; use the trained enhanced random forest model to conduct risk assessment on the new data to obtain the risk level, first combine the outputs of all weak classifiers, and calculate the final risk score Then, according to the risk score, the threshold is set to divide the risk level: risk score>0.7: high risk, 0.4<risk score≤0.7: medium risk, risk score≤0.4: low risk. According to the risk assessment results, the corresponding early warning mechanism is triggered, and early warning information is sent to relevant personnel through SMS, APP push, etc., to ensure that they can understand the situation in time and take corresponding measures.

[0038] Then, based on historical emergency event data and current event information and risk assessment results, a multi-layer perceptron (MLP) is used to predict the type and quantity of emergency resources required. MLP can deeply explore the potential patterns and relationships in the data to make more accurate predictions.

[0039] Finally, the genetic algorithm is used to plan the optimal resource transportation route. The algorithm can find the optimal solution among many possible routes, ensuring that materials can arrive at the scene quickly and accurately, and improving the efficiency of emergency response. First, all the transportation routes that need to be passed are represented as genetic codes; secondly, a fitness function is constructed to evaluate the pros and cons of each transportation route. This function comprehensively considers multiple factors such as transportation time, transportation cost, and transportation reliability, namely Where T is the transportation time, C is the transportation cost, and R is the reliability of transportation, with a value range of 0 to 1; then iteration is performed to improve the individual fitness of the population; finally, after the iteration, the transportation path represented by the individual with the highest fitness in the population is selected as the planned optimal transportation path.

[0040] The above description is only a preferred embodiment of the present invention and does not limit the present invention in other forms. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A modular deployment method for emergency resources with rapid and accurate early warning response, characterized in that: The following steps are involved: S1. First, use sensors, IoT devices, and social media multi-source data collection channels to collect information related to emergency events, including event type, location, impact range, and casualties, and clean and pre-process the collected data to ensure data quality and availability; S2. Secondly, the enhanced random forest algorithm is used to analyze the pre-processed data to assess the risk level of the emergency event. According to the risk assessment results, the corresponding early warning mechanism is triggered and early warning information is sent to relevant personnel through SMS and APP push; S3, then based on the historical emergency event data and the information and risk assessment results of the current event, the multi-layer perceptron MLP is used to predict the type and quantity of emergency resources required; S4. Finally, the genetic algorithm is used to plan the optimal resource transportation route to ensure that the materials arrive at the site quickly and accurately.

2. According to the method of claim 1, the modular deployment method of emergency resources for rapid and accurate early warning response is characterized in that: In step S1, the collected data is cleaned using the modified Laida criterion, which is specifically implemented as follows: S11. First, the collected data set is {X1,X2,...X n }, calculate the mean and standard deviation of the data; S12, then determine the interval of abnormal judgment. The traditional Laida criterion is that if X satisfies in is the mean, σ is the standard deviation, and the modified Laida criterion adjusts the multiple according to the actual situation, then the interval of abnormal judgment becomes S13. Finally, for each data point X in the data set i , determine whether it is within the outlier judgment interval. If it is not within the interval, it is considered an outlier, and the outlier is replaced by the mean value.

3. According to claim 1, a modular deployment method for emergency resources with rapid and accurate early warning response is characterized in that: The implementation process of step S2 using the enhanced random forest algorithm to evaluate the risk level of emergency events is as follows: S21. First, define the enhanced random forest model. The random forest is Where B is the number of trees, f b is the prediction result of the b-th tree, and the adaptive boosting algorithm AdaBoost is where h m (x) is the mth weak classifier, α m is the weight of the classifier; S22, then combine the random forest and adaptive boosting algorithm AdaBoost for training, initialize the weights of all training samples to equal values, train a decision tree as a weak classifier, use the current sample weights, and calculate the classification error rate of the weak classifier where ω i is the weight of the i-th sample, I is the indicator function, if the classification error is 1, the classification is correct is 0; then the weight α of the weak classifier is calculated according to the classification error rate m , Then update the sample weights so that the weights of samples with classification errors increase; then normalize the sample weights so that their sum is 1; finally, repeat the training of the weak classifier to the normalized weight step until the preset number of iterations is reached or the error rate reaches a threshold; S23. Use the trained enhanced random forest model to conduct risk assessment on new data to obtain the risk level. Combine the outputs of all weak classifiers to calculate the final risk score. Then, set the threshold according to the risk score and divide the risk level: risk score > 0.7: high risk, 0.4 < risk score ≤ 0.7: medium risk, risk score ≤ 0.4: low risk.

4. According to claim 1, a modular deployment method for emergency resources with rapid and accurate early warning response is characterized in that: The specific implementation operation of step S3 using the multi-layer perceptron MLP to predict the required emergency resource types and quantities is: S31. First, determine the model input type, encode the event type, quantify the risk level, and express the impact range and casualties with actual values; S32, secondly, dividing the data set into a training set, a validation set, and a test set in a ratio of 7:2:1, using the training set to train the selected deep learning model, and adjusting the parameters of the model by minimizing the mean square error; S33, then use the validation set to evaluate the trained model to check the accuracy and generalization ability of the model; S34. Finally, when the model evaluation is passed, the characteristic data of the current event is input into the trained model, and the model will output the predicted type and quantity of emergency resources.

5. According to the method of modular deployment of emergency resources for rapid and accurate early warning response according to claim 1, it is characterized in that: After predicting the type and quantity of emergency resources in step S3, the emergency resources are divided into medical resource module, rescue equipment module, and daily necessities module. According to the resource demand prediction results, the corresponding resources are quickly allocated from each module to complete the emergency allocation task.

6. According to claim 1, a modular deployment method for emergency resources with rapid and accurate early warning response is characterized in that: The steps for implementing the planning of the optimal resource transportation path using the genetic algorithm in step S4 are as follows: S41. First, all the transport paths that need to be taken are expressed as genetic codes; S42. Secondly, a fitness function is constructed to evaluate the pros and cons of each transportation route. This function comprehensively considers multiple factors such as transportation time, transportation cost, and transportation reliability, namely Where T is the transportation time, C is the transportation cost, and R is the transportation reliability, ranging from 0 to 1. The shorter the transportation time, the lower the cost, and the higher the reliability, the higher the value of the fitness function, and the better the corresponding path; S43, then randomly generate a certain number of initial transport path individuals to form an initial population, the encoding of the individuals is a randomly arranged node sequence; set the parameters of the genetic algorithm, the crossover probability is 0.75, and the mutation probability is 0.04; S44, then performing population evolution operations according to the set algorithm parameters, including selection, crossover and mutation; S45, then use the newly generated offspring individuals to replace some individuals in the current population to form a new population; repeat the population evolution operation, after 80 rounds of iterations, the individual fitness of the population will gradually increase; S46. Finally, after the iteration, the transportation path represented by the individual with the highest fitness in the population is the planned optimal transportation path.

Images