Emergency rescue method and system for coping with emergency fault event

Abstract

The invention discloses an emergency rescue method and system for dealing with an emergency fault event. The method comprises the following steps: delineating a monitoring range and independent operation individuals in the monitoring range; historical operation data of the independent operation individuals are collected, the historical operation data comprise monitoring parameters influencing normal work of the independent operation individuals and exception categories corresponding to the monitoring parameters, and the exception categories comprise exception and normality; training an exception judgment model for identifying exception categories based on the historical operation data; acquiring g real-time monitoring parameters of the individual running independently in unit time, inputting the g real-time monitoring parameters into the trained anomaly judgment model, outputting g'anomaly categories, and marking the real-time monitoring parameters with the anomaly categories as abnormal parameters; and calculating a danger degree based on the abnormal parameters, and generating a rescue scheme based on the danger degree. By balancing the risk between automatic rescue and manual rescue, the safety and timeliness of emergency rescue are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of data processing, and in particular relates to an emergency rescue method and system for coping with emergency failure events. Background Art

[0002] In the fields of public facilities, industrial production, etc., emergency rescue for emergency failure events is particularly important. It aims to quickly respond to equipment or system failures and reduce the impact of failures on production, business and public safety;

[0003] For example, the authorization announcement number is: CN113408823B discloses an emergency rescue method for urban emergencies. At present, the rescue methods for emergency failure events are mainly divided into automatic equipment rescue or manual rescue. Automatic rescue is a backup rescue configuration reserved for the equipment and automatic rescue executed by an automatic rescue program. Automatic rescue can perform rescue as soon as an emergency failure event occurs. However, whether it is an equipment emergency failure event involving public facilities or industrial production fields, automatic rescue is a reserved solution. Its risk assessment of emergency failure events and the flexibility of rescue are not as good as manual rescue, but the cost and timeliness of manual rescue are weaker than automatic rescue.

[0004] To this end, the present invention proposes an emergency rescue method and system for coping with emergency failure events. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention proposes an emergency rescue method and system for dealing with emergency failure events, which can balance the risks between automatic rescue and manual rescue, and improve the safety and timeliness of emergency rescue.

[0006] To achieve the above object, an emergency rescue method for emergency failure events is proposed according to the present invention, the method comprising:

[0007] Define the monitoring scope and the independently operating individuals within the monitoring scope;

[0008] Collect historical operation data of the independently operated individuals, the historical operation data including monitoring parameters that affect the normal operation of the independently operated individuals and abnormal categories corresponding to the monitoring parameters, the abnormal categories including abnormal and normal;

[0009] Anomaly determination models that are trained to identify anomaly categories based on historical operation data;

[0010] Obtain g real-time monitoring parameters of the independent operating individuals in unit time, input the g real-time monitoring parameters into the trained abnormality judgment model, output g′ abnormality categories, and mark the real-time monitoring parameters with abnormal categories as abnormal parameters;

[0011] The degree of danger is calculated based on abnormal parameters, and a rescue plan is generated based on the degree of danger.

[0012] Preferably, the method for calculating the degree of danger based on abnormal parameters includes:

[0013] Remove the real-time monitoring parameters with normal abnormal categories from the g′ abnormal categories to obtain g″ abnormal parameters, and mark the total number of g″ abnormal parameters as the abnormal amount;

[0014] Calculate the abnormality degree of g″ abnormal parameters one by one, and obtain g″′ abnormality degrees, g″′=g″≤g′=g;

[0015] Comprehensively analyze and calculate the g″′ abnormality levels to obtain the comprehensive abnormality level;

[0016] The degree of danger is calculated based on the comprehensive abnormality degree and abnormality amount.

[0017] Preferably, the rescue plan includes automatic rescue and manual rescue, and the method for generating the rescue plan based on the degree of danger includes:

[0018] Compare the danger level with a preset danger level threshold;

[0019] If the danger level is greater than the preset danger level threshold, the rescue plan is determined to be a manual rescue plan;

[0020] If the danger level is less than or equal to the preset danger level threshold, the rescue plan is determined to be an automatic rescue plan, the automatic rescue plan is immediately executed and the progress of the automatic rescue plan is monitored, and based on the progress of the automatic rescue plan, it is determined whether to switch to a manual rescue plan.

[0021] Preferably, the method for monitoring the progress of the automatic rescue solution includes:

[0022] When the automatic rescue plan is executed, the automatic rescue time, the abnormal quantity reduction coefficient and the comprehensive abnormal degree reduction coefficient are monitored in real time;

[0023] The rescue progress value is generated in real time based on the automatic rescue time, the abnormality reduction coefficient and the comprehensive abnormality degree reduction coefficient.

[0024] Preferably, the method for determining whether to switch to a manual rescue solution based on the progress of the automatic rescue solution includes:

[0025] Compare the rescue progress value with a preset rescue progress value threshold to determine whether to generate a stop instruction;

[0026] If the rescue progress value is greater than or equal to the preset rescue progress value threshold, a stop command is generated and manual rescue is immediately performed;

[0027] If the rescue progress value is less than the preset rescue progress value threshold, no stop instruction is generated, and the rescue progress value is continuously updated.

[0028] Preferably, the time taken for the automatic rescue is the time taken for the automatic rescue starting from the time when the automatic rescue scheme is started;

[0029] The abnormal amount reduction coefficient is the rate of decrease of the total amount of abnormal parameters of the independently operating individual during the automatic rescue process;

[0030] The comprehensive abnormality degree reduction coefficient is the rate of reduction of the comprehensive abnormality degree of the independently operating individual during the automatic rescue process.

[0031] Preferably, the training method of the abnormality determination model includes:

[0032] The label of the abnormal category corresponding to the monitoring parameter is normal is set to 0, and the label of the abnormal category corresponding to the monitoring parameter is abnormal is set to 1, and the monitoring parameters and the labels corresponding to the monitoring parameters are constructed as a data set of the machine learning model; the data set is divided into a training set, a validation set and a test set;

[0033] The training set is used as the input of the machine learning model, and the machine learning model uses the predicted label corresponding to the monitoring parameter as the output; the actual label corresponding to the monitoring parameter is used as the prediction target, and minimizing the loss function value of the machine learning model is used as the training target; when the loss function value of the machine learning model is less than or equal to the preset target loss value, the training is stopped, and the machine learning model obtained by training is used as the abnormality judgment model;

[0034] The machine learning model is a random forest, a deep learning model or a long short-term memory network.

[0035] Preferably, the automatic rescue is an automatic rescue procedure manually provided for an independently operating individual.

[0036] An emergency rescue system for dealing with emergency failure events, which is applied to the above-mentioned emergency rescue method for dealing with emergency failure events, comprises:

[0037] A monitoring unit is arranged on the independently operated individual and is used to monitor the real-time monitoring parameters of the independently operated individual;

[0038] A training module is used to collect historical operation data of independent operating individuals and train an abnormality determination model based on the historical operation data;

[0039] A processing unit, used for generating an abnormality category based on the real-time monitoring parameters and the trained abnormality determination model, and marking the real-time monitoring parameters with abnormal categories as abnormal parameters;

[0040] The rescue identification module is used to calculate the degree of danger based on abnormal parameters and generate a rescue plan based on the degree of danger.

[0041] Compared with the prior art, the beneficial effects of the present invention are: through the abnormal judgment model trained based on historical operation data, the abnormal parameters and the number of abnormal parameters of independently operating individuals are identified, and the degree of danger of the independently operating individuals is determined based on the abnormal parameters and the number of abnormal parameters, and whether to execute the automatic rescue plan is determined according to the degree of danger, and the rescue progress value of the automatic rescue is continuously judged. If the automatic rescue effect is not ideal, manual rescue is promptly switched to ensure the flexibility and controllability of the rescue process, and avoid potential risks caused by slow rescue progress. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A flowchart of an emergency rescue method for emergency failure events according to an embodiment of the present invention;

[0043] Figure 2 A flowchart of an emergency rescue method for dealing with an emergency failure event according to another embodiment of the present invention;

[0044] Figure 3 A module connection relationship diagram of an emergency rescue system for coping with emergency failure events according to the present invention;

[0045] Figure 4 A schematic diagram of n independently operating individuals of an emergency rescue method for coping with an emergency failure event according to the present invention. DETAILED DESCRIPTION

[0046] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] 1. Reference Figure 1 As shown, an emergency rescue method for dealing with an emergency failure event, the method includes:

[0048] Define the monitoring scope and the independently operating individuals within the monitoring scope; illustratively, for example, the unit of the monitoring scope is a factory workshop, that is, each equipment in the factory workshop is monitored, then the rescue target is each equipment in the factory workshop, and each equipment in the factory workshop is an independently operating individual; for another example, the unit of the monitoring scope is the elevator equipment in a residential area, and the rescue target is the emergency failure event of each elevator equipment in the residential area, and each elevator equipment is an independently operating individual.

[0049] The historical operation data of the independently operating individuals are collected. The historical operation data include the monitoring parameters that affect the normal operation of the independently operating individuals and the abnormal categories corresponding to the monitoring parameters. The abnormal categories include abnormal and normal. The monitoring parameters include the status data, electrical data and environmental data of the independently operating individuals.

[0050] In some embodiments, status data includes temperature, vibration, and noise during operation of the independently operating individual; abnormal changes in temperature, vibration, or noise can reflect a failure of the independently operating individual. For example, if the temperature, vibration, or noise increases compared to normal values, it means that a jam or increased friction occurs at a certain position of the independently operating individual; electrical data includes power, current, and voltage during operation of the independently operating individual. Optionally, the power, current, and voltage reflect the working status and load conditions of the independently operating individual. For example, as the load of the equipment gradually increases from conventional load, the power, current, and voltage of the equipment will also increase accordingly. If the power, current, and voltage approach zero, it means that the independent operating individual has a shutdown failure or a power failure; environmental data includes ambient temperature, ambient humidity, and load, etc. Such data are used to reflect the working environment of the independently operating individual.

[0051] An abnormality determination model for identifying abnormal categories is trained based on historical operation data. Specifically, the training method of the abnormality determination model includes:

[0052] Collect multiple groups of historical operation data, set the label of the abnormal category corresponding to the monitoring parameters in the historical operation data as normal to 0, set the label of the abnormal category corresponding to the monitoring parameters as abnormal to 1, and construct the monitoring parameters and the labels corresponding to the monitoring parameters as the data set of the machine learning model; the data set is divided into a training set, a validation set and a test set, such as dividing the data set into 70% training set, 15% validation set and 15% test set; use the training set as the input of the machine learning model, and the machine learning model uses the predicted labels corresponding to the monitoring parameters as the output; use the actual labels corresponding to the monitoring parameters as the prediction targets, and use minimizing the loss function value of the machine learning model as the training target; stop training when the loss function value of the machine learning model is less than or equal to the preset target loss value, and use the trained machine learning model as the abnormality judgment model;

[0053] The machine learning model loss function can be mean square error (MSE) or cross entropy (CE);

[0054] For example, the mean square error (MSE) is calculated by dividing the loss function value by The model is trained with minimization as the goal, so that the machine learning model can better fit the data, thereby improving performance and accuracy; i is the monitoring parameter group number; u is the number of monitoring parameter groups; y i is the label corresponding to the i-th group of monitoring parameters, is the predicted label based on the i-th set of monitoring parameters.

[0055] The first machine learning model can be any one of a random forest, a deep learning model, or a long short-term memory network.

[0056] Obtain g real-time monitoring parameters of independent operating individuals in unit time, and input the g real-time monitoring parameters into the trained abnormality judgment model, where g is an integer greater than or equal to 1, output g′ abnormal categories, and mark the real-time monitoring parameters with abnormal categories as abnormal parameters.

[0057] Specifically, the method for calculating the degree of danger based on abnormal parameters includes:

[0058] Collect g real-time monitoring parameters, input the g real-time monitoring parameters into the anomaly judgment model, and output g′ anomaly categories;

[0059] Mark the real-time monitoring parameters with abnormal categories as abnormal parameters, remove the real-time monitoring parameters with normal abnormal categories, obtain g″ abnormal parameters, and mark the total number of g″ abnormal parameters as abnormal quantity;

[0060] Calculate the abnormal degree of g″ abnormal parameters one by one to obtain g″′ abnormal degrees, g″′=g″≤g′=g, and calculate the g″′ abnormal degrees formulaically to obtain the comprehensive abnormal degree, and calculate the degree of danger based on the comprehensive abnormal degree and the abnormal amount.

[0061] Specifically, the abnormality degree is calculated as follows: Among them, YC p is the abnormal degree of the Pth abnormal parameter, x P The Pth abnormal parameter, x′ P is the normal value corresponding to the Pth abnormal parameter. Each normal value is evaluated by the equipment operation manual or professional technicians. For example, the independent operation unit is the factory workshop equipment, and the monitoring parameter is the factory workshop equipment temperature. The equipment manual of the factory workshop equipment stipulates that the normal operating temperature of the factory workshop equipment is 70°C. Scenario 1: The abnormality is small: Scenario 2: Greater abnormality: Note that in scenario one, it means that the factory workshop equipment is working in a near-normal state and immediate action may not be required; in scenario two, the temperature is abnormally large, which means that the factory workshop equipment is high and immediate action may be required to avoid equipment damage or safety hazards.

[0062] The calculation method of comprehensive abnormality degree is: Among them, ZY is the abnormal amount.

[0063] The calculation method of the degree of danger is: XQ = ω 1 ×ZY+ω 2 ×ZH, where ω 1and ω 2 The preset weights for the amount of anomaly and the overall degree of anomaly.

[0064] More specifically, the rescue plan includes automatic rescue and manual rescue. The method for generating a rescue plan based on the degree of danger includes: comparing the degree of danger with a preset danger level threshold. The preset danger level threshold is set by technical personnel in this field according to specific actual conditions and is not specifically limited here; if the degree of danger is greater than the preset danger level threshold, the rescue plan is determined to be a manual rescue plan; if the degree of danger is less than or equal to the preset danger level threshold, the rescue plan is determined to be an automatic rescue plan, the automatic rescue plan is immediately executed and the progress of the automatic rescue plan is monitored, and based on the progress of the automatic rescue plan, it is determined whether to switch to a manual rescue plan.

[0065] Automatic rescue is an automatic rescue program provided by humans for independent operating individuals. For example, when an emergency failure occurs, a series of tasks are automatically started and executed to respond to emergencies through automatic control, adjustment or modification of the operating status by computers or PLCs. The core goal of the program is to reduce equipment downtime, avoid further damage to the equipment, ensure safety, and do not rely on manual operation or intervention;

[0066] Exemplary, for example, for an emergency fault event in which the independent operating individual is an elevator, if the emergency fault event is an emergency stop caused by a power outage, the fault characteristics are that the independent operating individual stops running, and the monitoring parameters such as current and voltage tend to zero; the automatic rescue plan is: the automatic rescue program controls the backup power supply to supply power to the independent operating individual, starts the ventilation and emergency lighting equipment to ensure the comfort of the passengers, and controls the elevator to level the floors, moves to the nearest floor and opens the elevator door to ensure the safe departure of the passengers; if the emergency fault event is that a foreign object is stuck in the floor elevator door, the fault characteristics are that after the elevator reaches the designated floor, the door cannot be opened normally and the door is stuck; the automatic rescue plan The solution is that the automatic rescue program automatically detects and marks the floor where the fault occurs. When the equipment reaches this floor, it skips this floor and goes to the next or previous floor, and sends fault alarm information to the outside world (such as property or maintenance personnel) to remind them to deal with the door card foreign body problem in time; if the emergency fault event is stall, the fault characteristics are that the elevator cannot run smoothly, and abnormal monitoring parameters such as vibration and noise occur; the automatic rescue plan is that the automatic rescue program triggers the elevator's safety brake device, stops the equipment operation, and controls the start and stop of the safety brake device, so that the elevator uses gravity to move smoothly to the nearest floor, automatically opens the elevator door and guides passengers to leave safely;

[0067] For another example, in the event of an emergency failure in which the independently operating individual is a factory workshop equipment, if the emergency failure event is that the factory workshop equipment automatically shuts down due to overtemperature, the failure characteristics are: abnormal ambient temperature or temperature of the independently operating individual in the monitoring parameters; the automatic rescue plan is: the automatic rescue program starts the backup cooling system or cooling water circulation system of the factory workshop equipment.

[0068] It is worth mentioning that the automatic rescue program provided in this embodiment needs to be implemented in conjunction with a specific redundant configuration designed for independent operation, such as a backup power supply, a backup cooling system, etc. The specific configuration is not specifically limited here.

[0069] In some embodiments, see Figure 2 As shown, the method for monitoring the progress of the automatic rescue plan and determining whether to terminate the automatic rescue includes: when the automatic rescue plan is executed, real-time monitoring of the automatic rescue time, the abnormality reduction coefficient and the comprehensive abnormality degree reduction coefficient; based on the automatic rescue time, the abnormality reduction coefficient and the comprehensive abnormality degree reduction coefficient, a rescue progress value is generated in real time, and the rescue progress value is compared with a preset rescue progress value threshold. The preset rescue progress value threshold is determined by technical personnel in this field based on historical rescue data statistics or experimental data fitting to determine whether to generate a stop command; if the rescue progress value is greater than or equal to the preset rescue progress value threshold, a stop command is generated and manual rescue is immediately switched; if the rescue progress value is less than the preset rescue progress value threshold, no stop command is generated, and the rescue progress value is continuously updated.

[0070] The rescue progress value reflects the effectiveness of the current automatic rescue progress and its potential risks. It combines the rescue time, the abnormality reduction coefficient and the comprehensive abnormality degree reduction coefficient. By combining these factors, a quantitative risk assessment can be obtained. Specifically, the time taken for automatic rescue is the time taken for automatic rescue from the start of the automatic rescue plan, that is, the automatic rescue time = current time - rescue start time; the abnormality reduction coefficient is the rate of decrease of the total amount of abnormal parameters of independently operating individuals during the execution of automatic rescue; illustratively, during the execution of automatic rescue, the total amount of abnormal parameters gradually decreases, indicating that the automatic rescue plan is effective; the comprehensive abnormality degree reduction coefficient is the rate of decrease of the comprehensive abnormality degree of independently operating individuals during the automatic rescue process. This coefficient evaluates the changes in the comprehensive abnormality degree of independent individuals.

[0071] Specifically, the calculation expression of the abnormal quantity reduction coefficient is: Among them, ZY CS is the initial abnormal amount, i.e. the abnormal amount when the automatic rescue starts, ZY CS =ZY, ZY′ is the abnormal amount calculated in real time after the automatic rescue starts as described above, and ZY″ is the abnormal amount reduction coefficient.

[0072] The calculation expression of the comprehensive abnormality reduction coefficient is: Among them, ZH CS is the initial comprehensive abnormality level, that is, the comprehensive abnormality level calculated when the automatic rescue starts, ZH CS=ZH, ZH′ is the comprehensive abnormality degree calculated in real time after the automatic rescue starts as described above, and ZH″ is the comprehensive abnormality degree reduction coefficient.

[0073] The calculation expression of the rescue progress value is: α, β, γ are preset weight coefficients, which respectively control the influence of time consumption, abnormal quantity reduction coefficient, and abnormal degree reduction coefficient on the rescue progress value, and α+β+γ=1, T max The maximum allowed time for the automatic rescue plan to execute, in seconds (s), T auto The time consumed for automatic rescue is the time consumed for automatic rescue starting from the start of the automatic rescue scheme.

[0074] The purpose is to ensure that automatic rescue can be effectively executed and its potential risks can be discovered in time by real-time monitoring and evaluation of key parameters in the automatic rescue process. When the rescue progress value exceeds the preset rescue progress value threshold, the system can automatically stop automatic rescue and switch to manual rescue to avoid further risks or equipment damage; when the rescue progress value is lower than the preset rescue progress value threshold, automatic rescue continues to be performed, thereby improving the effectiveness and efficiency of rescue. Through this dynamic adjustment and monitoring mechanism, the most appropriate emergency response can be achieved in complex fault situations, thereby maximizing the protection of equipment recovery and system stability.

[0075] 2. Reference Figure 3 and Figure 4 As shown, an emergency rescue system for dealing with emergency failure events is applied to the above-mentioned emergency rescue method for dealing with emergency failure events. The system includes a monitoring unit, a training module, a processing unit, a rescue judgment module and a supervision module, wherein each module is connected by wire and / or wireless.

[0076] The monitoring unit is arranged on the independently operating individual and is used to monitor the real-time monitoring parameters of the independently operating individual. The monitoring unit of this embodiment includes various sensors for monitoring various real-time monitoring parameters, which depends on the type of monitoring parameters and is not specifically limited here.

[0077] The training module is used to collect the historical operation data of the independent operation individuals and train the abnormality determination model based on the historical operation data; the training method of the abnormality determination model includes:

[0078] Collect multiple groups of historical operation data, set the label of the abnormal category corresponding to the monitoring parameters in the historical operation data as normal to 0, set the label of the abnormal category corresponding to the monitoring parameters as abnormal to 1, and construct the monitoring parameters and the labels corresponding to the monitoring parameters as the data set of the machine learning model; the data set is divided into a training set, a validation set and a test set, such as dividing the data set into 70% training set, 15% validation set and 15% test set; use the training set as the input of the machine learning model, and the machine learning model uses the predicted labels corresponding to the monitoring parameters as the output; use the actual labels corresponding to the monitoring parameters as the prediction targets, and use minimizing the loss function value of the machine learning model as the training target; stop training when the loss function value of the machine learning model is less than or equal to the preset target loss value, and use the trained machine learning model as the abnormality judgment model;

[0079] The machine learning model loss function can be mean square error (MSE) or cross entropy (CE);

[0080] For example, the mean square error (MSE) is calculated by dividing the loss function value by The model is trained with minimization as the goal, so that the machine learning model can better fit the data, thereby improving performance and accuracy; i is the monitoring parameter group number; u is the number of monitoring parameter groups; y i is the label corresponding to the i-th group of monitoring parameters, is the predicted label based on the i-th set of monitoring parameters.

[0081] The first machine learning model can be any one of a random forest, a deep learning model or a long short-term memory network; other model parameters of the first machine learning model, such as the depth of the long short-term memory network, the number of neurons in each layer, the activation function used by the long short-term memory network, and the optimization of the loss function, are all obtained through actual engineering implementation and continuous experimental tuning.

[0082] The processing unit is used to generate an abnormality category based on the real-time monitoring parameters and the trained abnormality determination model, and mark the real-time monitoring parameters with abnormal categories as abnormal parameters.

[0083] The rescue identification module is used to calculate the degree of danger based on abnormal parameters and generate a rescue plan based on the degree of danger. Specifically, the method for calculating the degree of danger based on abnormal parameters includes: collecting g real-time monitoring parameters, inputting the g real-time monitoring parameters into an abnormal determination model, and outputting g′ abnormal categories; marking the real-time monitoring parameters with abnormal categories as abnormal parameters, removing the real-time monitoring parameters with normal abnormal categories, obtaining g″ abnormal parameters, and marking the total number of g″ abnormal parameters as abnormal quantity; calculating the abnormal degree of g″ abnormal parameters one by one, obtaining g″′ abnormal degrees, g″′=g″≤g′=g, formulating the g″′ abnormal degrees to obtain a comprehensive abnormal degree, and calculating the degree of danger based on the comprehensive abnormal degree and the abnormal quantity;

[0084] The method for generating a rescue plan based on the degree of danger includes: comparing the degree of danger with a preset danger level threshold, the preset danger level threshold is set by technical personnel in this field based on a large amount of experimental data, and is not specifically limited here; if the degree of danger is greater than the preset danger level threshold, the rescue plan is determined to be a manual rescue plan; if the degree of danger is less than or equal to the preset danger level threshold, the rescue plan is determined to be an automatic rescue plan, the automatic rescue plan is immediately executed and the progress of the automatic rescue plan is monitored, and based on the progress of the automatic rescue plan, it is determined whether to switch to a manual rescue plan.

[0085] See also Figure 4 As shown, in some embodiments, the system also includes a supervision module, which is used to monitor the progress of the automatic rescue plan. The method for monitoring the progress of the automatic rescue plan and determining whether to terminate the automatic rescue includes: when the automatic rescue plan is executed, real-time monitoring of the automatic rescue time, the abnormality reduction coefficient and the comprehensive abnormality degree reduction coefficient; based on the automatic rescue time, the abnormality reduction coefficient and the comprehensive abnormality degree reduction coefficient, a rescue progress value is generated in real time, and the rescue progress value is compared with a preset rescue progress value threshold. The preset rescue progress value threshold is determined by technical personnel in this field based on historical rescue data statistics or experimental data fitting to determine whether to generate a stop command; if the rescue progress value is greater than or equal to the preset rescue progress value threshold, a stop command is generated and manual rescue is immediately switched; if the rescue progress value is less than the preset rescue progress value threshold, no stop command is generated, and the rescue progress value is continuously updated.

[0086] In addition, the parts of the above-mentioned technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.

[0087] The specific implementation modes as described above further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation mode of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0088] The above preset parameters or preset thresholds are all set by technicians in this field according to actual conditions or obtained by simulating a large amount of data.

[0089] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An emergency rescue method for emergency failure events, characterized in that: Methods include: Define the monitoring scope and the independently operating individuals within the monitoring scope; Collect historical operation data of the independently operated individuals, the historical operation data including monitoring parameters that affect the normal operation of the independently operated individuals and abnormal categories corresponding to the monitoring parameters, the abnormal categories including abnormal and normal; Anomaly determination models that are trained to identify anomaly categories based on historical operation data; Obtain g real-time monitoring parameters of the independent operating individuals in unit time, input the g real-time monitoring parameters into the trained abnormality judgment model, output g′ abnormality categories, and mark the real-time monitoring parameters with abnormal categories as abnormal parameters; The degree of danger is calculated based on abnormal parameters, and a rescue plan is generated based on the degree of danger.

2. The emergency rescue method for emergency failure events according to claim 1, characterized in that: Methods for calculating the severity of danger based on abnormal parameters include: Remove the real-time monitoring parameters with normal abnormal categories from the g′ abnormal categories to obtain g″ abnormal parameters, and mark the total number of g″ abnormal parameters as the abnormal amount; Calculate the abnormality degree of g″ abnormal parameters one by one, and obtain g″′ abnormality degrees, g″′=g″≤g′=g; Comprehensively analyze and calculate the g″′ abnormality levels to obtain the comprehensive abnormality level; The degree of danger is calculated based on the comprehensive abnormality degree and abnormality amount.

3. The emergency rescue method for emergency failure events according to claim 2, characterized in that: The rescue plan includes automatic rescue and manual rescue. The methods for generating rescue plans based on the degree of danger include: Compare the danger level with a preset danger level threshold; If the danger level is greater than the preset danger level threshold, the rescue plan is determined to be a manual rescue plan; If the danger level is less than or equal to the preset danger level threshold, the rescue plan is determined to be an automatic rescue plan, the automatic rescue plan is immediately executed and the progress of the automatic rescue plan is monitored, and based on the progress of the automatic rescue plan, it is determined whether to switch to a manual rescue plan.

4. The emergency rescue method for emergency failure events according to claim 3, characterized in that: Methods for monitoring the progress of automatic rescue solutions include: When the automatic rescue plan is executed, the automatic rescue time, the abnormal quantity reduction coefficient and the comprehensive abnormal degree reduction coefficient are monitored in real time; The rescue progress value is generated in real time based on the automatic rescue time, the abnormality reduction coefficient and the comprehensive abnormality degree reduction coefficient.

5. The emergency rescue method for emergency failure events according to claim 4, characterized in that: The method for determining whether to switch to a manual rescue solution based on the progress of the automatic rescue solution includes: Compare the rescue progress value with a preset rescue progress value threshold to determine whether to generate a stop instruction; If the rescue progress value is greater than or equal to the preset rescue progress value threshold, a stop command is generated and manual rescue is immediately performed; If the rescue progress value is less than the preset rescue progress value threshold, no stop instruction is generated, and the rescue progress value is continuously updated.

6. The emergency rescue method for emergency failure events according to claim 4, characterized in that: The time taken for automatic rescue is the time taken for automatic rescue starting from the time when the automatic rescue scheme is started; The abnormal amount reduction coefficient is the rate of decrease of the total amount of abnormal parameters of the independently operating individual during the automatic rescue process; The comprehensive abnormality degree reduction coefficient is the rate of reduction of the comprehensive abnormality degree of the independently operating individual during the automatic rescue process.

7. The emergency rescue method for emergency failure events according to claim 1, characterized in that: The training method of the abnormality determination model includes: The label of the abnormal category corresponding to the monitoring parameter is normal is set to 0, and the label of the abnormal category corresponding to the monitoring parameter is abnormal is set to 1, and the monitoring parameters and the labels corresponding to the monitoring parameters are constructed as a data set of the machine learning model; the data set is divided into a training set, a validation set and a test set; The training set is used as the input of the machine learning model, and the machine learning model uses the predicted label corresponding to the monitoring parameter as the output; the actual label corresponding to the monitoring parameter is used as the prediction target, and minimizing the loss function value of the machine learning model is used as the training target; when the loss function value of the machine learning model is less than or equal to the preset target loss value, the training is stopped, and the machine learning model obtained by training is used as the abnormality judgment model; The machine learning model is a random forest, a deep learning model or a long short-term memory network.

8. The emergency rescue method for emergency failure events according to claim 1, characterized in that: The monitoring parameters include status data, electrical data and environmental data; the status data include temperature, vibration and noise when the independent individual is running; the electrical data include power, current and voltage when the independent individual is running; the environmental data include ambient temperature, ambient humidity and load.

9. The emergency rescue method for emergency failure events according to claim 3, characterized in that: The automatic rescue is an automatic rescue program manually provided for an independently operating individual.

10. An emergency rescue system for emergency failure events, which is applied to an emergency rescue method for emergency failure events as claimed in any one of claims 1 to 9, characterized in that: The system includes: A monitoring unit is arranged on the independently operated individual and is used to monitor the real-time monitoring parameters of the independently operated individual; A training module is used to collect historical operation data of independent operating individuals and train an abnormality determination model based on the historical operation data; A processing unit, used for generating an abnormality category based on the real-time monitoring parameters and the trained abnormality determination model, and marking the real-time monitoring parameters with abnormal categories as abnormal parameters; The rescue identification module is used to calculate the degree of danger based on abnormal parameters and generate a rescue plan based on the degree of danger.

Images