Intelligent early warning method and system based on hydrogen leakage danger degree in industrial production

Abstract

The present application relates to hydrogen leakage intelligent early warning technical field, a kind of intelligent early warning method and system based on hydrogen leakage dangerous degree in industrial production, method includes: in hydrogen leakage historical data set sequentially extracting hydrogen leakage historical data, using interval leakage concentration difference, leakage ambient temperature, leakage ambient humidity, leakage ambient pressure and target leakage concentration to train initial BP neural network, obtain target BP neural network, judge whether current hydrogen concentration is greater than first leakage concentration, if greater than, according to first leakage concentration and second leakage concentration calculate current leakage concentration difference, current leakage concentration difference, current ambient temperature, current ambient humidity and current ambient pressure are input into target BP neural network, obtain predicted leakage concentration, judge whether predicted leakage concentration is greater than early warning leakage concentration, if greater than, hydrogen leakage intelligent early warning is carried out.The present application can solve the problem that the monitoring precision is low and the monitoring effect is poor in the current hydrogen concentration monitoring.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hydrogen leakage intelligent early warning, and in particular to an intelligent early warning method and system based on hydrogen leakage danger level in industrial production. BACKGROUND

[0002] Hydrogen, as a clean and efficient energy carrier, has been widely used in industrial production, such as petroleum chemical industry, semiconductor manufacturing, fuel cell production and other fields. However, hydrogen has the characteristics of flammability and explosiveness, with a low explosion limit (LEL) of 4% by volume concentration, and a fast diffusion speed and low ignition energy. Once leakage occurs, it is easy to cause fire or explosion accidents, causing serious threats to personnel safety and equipment property.

[0003] The hydrogen production plant in industrial production is very large, with numerous pipelines. It is usually necessary to configure hydrogen detectors and other detection equipment to detect the hydrogen concentration in the working environment and comprehensively judge the danger level. However, if the early warning concentration is set too high, when the hydrogen detector detects that the hydrogen concentration is greater than the early warning concentration and alarms, a serious hydrogen leakage event may have occurred in the working environment. If the early warning concentration is set too low, the hydrogen detector may frequently alarm, affecting work. Therefore, the existing hydrogen detection method for large industrial production plants does not comprehensively consider the influence of the current environmental temperature, the current environmental humidity and the current environmental air pressure, the algorithm is not matched, and there is a problem of low monitoring precision and poor monitoring effect in monitoring the hydrogen concentration. SUMMARY

[0004] The present application provides an intelligent early warning method and system based on hydrogen leakage danger level in industrial production, which has high monitoring precision and better monitoring effect for a large amount of hydrogen in industrial production.

[0005] To achieve the above-mentioned purpose, the present application provides an intelligent early warning method based on hydrogen leakage danger level in industrial production, comprising:

[0006] Extracting hydrogen leakage historical data in a pre-constructed hydrogen leakage historical data set in sequence, and extracting interval leakage concentration difference, target leakage concentration, leakage environment temperature, leakage environment humidity and leakage environment air pressure in the hydrogen leakage historical data according to a pre-set first leakage concentration, interval sampling duration and leakage processing duration;

[0007] Training a pre-constructed initial BP neural network using the interval leakage concentration difference, the leakage environment temperature, the leakage environment humidity, the leakage environment air pressure and the target leakage concentration to obtain a target BP neural network, wherein the input layer of the initial BP neural network comprises a leakage concentration difference node, an environment temperature node, an environment humidity node and an environment air pressure node, and the output layer comprises a concentration prediction node;

[0008] monitoring a current hydrogen concentration;

[0009] determining whether the current hydrogen concentration is greater than the first leakage concentration;

[0010] if the current hydrogen concentration is not greater than the first leakage concentration, returning to the step of monitoring the current hydrogen concentration;

[0011] if the current hydrogen concentration is greater than the first leakage concentration, collecting a second leakage concentration according to the interval sampling duration, and calculating a current leakage concentration difference according to the first leakage concentration and the second leakage concentration, wherein the current leakage concentration difference refers to a difference between the second leakage concentration and the first leakage concentration;

[0012] collecting a current environment temperature, a current environment humidity, and a current environment pressure, inputting the current leakage concentration difference, the current environment temperature, the current environment humidity, and the current environment pressure into the target BP neural network, and obtaining a predicted leakage concentration;

[0013] determining whether the predicted leakage concentration is greater than a preset early warning leakage concentration;

[0014] if the predicted leakage concentration is not greater than the early warning leakage concentration, returning to the step of monitoring the current hydrogen concentration;

[0015] if the predicted leakage concentration is greater than the early warning leakage concentration, performing hydrogen leakage intelligent early warning.

[0016] Optionally, the extracting, according to the preset first leakage concentration, the interval sampling duration, and the leakage processing duration, the interval leakage concentration difference, the target leakage concentration, the leakage environment temperature, the leakage environment humidity, and the leakage environment pressure from the hydrogen leakage historical data comprises:

[0017] identifying a historical first leakage time of the first leakage concentration in the hydrogen leakage historical data;

[0018] calculating a historical second leakage time according to the historical first leakage time and the interval sampling duration, wherein the interval sampling duration is equal to a time difference between the historical first leakage time and the historical second leakage time;

[0019] extracting a historical second leakage concentration from the hydrogen leakage historical data according to the historical second leakage time;

[0020] calculating an interval leakage concentration difference according to the first leakage concentration and the historical second leakage concentration, wherein the interval leakage concentration difference is equal to a difference between the historical second leakage concentration and the first leakage concentration;

[0021] extracting a target leakage concentration from the hydrogen leakage historical data according to the historical second leakage time and the leakage processing duration;

[0022] extracting, from the hydrogen leakage history data, a leakage ambient temperature, a leakage ambient humidity, and a leakage ambient pressure at the second leakage time.

[0023] Optionally, before the extracting, from the hydrogen leakage history data, an interval leakage concentration difference, a target leakage concentration, a leakage ambient temperature, a leakage ambient humidity, and a leakage ambient pressure according to the preset first leakage concentration, an interval sampling duration, and a leakage processing duration, the method further comprises:

[0024] extracting test interval durations in sequence in a preset interval duration gradient sequence;

[0025] extracting hydrogen leakage history data in sequence in the hydrogen leakage history data set;

[0026] identifying a pre-warning leakage time in the hydrogen leakage history data according to the pre-warning leakage concentration, wherein the pre-warning leakage time refers to a leakage time corresponding to a first occurrence of the pre-warning leakage concentration in the hydrogen leakage history data;

[0027] calculating a cutoff sampling time using the pre-warning leakage time according to the leakage processing duration and a preset safety reserved duration, wherein a time difference between the pre-warning leakage time and the cutoff sampling time is equal to a sum of the leakage processing duration and the safety reserved duration;

[0028] determining a sampling value interval according to the historical first leakage time and the cutoff sampling time;

[0029] calculating a starting sampling time using the following formula according to the test interval duration and the historical first leakage time:

[0030] T start =T1+t,T start ∈(T1,T end )

[0031] wherein T start represents the starting sampling time, T1 represents the historical first leakage time, t represents the test interval duration, (T1, T end ) represents the sampling value interval, and T end represents the cutoff sampling time;

[0032] calculating an interval test concentration difference using the hydrogen leakage history data according to the first leakage concentration and the starting sampling time;

[0033] calculating a target test time according to the starting sampling time and the leakage processing duration, wherein a difference between the target test time and the starting sampling time is equal to the leakage processing duration;

[0034] extracting a target test concentration in the hydrogen leakage history data according to the target test time;

[0035] training the pre-constructed initial BP neural network according to the interval test concentration difference, the leakage environment temperature, the leakage environment humidity, the leakage environment air pressure and the target test concentration to obtain an iterative BP neural network;

[0036] judging whether the iterative BP neural network is trained completely;

[0037] if the iterative BP neural network is not trained completely, returning to the step of extracting the hydrogen leakage history data in the hydrogen leakage history data set in sequence;

[0038] if the iterative BP neural network is trained completely, obtaining a BP neural network to be verified;

[0039] verifying the prediction accuracy of the BP neural network to be verified according to the first leakage concentration, the test interval length and the leakage processing length to obtain the prediction accuracy;

[0040] collecting the prediction accuracy corresponding to each test interval length to obtain a prediction accuracy set, and extracting a maximum prediction accuracy in the prediction accuracy set;

[0041] identifying the target interval length corresponding to the maximum prediction accuracy, and taking the target interval length as an interval sampling length.

[0042] Optionally, the extracting a target leakage concentration in the hydrogen leakage history data according to the historical second leakage time and the leakage processing length comprises:

[0043] calculating a target leakage time according to the historical second leakage time and the leakage processing length, wherein the leakage processing length is equal to the difference between the target leakage time and the historical second leakage time;

[0044] extracting a target leakage concentration in the hydrogen leakage history data according to the target leakage time.

[0045] Optionally, the training the pre-constructed initial BP neural network according to the interval leakage concentration difference, the leakage environment temperature, the leakage environment humidity, the leakage environment air pressure and the target leakage concentration to obtain a target BP neural network comprises:

[0046] inputting the interval leakage concentration difference, the leakage environment temperature, the leakage environment humidity and the leakage environment air pressure into a leakage concentration difference node, an environment temperature node, an environment humidity node and an environment air pressure node of the initial BP neural network respectively to obtain an iterative prediction leakage concentration;

[0047] calculating a prediction error according to the iteratively predicted leakage concentration and the target leakage concentration;

[0048] determining whether the prediction error is less than a preset error threshold value;

[0049] if the prediction error is not less than the error threshold value, performing feedback adjustment on the initial BP neural network according to the prediction error, and returning to the step of sequentially extracting hydrogen leakage historical data in the pre-constructed hydrogen leakage historical data set;

[0050] if the prediction error is less than the error threshold value, obtaining a target BP neural network.

[0051] Optionally, the collecting the second leakage concentration according to the interval sampling duration comprises:

[0052] obtaining a current time, and calculating a second sampling time according to the current time and the interval sampling duration;

[0053] collecting the second leakage concentration according to the second sampling time.

[0054] Optionally, the calculating the interval test concentration difference by using the hydrogen leakage historical data according to the first leakage concentration and the starting sampling time comprises:

[0055] extracting a starting sampling concentration in the hydrogen leakage historical data according to the starting sampling time;

[0056] calculating an interval test concentration difference according to the starting sampling concentration and the first leakage concentration, wherein the interval test concentration difference is equal to the difference between the starting sampling concentration and the first leakage concentration.

[0057] Optionally, the performing prediction accuracy verification on the to-be-verified BP neural network according to the first leakage concentration, the test interval duration and the leakage processing duration comprises:

[0058] randomly extracting batch hydrogen leakage verification data in the hydrogen leakage historical data set, wherein the number of hydrogen leakage verification data in the batch hydrogen leakage verification data is equal to a preset batch number;

[0059] extracting a verification data set in the batch hydrogen leakage verification data according to the first leakage concentration, the test interval duration and the leakage processing duration, wherein the verification data in the verification data set comprises an interval verification concentration difference, a target verification concentration, a verification environment temperature, a verification environment humidity and a verification environment air pressure;

[0060] performing prediction accuracy verification on the to-be-verified BP neural network by using the verification data set, and obtaining a prediction accuracy.

[0061] Optionally, the prediction accuracy checking of the to-be-checked BP neural network by using the checking data set comprises:

[0062] extracting checking data in the checking data set in sequence;

[0063] inputting the interval verification concentration difference, the verification environment temperature, the verification environment humidity and the verification environment air pressure in the checking data into the to-be-checked BP neural network to obtain a to-be-checked predicted concentration;

[0064] calculating the reciprocal of the prediction difference value of the to-be-checked predicted concentration and the target verification concentration to obtain a prediction difference reciprocal set, wherein the prediction difference reciprocal is the reciprocal of the difference value of the to-be-checked predicted concentration and the target verification concentration;

[0065] calculating the prediction accuracy by using the following formula according to the prediction difference reciprocal set:

[0066]

[0067] wherein z represents the prediction accuracy, c1 represents the first prediction difference reciprocal in the prediction difference reciprocal set, c2 represents the second prediction difference reciprocal in the prediction difference reciprocal set, c m represents the mth prediction difference reciprocal in the prediction difference reciprocal set.

[0068] To achieve the above object, the application further provides an intelligent early warning system based on the hydrogen leakage danger degree in industrial production, comprising:

[0069] a neural network training module, configured to extract hydrogen leakage historical data in a pre-constructed hydrogen leakage historical data set in sequence, extract an interval leakage concentration difference, a target leakage concentration, a leakage environment temperature, a leakage environment humidity and a leakage environment air pressure in the hydrogen leakage historical data according to a pre-set first leakage concentration, an interval sampling time length and a leakage processing time length, and train a pre-constructed initial BP neural network by using the interval leakage concentration difference, the leakage environment temperature, the leakage environment humidity, the leakage environment air pressure and the target leakage concentration to obtain a target BP neural network, wherein the input layer of the initial BP neural network comprises a leakage concentration difference node, an environment temperature node, an environment humidity node and an environment air pressure node, and the output layer comprises a concentration prediction node;

[0070] a hydrogen concentration judgment module, configured to monitor a current hydrogen concentration and judge whether the current hydrogen concentration is greater than the first leakage concentration, and return to the step of monitoring the current hydrogen concentration if the current hydrogen concentration is not greater than the first leakage concentration;

[0071] The leakage concentration comparison module is used for, if the current hydrogen concentration is greater than the first leakage concentration, collecting a second leakage concentration according to the interval sampling duration, calculating a current leakage concentration difference according to the first leakage concentration and the second leakage concentration, wherein the current leakage concentration difference refers to a difference between the second leakage concentration and the first leakage concentration; collecting a current environment temperature, a current environment humidity and a current environment air pressure, inputting the current leakage concentration difference, the current environment temperature, the current environment humidity and the current environment air pressure into the target BP neural network to obtain a predicted leakage concentration; judging whether the predicted leakage concentration is greater than a preset early warning leakage concentration; if the predicted leakage concentration is not greater than the early warning leakage concentration, returning to the step of monitoring the current hydrogen concentration.

[0072] The leakage intelligent early warning module is used for, if the predicted leakage concentration is greater than the early warning leakage concentration, performing hydrogen leakage intelligent early warning.

[0073] To solve the above problems, the application further provides an electronic device, which comprises:

[0074] a memory for storing at least one instruction; and

[0075] a processor for executing the instruction stored in the memory to realize the intelligent early warning method based on the hydrogen leakage danger degree in industrial production.

[0076] To solve the above problems, the application further provides a computer readable storage medium, which stores at least one instruction, and the at least one instruction is executed by a processor in an electronic device to realize the intelligent early warning method based on the hydrogen leakage danger degree in industrial production.

[0077] The method of the application is to determine whether to perform hydrogen leakage early warning by predicting the hydrogen concentration through the target BP neural network and judging the size of the predicted concentration and the early warning leakage concentration, therefore, the training of the initial BP neural network is required first, before the training, the hydrogen leakage historical data is extracted in the pre-constructed hydrogen leakage historical data set in sequence, then the interval leakage concentration difference, the target leakage concentration, the leakage environment temperature, the leakage environment humidity and the leakage environment air pressure are extracted from the hydrogen leakage historical data according to the preset first leakage concentration, interval sampling duration and leakage processing duration, at this time, the interval leakage concentration difference, the leakage environment temperature, the leakage environment humidity, the leakage environment air pressure and the target leakage concentration are used to train the pre-constructed initial BP neural network to obtain the target BP neural network, wherein the input layer of the initial BP neural network comprises a leakage concentration difference node, an environment temperature node, an environment humidity node and an environment air pressure node, and the output layer comprises a concentration prediction node.

[0078] In order to avoid frequent prediction of the hydrogen concentration, a first leakage concentration is set, and the current hydrogen concentration is monitored to determine whether the current hydrogen concentration is greater than the first leakage concentration, if the current hydrogen concentration is not greater than the first leakage concentration, the step of monitoring the current hydrogen concentration is returned, if the current hydrogen concentration is greater than the first leakage concentration, a second leakage concentration is collected according to the interval sampling duration, and the current leakage concentration difference is calculated according to the first leakage concentration and the second leakage concentration, wherein the current leakage concentration difference refers to the difference between the second leakage concentration and the first leakage concentration, due to the influence of environmental factors, the current environmental temperature, the current environmental humidity and the current environmental pressure are collected, and the current leakage concentration difference, the current environmental temperature, the current environmental humidity and the current environmental pressure are input into the target BP neural network to obtain the predicted leakage concentration, at this time, it is determined whether the predicted leakage concentration is greater than the preset early warning leakage concentration, if the predicted leakage concentration is not greater than the early warning leakage concentration, the step of monitoring the current hydrogen concentration is returned, if the predicted leakage concentration is greater than the early warning leakage concentration, the hydrogen leakage intelligent early warning is performed.

[0079] Therefore, the detection method and system of the present application not only detect the hydrogen leakage situation, but also comprehensively detect the current environmental temperature, the current environmental humidity and the current environmental pressure, and use a unique new algorithm based on historical data, so that the monitoring accuracy of the hydrogen concentration in industrial production is high, the alarm reaction speed is fast, the intelligent level is high, and more environmental factors are considered than the existing hydrogen leakage monitoring method. BRIEF DESCRIPTION OF DRAWINGS

[0080] Figure 1 The flowchart of the intelligent early warning method based on the hydrogen leakage danger degree in industrial production provided by an embodiment of the present application is shown;

[0081] Figure 2 The functional module diagram of the intelligent early warning system based on the hydrogen leakage danger degree in industrial production provided by an embodiment of the present application is shown;

[0082] Figure 3 The structural diagram of the electronic device for realizing the intelligent early warning method based on the hydrogen leakage danger degree in industrial production provided by an embodiment of the present application is shown.

[0083] The implementation of the present application, the functional characteristics and advantages will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0084] It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0085] The embodiment of the present application provides an intelligent early warning method based on the hydrogen leakage danger degree in industrial production. The execution subject of the intelligent early warning method based on the hydrogen leakage danger degree in industrial production includes but is not limited to at least one of the electronic devices capable of being configured to execute the method provided by the embodiment of the present application, such as a server, a terminal and the like. In other words, the intelligent early warning method based on the hydrogen leakage danger degree in industrial production can be executed by the software or hardware installed in the terminal device or the server device, and the software can be a blockchain platform. The server includes but is not limited to a single server, a server cluster, a cloud server or a cloud server cluster and the like.

[0086] Referring to Figure 1 As shown in the figure, the flowchart of the intelligent early warning method based on the hydrogen leakage danger degree in industrial production provided by the embodiment of the present application. In the embodiment, the intelligent early warning method based on the hydrogen leakage danger degree in industrial production includes:

[0087] S1, extracting hydrogen leakage historical data in a pre-constructed hydrogen leakage historical data set in sequence, and extracting interval leakage concentration difference, target leakage concentration, leakage environment temperature, leakage environment humidity and leakage environment air pressure in the hydrogen leakage historical data according to a preset first leakage concentration, interval sampling duration and leakage processing duration.

[0088] As can be understood, the hydrogen leakage historical data set refers to a collection of monitoring data of past hydrogen leakage accidents, and the monitoring data includes hydrogen real-time concentration, environment humidity, environment temperature and environment air pressure when the hydrogen leakage accident occurs. Since the environment humidity, environment temperature and environment air pressure will affect the explosion concentration of hydrogen when the hydrogen explosion occurs, therefore, the environment temperature, environment humidity and environment air pressure need to be taken as the influencing factors of the hydrogen leakage danger degree in industrial production. The first leakage concentration refers to a preset hydrogen leakage concentration threshold value for starting hydrogen concentration prediction, for example, 1% volume concentration. The interval sampling duration refers to the interval duration for collecting hydrogen leakage concentration after collecting the first leakage concentration. The leakage processing duration refers to a preset processing duration for which the staff can properly handle the leakage event. The interval leakage concentration difference refers to the concentration difference of two hydrogen leakage concentrations collected according to the interval sampling duration. The target leakage concentration refers to the leakage concentration of hydrogen after the interval sampling duration and the leakage processing duration after collecting the first leakage concentration. The leakage environment temperature refers to the environment temperature of the hydrogen leakage site, the leakage environment humidity refers to the environment humidity of the hydrogen leakage site, and the leakage environment air pressure refers to the environment air pressure of the hydrogen leakage site.

[0089] In the embodiment of the present application, the interval leakage concentration difference, the target leakage concentration, the leakage environment temperature, the leakage environment humidity and the leakage environment pressure are extracted from the hydrogen leakage history data according to the preset first leakage concentration, the interval sampling duration and the leakage processing duration, which comprises:

[0090] The historical first leakage time of the first leakage concentration is identified in the hydrogen leakage history data;

[0091] The historical second leakage time is calculated according to the historical first leakage time and the interval sampling duration, wherein the interval sampling duration is equal to the time difference between the historical first leakage time and the historical second leakage time;

[0092] The historical second leakage concentration is extracted from the hydrogen leakage history data according to the historical second leakage time;

[0093] The interval leakage concentration difference is calculated according to the first leakage concentration and the historical second leakage concentration, wherein the interval leakage concentration difference is equal to the difference between the historical second leakage concentration and the first leakage concentration;

[0094] The target leakage concentration is extracted from the hydrogen leakage history data according to the historical second leakage time and the leakage processing duration;

[0095] The leakage environment temperature, the leakage environment humidity and the leakage environment pressure of the second leakage time are extracted from the hydrogen leakage history data.

[0096] Further, the historical first leakage time refers to the corresponding time of the first leakage concentration in the hydrogen leakage history data, the historical second leakage time refers to the corresponding time of the historical second leakage concentration in the hydrogen leakage history data, and the historical second leakage concentration refers to the hydrogen leakage concentration collected at the time point of the historical first leakage time plus the interval sampling duration.

[0097] In the embodiment of the present application, the target leakage concentration is extracted from the hydrogen leakage history data according to the historical second leakage time and the leakage processing duration, which comprises:

[0098] The target leakage time is calculated according to the historical second leakage time and the leakage processing duration, wherein the leakage processing duration is equal to the difference between the target leakage time and the historical second leakage time;

[0099] The target leakage concentration is extracted from the hydrogen leakage history data according to the target leakage time.

[0100] In the embodiment of the present application, before extracting the interval leakage concentration difference, the target leakage concentration, the leakage environment temperature, the leakage environment humidity and the leakage environment air pressure from the hydrogen leakage history data according to the preset first leakage concentration, the interval sampling duration and the leakage processing duration, the method further comprises:

[0101] extracting the test interval duration in sequence in the preset interval duration gradient sequence;

[0102] extracting the hydrogen leakage history data in sequence in the hydrogen leakage history data set;

[0103] identifying the early warning leakage time in the hydrogen leakage history data according to the early warning leakage concentration, wherein the early warning leakage time refers to the leakage time corresponding to the first occurrence of the early warning leakage concentration in the hydrogen leakage history data;

[0104] calculating the cut-off sampling time by using the early warning leakage time according to the leakage processing duration and the preset safety reserved duration, wherein the time difference between the early warning leakage time and the cut-off sampling time is equal to the sum of the leakage processing duration and the safety reserved duration;

[0105] determining the sampling value interval according to the historical first leakage time and the cut-off sampling time;

[0106] calculating the starting sampling time by using the following formula according to the test interval duration and the historical first leakage time:

[0107] T start = T1 + t, T start ∈ (T1, Tend)

[0108] wherein T start represents the starting sampling time, T1 represents the historical first leakage time, t represents the test interval duration, (T1, T end ) represents the sampling value interval, and T end represents the cut-off sampling time;

[0109] calculating the interval test concentration difference by using the hydrogen leakage history data according to the first leakage concentration and the starting sampling time;

[0110] calculating the target test time according to the starting sampling time and the leakage processing duration, wherein the difference between the target test time and the starting sampling time is equal to the leakage processing duration;

[0111] extracting the target test concentration from the hydrogen leakage history data according to the target test time;

[0112] training the pre-constructed initial BP neural network according to the interval test concentration difference, the leakage ambient temperature, the leakage ambient humidity, the leakage ambient pressure, and the target test concentration to obtain an iterative BP neural network;

[0113] determining whether the iterative BP neural network is trained;

[0114] If the iterative BP neural network is not trained, returning to the step of sequentially extracting the hydrogen leakage historical data in the hydrogen leakage historical data set;

[0115] If the iterative BP neural network is trained, obtaining a BP neural network to be verified;

[0116] performing prediction accuracy verification on the BP neural network to be verified according to the first leakage concentration, the test interval length, and the leakage processing length to obtain prediction accuracy;

[0117] collecting the prediction accuracy corresponding to each test interval length to obtain a prediction accuracy set, and extracting the maximum prediction accuracy in the prediction accuracy set;

[0118] identifying the target interval length corresponding to the maximum prediction accuracy, and taking the target interval length as an interval sampling length.

[0119] Understandably, the interval length gradient sequence refers to a length gradient sequence preset by a user to test the interval sampling length, for example, 1s, 20s, 39s, 58s, 77s, and the like. The early warning leakage time refers to a time corresponding to the first occurrence of the early warning leakage concentration in the hydrogen leakage historical data. The safety reserved length refers to a time fault tolerance value preset by a user, for example, 5min. The cut-off sampling time refers to the maximum value of the sampling value interval of the starting sampling time. The sampling value interval refers to the interval of the sampling time of the starting sampling time. The interval test concentration difference refers to the difference between the first leakage concentration and the hydrogen concentration corresponding to the starting sampling time in the hydrogen leakage historical data. The target test time refers to the sampling time of the target test concentration. The target test concentration refers to the hydrogen concentration corresponding to the target test time in the hydrogen leakage historical data. The initial BP neural network refers to an untrained BP neural network, and the initial BP neural network includes an input layer, a hidden layer, and an output layer, which is described in detail in S2. The iterative BP neural network refers to the initial BP neural network in the iterative training. The BP neural network to be verified refers to a completed training initial BP neural network that needs to be verified for prediction accuracy. The prediction accuracy refers to the accuracy of the hydrogen concentration predicted by the BP neural network to be verified.

[0120] In the embodiment of the present application, the interval test concentration difference is calculated by using the hydrogen leakage history data according to the first leakage concentration and the starting sampling time, and the method comprises the following steps:

[0121] The starting sampling concentration is extracted from the hydrogen leakage history data according to the starting sampling time;

[0122] The interval test concentration difference is calculated according to the starting sampling concentration and the first leakage concentration, wherein the interval test concentration difference is equal to the difference between the starting sampling concentration and the first leakage concentration.

[0123] It can be understood that the starting sampling concentration refers to the hydrogen concentration corresponding to the starting sampling time in the hydrogen leakage history data.

[0124] In the embodiment of the present application, the prediction accuracy of the BP neural network to be verified is verified according to the first leakage concentration, the test interval length and the leakage processing length, and the prediction accuracy is obtained, and the method comprises the following steps:

[0125] The batch hydrogen leakage verification data is randomly extracted from the hydrogen leakage history data set, wherein the number of hydrogen leakage verification data in the batch hydrogen leakage verification data is equal to the preset batch number;

[0126] The verification data set is extracted from the batch hydrogen leakage verification data according to the first leakage concentration, the test interval length and the leakage processing length, wherein the verification data in the verification data set comprises an interval verification concentration difference, a target verification concentration, a verification environment temperature, a verification environment humidity and a verification environment pressure;

[0127] The prediction accuracy of the BP neural network to be verified is verified by using the verification data set, and the prediction accuracy is obtained.

[0128] Further, the batch hydrogen leakage verification data refers to a set composed of a plurality of hydrogen leakage history data randomly extracted from the hydrogen leakage history data set. The preset batch number can be 50. The interval verification concentration difference refers to an interval leakage concentration difference used for verifying the prediction accuracy of the BP neural network to be verified. The extraction method of the verification data set can refer to the step of extracting an interval leakage concentration difference, a target leakage concentration, a leakage environment temperature, a leakage environment humidity and a leakage environment pressure from the hydrogen leakage history data according to the preset first leakage concentration, interval sampling length and leakage processing length, which is not described herein again. The interval verification concentration difference refers to an interval leakage concentration difference used for verifying the prediction accuracy. The target verification concentration refers to a target leakage concentration used for verifying the prediction accuracy. The verification environment temperature refers to a leakage environment temperature used for verifying the prediction accuracy. The verification environment humidity refers to a leakage environment humidity used for verifying the prediction accuracy. The verification environment pressure refers to a leakage environment pressure used for verifying the prediction accuracy.

[0129] In the embodiment of the present application, the prediction accuracy checking of the BP neural network to be checked by using the checking data set comprises:

[0130] checking data are extracted from the checking data set in sequence;

[0131] the interval verification concentration difference, the verification environment temperature, the verification environment humidity and the verification environment air pressure in the checking data are input into the BP neural network to be checked to obtain a predicted concentration to be checked;

[0132] a reciprocal of a prediction difference value of the predicted concentration to be checked and the target verification concentration is calculated to obtain a reciprocal of a prediction difference value set, wherein the reciprocal of the prediction difference value is a reciprocal of a difference value between the predicted concentration to be checked and the target verification concentration;

[0133] the prediction accuracy is calculated by using the following formula according to the reciprocal of the prediction difference value set:

[0134]

[0135] wherein z represents the prediction accuracy, c1 represents the first reciprocal of the prediction difference value in the reciprocal of the prediction difference value set, c2 represents the second reciprocal of the prediction difference value in the reciprocal of the prediction difference value set, and cm represents the mthreciprocal of the prediction difference value in the reciprocal of the prediction difference value set. m The mthreciprocal of the prediction difference value in the reciprocal of the prediction difference value set is represented by cm.

[0136] S2, the initial BP neural network is trained by using the interval leakage concentration difference, the leakage environment temperature, the leakage environment humidity, the leakage environment air pressure and the target leakage concentration to obtain a target BP neural network.

[0137] It should be understood that the target BP neural network refers to the initial BP neural network after training.

[0138] In detail, the input layer of the initial BP neural network comprises a leakage concentration difference node, an environment temperature node, an environment humidity node and an environment air pressure node, and the output layer comprises a concentration prediction node.

[0139] In the embodiment of the present application, the initial BP neural network is trained by using the interval leakage concentration difference, the leakage environment temperature, the leakage environment humidity, the leakage environment air pressure and the target leakage concentration to obtain a target BP neural network, which comprises:

[0140] the interval leakage concentration difference, the leakage environment temperature, the leakage environment humidity and the leakage environment air pressure are respectively input into the leakage concentration difference node, the environment temperature node, the environment humidity node and the environment air pressure node of the initial BP neural network to obtain an iterative predicted leakage concentration;

[0141] calculating a prediction error according to the iteratively predicted leak concentration and the target leak concentration;

[0142] determining whether the prediction error is less than a preset error threshold value;

[0143] If the prediction error is not less than the error threshold value, the initial BP neural network is adjusted according to the prediction error, and the step of sequentially extracting hydrogen leak historical data in the pre-constructed hydrogen leak historical data set is returned.

[0144] If the prediction error is less than the error threshold value, a target BP neural network is obtained.

[0145] Understandably, the iteratively predicted leak concentration refers to the hydrogen leak concentration prediction value output from the concentration prediction node after different interval leak concentration differences, leak environment temperatures, leak environment humidities and leak environment pressures corresponding to the hydrogen leak historical data are input into the initial BP neural network each time. The prediction error refers to the difference between the iteratively predicted leak concentration and the target leak concentration.

[0146] S3, monitoring the current hydrogen concentration.

[0147] Understandably, the current hydrogen concentration refers to the real-time concentration of hydrogen in the current environment.

[0148] Further, the current hydrogen concentration can also be determined by an image acquisition device using infrared thermal imaging technology to acquire images of the hydrogen leak site for image recognition to determine the leak position, leak amount and leak severity of the hydrogen storage device to determine whether the storage device has leaked. After image recognition, when it is determined that the storage device has leaked, the approximate leak position can be determined according to infrared thermal imaging, and a pipeline robot can be controlled to verify along the pipeline line to collect the sound size of the leak point to assist in determining the leak amount and severity.

[0149] S4, determining whether the current hydrogen concentration is greater than the first leak concentration.

[0150] If the current hydrogen concentration is not greater than the first leak concentration, the step of monitoring the current hydrogen concentration is returned.

[0151] Understandably, when the current hydrogen concentration is not greater than the first leak concentration, it indicates that the current hydrogen concentration needs to be warned, and therefore, the current hydrogen concentration can be monitored again.

[0152] If the current hydrogen concentration is greater than the first leak concentration, S5 is executed, a second leak concentration is acquired according to the interval sampling duration, and a current leak concentration difference is calculated according to the first leak concentration and the second leak concentration.

[0153] In detail, the current leakage concentration difference refers to a difference between the second leakage concentration and the first leakage concentration.

[0154] The second leakage concentration is explainable, and refers to a hydrogen leakage concentration collected at a time point after the interval sampling duration.

[0155] In the embodiment of the present application, the second leakage concentration is collected according to the interval sampling duration, comprising:

[0156] The current time is obtained, and the second sampling time is calculated according to the current time and the interval sampling duration.

[0157] The second leakage concentration is collected according to the second sampling time.

[0158] The second sampling time is explainable, and refers to a sampling time of the second leakage concentration.

[0159] S6, collect the current environment temperature, the current environment humidity and the current environment air pressure, input the current leakage concentration difference, the current environment temperature, the current environment humidity and the current environment air pressure into the target BP neural network, and obtain the predicted leakage concentration.

[0160] Further, the current environment temperature, the current environment humidity and the current environment air pressure respectively refer to the temperature, the humidity and the air pressure in the current environment. The predicted leakage concentration refers to a hydrogen leakage concentration at the second sampling time interval leakage processing duration predicted by the target BP neural network according to the current leakage concentration difference, the current environment temperature, the current environment humidity and the current environment air pressure.

[0161] S7, judge whether the predicted leakage concentration is greater than a preset warning leakage concentration.

[0162] The warning leakage concentration is explainable, and refers to a hydrogen concentration that needs to be alarmed. For example: 4% volume concentration.

[0163] If the predicted leakage concentration is not greater than the warning leakage concentration, return to the step of monitoring the current hydrogen concentration.

[0164] If the predicted leakage concentration is greater than the warning leakage concentration, perform S8, and perform hydrogen leakage intelligent warning.

[0165] The application is to predict the hydrogen concentration by the target BP neural network, and to determine whether to carry out the hydrogen leakage early warning by judging the size of the predicted concentration and the early warning leakage concentration. Therefore, the initial BP neural network needs to be trained first. Before the training, the hydrogen leakage historical data is extracted in the pre-constructed hydrogen leakage historical data set in turn, and then the interval leakage concentration difference, target leakage concentration, leakage environment temperature, leakage environment humidity and leakage environment air pressure are extracted in the hydrogen leakage historical data according to the preset first leakage concentration, interval sampling duration and leakage processing duration. At this time, the interval leakage concentration difference, leakage environment temperature, leakage environment humidity, leakage environment air pressure and target leakage concentration are used to train the pre-constructed initial BP neural network to obtain the target BP neural network. The input layer of the initial BP neural network includes a leakage concentration difference node, an environment temperature node, an environment humidity node and an environment air pressure node, and the output layer includes a concentration prediction node. In order to avoid frequent prediction of hydrogen concentration, the first leakage concentration needs to be set, and the current hydrogen concentration is monitored to determine whether the current hydrogen concentration is greater than the first leakage concentration. If the current hydrogen concentration is not greater than the first leakage concentration, the step of monitoring the current hydrogen concentration is returned. If the current hydrogen concentration is greater than the first leakage concentration, the second leakage concentration is collected according to the interval sampling duration, and the current leakage concentration difference is calculated according to the first leakage concentration and the second leakage concentration. Due to the influence of environmental factors, the current environment temperature, current environment humidity and current environment air pressure need to be collected, and the current leakage concentration difference, current environment temperature, current environment humidity and current environment air pressure are input into the target BP neural network to obtain the predicted leakage concentration. At this time, it is judged whether the predicted leakage concentration is greater than the preset early warning leakage concentration. If the predicted leakage concentration is not greater than the early warning leakage concentration, the step of monitoring the current hydrogen concentration is returned. If the predicted leakage concentration is greater than the early warning leakage concentration, the intelligent hydrogen leakage early warning is carried out.

[0166] Therefore, the detection method and system of the application have high monitoring precision, fast alarm response speed and high intelligent level for the hydrogen concentration in industrial production, consider more environmental factors than the existing hydrogen leakage monitoring method, have 12-16% higher monitoring precision than the existing detection technology, meet the environmental protection and safety requirements of large-scale hydrogen industrial production, reduce the false alarm frequency by 6-9%, do not need to frequently shut down the equipment for inspection, and greatly improve the production efficiency.

[0167] As Figure 2 shown, it is a functional module diagram of the intelligent early warning system based on the hydrogen leakage danger degree in industrial production according to an embodiment of the application.

[0168] The intelligent early warning system 100 based on the danger degree of hydrogen leakage in industrial production can be installed in an electronic device. According to the functions to be realized, the intelligent early warning system 100 based on the danger degree of hydrogen leakage in industrial production can include a neural network training module 101, a hydrogen concentration judgment module 102, a leakage concentration comparison module 103, and a leakage intelligent early warning module 104. The modules in the present application can also be referred to as units, which refer to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, and are stored in the memory of the electronic device.

[0169] The neural network training module 101 is used to sequentially extract hydrogen leakage historical data in a pre-constructed hydrogen leakage historical data set, extract interval leakage concentration difference, target leakage concentration, leakage environment temperature, leakage environment humidity, and leakage environment air pressure in the hydrogen leakage historical data according to a preset first leakage concentration, interval sampling duration, and leakage processing duration; train the pre-constructed initial BP neural network using the interval leakage concentration difference, leakage environment temperature, leakage environment humidity, leakage environment air pressure, and target leakage concentration to obtain a target BP neural network, wherein the input layer of the initial BP neural network includes a leakage concentration difference node, an environment temperature node, an environment humidity node, and an environment air pressure node, and the output layer includes a concentration prediction node;

[0170] The hydrogen concentration judgment module 102 is used to monitor the current hydrogen concentration and judge whether the current hydrogen concentration is greater than the first leakage concentration; if the current hydrogen concentration is not greater than the first leakage concentration, return to the step of monitoring the current hydrogen concentration;

[0171] The leakage concentration comparison module 103 is used to, if the current hydrogen concentration is greater than the first leakage concentration, collect a second leakage concentration according to the interval sampling duration, calculate a current leakage concentration difference according to the first leakage concentration and the second leakage concentration, wherein the current leakage concentration difference refers to the difference between the second leakage concentration and the first leakage concentration; collect the current environment temperature, the current environment humidity, and the current environment air pressure, input the current leakage concentration difference, the current environment temperature, the current environment humidity, and the current environment air pressure into the target BP neural network to obtain a predicted leakage concentration; judge whether the predicted leakage concentration is greater than a preset early warning leakage concentration; if the predicted leakage concentration is not greater than the early warning leakage concentration, return to the step of monitoring the current hydrogen concentration;

[0172] The leakage intelligent early warning module 104 is used to, if the predicted leakage concentration is greater than the early warning leakage concentration, perform hydrogen leakage intelligent early warning.

[0173] In detail, the modules in the intelligent early warning system 100 based on the hydrogen leakage danger degree in industrial production in the embodiments of the present application adopt the same technical means as the intelligent early warning method based on the hydrogen leakage danger degree in industrial production in the above-mentioned Figure 1 and can produce the same technical effects, which will not be described here.

[0174] As shown in Figure 3 , it is a structural schematic diagram of an electronic device for implementing the intelligent early warning method based on the hydrogen leakage danger degree in industrial production according to an embodiment of the present application.

[0175] The electronic device 1 can include a processor 10, a memory 11 and a bus 12, and can further include a computer program stored in the memory 11 and executable on the processor 10, such as an intelligent early warning method program based on the hydrogen leakage danger degree in industrial production.

[0176] The memory 11 includes at least one type of readable storage medium, including a flash memory, a mobile hard disk, a multimedia card, a card-type memory (for example, an SD or DX memory, etc.), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, for example, a mobile hard disk of the electronic device 1. In other embodiments, the memory 11 can also be an external storage device of the electronic device 1, for example, a plug-in mobile hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Further, the memory 11 includes an internal storage unit of the electronic device 1 and also includes an external storage device. The memory 11 can be used not only to store application software and various data installed in the electronic device 1, such as the code of the intelligent early warning method program based on the hydrogen leakage danger degree in industrial production, but also to temporarily store data that has been output or will be output.

[0177] The processor 10 can be composed of integrated circuits in some embodiments, for example, can be composed of a single packaged integrated circuit, or can be composed of multiple packaged integrated circuits with the same function or different functions, including one or more combinations of central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 10 is the control core of the electronic device, which connects various components of the entire electronic device through various interfaces and lines, executes programs or modules stored in the memory 11 (such as the intelligent early warning method program based on the risk degree of hydrogen leakage in industrial production), and calls data stored in the memory 11 to perform various functions and process data of the electronic device 1.

[0178] The bus 12 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 can be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to realize the connection and communication between the memory 11, the at least one processor 10, etc.

[0179] Figure 3 Only the electronic device with components is shown, and those skilled in the art can understand that, Figure 3 The structure shown does not constitute a limitation on the electronic device 1, and can include fewer or more components than shown, or combine certain components, or different component arrangements.

[0180] For example, although not shown, the electronic device 1 can also include a power supply (such as a battery) for powering various components. Preferably, the power supply can be logically connected to the at least one processor 10 through a power management device, so as to realize functions such as charge management, discharge management, and power consumption management through the power management device. The power supply can also include one or more direct current or alternating current power supplies, recharging devices, power supply fault detection circuits, power supply converters or inverters, power supply status indicators, and any other components. The electronic device 1 can also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which are not described here.

[0181] Further, the electronic device 1 can also include a network interface, which can optionally include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), and is usually used to establish a communication connection between the electronic device 1 and other electronic devices.

[0182] Optionally, the electronic device 1 can also include a user interface, which can be a display, an input unit such as a keyboard, and optionally a standard wired interface, a wireless interface. Optionally, in some embodiments, the display can be an LED display, a liquid crystal display, a touch liquid crystal display, an OLED (Organic Light-Emitting Diode) touch, etc. Among them, the display can also be appropriately called a display screen or a display unit, which is used to display the information processed in the electronic device 1 and to display the visualized user interface.

[0183] The program of the intelligent early warning method based on the risk degree of hydrogen leakage in industrial production stored in the memory 11 in the electronic device 1 is a combination of multiple instructions, which can realize the following functions when running in the processor 10:

[0184] Extracting hydrogen leakage historical data in a pre-constructed hydrogen leakage historical data set, and extracting interval leakage concentration difference, target leakage concentration, leakage environment temperature, leakage environment humidity, and leakage environment pressure in the hydrogen leakage historical data according to a pre-set first leakage concentration, interval sampling duration, and leakage processing duration;

[0185] Training the pre-constructed initial BP neural network using the interval leakage concentration difference, leakage environment temperature, leakage environment humidity, leakage environment pressure, and target leakage concentration to obtain a target BP neural network, wherein the input layer of the initial BP neural network includes a leakage concentration difference node, an environment temperature node, an environment humidity node, and an environment pressure node, and the output layer includes a concentration prediction node;

[0186] Monitoring the current hydrogen concentration;

[0187] Determining whether the current hydrogen concentration is greater than the first leakage concentration;

[0188] If the current hydrogen concentration is not greater than the first leakage concentration, returning to the step of monitoring the current hydrogen concentration;

[0189] If the current hydrogen concentration is greater than the first leakage concentration, collecting a second leakage concentration according to the interval sampling duration, and calculating a current leakage concentration difference according to the first leakage concentration and the second leakage concentration, wherein the current leakage concentration difference refers to the difference between the second leakage concentration and the first leakage concentration;

[0190] Collecting the current environment temperature, the current environment humidity, and the current environment pressure, inputting the current leakage concentration difference, the current environment temperature, the current environment humidity, and the current environment pressure into the target BP neural network to obtain a predicted leakage concentration;

[0191] determining whether the predicted leakage concentration is greater than a preset early warning leakage concentration;

[0192] if the predicted leakage concentration is not greater than the early warning leakage concentration, returning to the step of monitoring the current hydrogen concentration;

[0193] if the predicted leakage concentration is greater than the early warning leakage concentration, performing intelligent hydrogen leakage warning.

[0194] Specifically, the specific implementation method of the processor 10 to the above instructions can refer to Figures 1 to 3 The description of related steps in the corresponding embodiments will not be repeated here.

[0195] Further, the modules / units integrated in the electronic device 1, if realized in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. The computer readable storage medium can be volatile or non-volatile. For example, the computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory).

[0196] The application also provides a computer readable storage medium, the readable storage medium stores a computer program, the computer program can realize the following when being executed by the processor of the electronic device:

[0197] extracting hydrogen leakage historical data in a pre-constructed hydrogen leakage historical data set in sequence, and extracting interval leakage concentration difference, target leakage concentration, leakage environment temperature, leakage environment humidity, and leakage environment air pressure in the hydrogen leakage historical data according to a preset first leakage concentration, interval sampling duration, and leakage processing duration;

[0198] training the pre-constructed initial BP neural network by using the interval leakage concentration difference, leakage environment temperature, leakage environment humidity, leakage environment air pressure, and target leakage concentration to obtain a target BP neural network, wherein the input layer of the initial BP neural network includes a leakage concentration difference node, an environment temperature node, an environment humidity node, and an environment air pressure node, and the output layer includes a concentration prediction node;

[0199] monitoring a current hydrogen concentration;

[0200] determining whether the current hydrogen concentration is greater than the first leakage concentration;

[0201] if the current hydrogen concentration is not greater than the first leakage concentration, returning to the step of monitoring the current hydrogen concentration;

[0202] If the current hydrogen concentration is greater than the first leakage concentration, a second leakage concentration is collected according to the interval sampling duration, and a current leakage concentration difference is calculated according to the first leakage concentration and the second leakage concentration, wherein the current leakage concentration difference refers to a difference between the second leakage concentration and the first leakage concentration;

[0203] The current environmental temperature, the current environmental humidity and the current environmental pressure are collected, and the current leakage concentration difference, the current environmental temperature, the current environmental humidity and the current environmental pressure are input into the target BP neural network to obtain a predicted leakage concentration;

[0204] It is judged whether the predicted leakage concentration is greater than a preset early warning leakage concentration;

[0205] If the predicted leakage concentration is not greater than the early warning leakage concentration, the step of monitoring the current hydrogen concentration is returned to;

[0206] If the predicted leakage concentration is greater than the early warning leakage concentration, hydrogen leakage intelligent early warning is performed.

[0207] In several embodiments provided in the present application, it should be understood that the disclosed devices, systems and methods can be implemented in other manners. For example, the system embodiments described above are merely illustrative, and actual implementation can have other division manners.

[0208] The modules described as separate components can or can not be physically separated, and the components displayed as modules can or can not be physical units, i.e., can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment scheme according to actual needs.

[0209] In addition, each functional module in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of hardware plus software functional modules.

[0210] It is obvious for those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and the present application can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application.

[0211] Finally, it should be noted that the above embodiments are merely used to illustrate the technical solutions of the present application but not limit the present application, and although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical solutions of the present application.

Claims

1. A smart early warning method based on the degree of hydrogen leakage hazard in industrial production, characterized in that, The method includes: Historical hydrogen leak data are extracted sequentially from a pre-constructed historical hydrogen leak dataset. Based on a preset first leak concentration, interval sampling time, and leak processing time, the interval leak concentration difference, target leak concentration, leak ambient temperature, leak ambient humidity, and leak ambient pressure are extracted from the historical hydrogen leak data. The pre-constructed initial BP neural network is trained using the interval leakage concentration difference, leakage ambient temperature, leakage ambient humidity, leakage ambient air pressure and target leakage concentration to obtain the target BP neural network. The input layer of the initial BP neural network includes: leakage concentration difference node, ambient temperature node, ambient humidity node and ambient air pressure node, and the output layer includes: concentration prediction node. Monitor the current hydrogen concentration; Determine whether the current hydrogen concentration is greater than the first leakage concentration; If the current hydrogen concentration is not greater than the first leakage concentration, then return to the above steps of monitoring the current hydrogen concentration; If the current hydrogen concentration is greater than the first leakage concentration, then the second leakage concentration is collected according to the interval sampling time, and the current leakage concentration difference is calculated based on the first leakage concentration and the second leakage concentration, wherein the current leakage concentration difference refers to the difference between the second leakage concentration and the first leakage concentration; The current ambient temperature, current ambient humidity, and current ambient air pressure are collected, and the current leakage concentration difference, current ambient temperature, current ambient humidity, and current ambient air pressure are input into the target BP neural network to obtain the predicted leakage concentration. Determine whether the predicted leakage concentration is greater than the preset warning leakage concentration; If the predicted leakage concentration is not greater than the warning leakage concentration, then return to the above steps of monitoring the current hydrogen concentration; If the predicted leakage concentration is greater than the warning leakage concentration, a smart warning for hydrogen leakage will be issued.

2. The intelligent early warning method based on the degree of hydrogen leakage hazard in industrial production as described in claim 1, characterized in that, The step of extracting the interval leakage concentration difference, target leakage concentration, leakage ambient temperature, leakage ambient humidity, and leakage ambient pressure from the hydrogen leakage historical data based on the preset first leakage concentration, interval sampling time, and leakage processing time includes: Identify the first historical leakage time of the first leakage concentration from the hydrogen leakage history data; The historical second leakage time is calculated based on the historical first leakage time and the interval sampling duration, wherein the interval sampling duration is equal to the time difference between the historical first leakage time and the historical second leakage time; Based on the historical second leak time, the historical second leak concentration is extracted from the historical hydrogen leak data; The interval leakage concentration difference is calculated based on the first leakage concentration and the historical second leakage concentration, wherein the interval leakage concentration difference is equal to the difference between the historical second leakage concentration and the first leakage concentration; Based on the historical second leak time and leak handling duration, the target leak concentration is extracted from the historical hydrogen leak data. Extract the ambient temperature, ambient humidity, and ambient pressure of the second leakage time from the historical hydrogen leakage data.

3. The intelligent early warning method based on the degree of hydrogen leakage hazard in industrial production as described in claim 2, characterized in that, Before extracting the interval leakage concentration difference, target leakage concentration, leakage ambient temperature, leakage ambient humidity, and leakage ambient pressure from the hydrogen leakage historical data based on the preset first leakage concentration, interval sampling time, and leakage processing time, the method further includes: The test interval duration is extracted sequentially from the preset interval duration gradient sequence; Extract historical hydrogen leak data sequentially from the aforementioned historical hydrogen leak dataset; The warning leakage time is identified in the historical hydrogen leakage data based on the warning leakage concentration, wherein the warning leakage time refers to the leakage time corresponding to the first appearance of the warning leakage concentration in the historical hydrogen leakage data; Based on the leakage handling time and the preset safety reserve time, the cutoff sampling time is calculated using the early warning leakage time, wherein the time difference between the early warning leakage time and the cutoff sampling time is equal to the sum of the leakage handling time and the safety reserve time; The sampling value range is determined based on the historical first leakage time and the sampling cutoff time. Based on the test interval and the historical first leakage time, the starting sampling time is calculated using the following formula: T start =T1+t,T start ∈(T1,T end ) Among them, T start T1 represents the initial sampling time, T1 represents the first historical leak time, and t represents the test interval duration. end ) represents the sampling range, T end Indicates the sampling deadline; Based on the first leakage concentration and the initial sampling time, the interval test concentration difference is calculated using the historical hydrogen leakage data. The target test time is calculated based on the initial sampling time and the leakage handling duration, wherein the difference between the target test time and the initial sampling time is equal to the leakage handling duration; The target test concentration is extracted from the historical hydrogen leak data based on the target test time. The pre-constructed initial BP neural network is trained based on the interval test concentration difference, leakage environment temperature, leakage environment humidity, leakage environment air pressure and target test concentration to obtain an iterative BP neural network; Determine whether the iterative BP neural network has been trained. If the iterative BP neural network has not been trained, return to the steps described above for sequentially extracting historical hydrogen leak data from the historical hydrogen leak dataset. If the iterative BP neural network training is completed, the BP neural network to be verified is obtained; The prediction accuracy of the BP neural network to be verified is obtained by verifying the first leakage concentration, the test interval duration, and the leakage handling duration. The prediction accuracy corresponding to each test interval is collected to obtain a prediction accuracy set, and the maximum prediction accuracy is extracted from the prediction accuracy set. Identify the target interval duration corresponding to the maximum prediction accuracy, and use the target interval duration as the interval sampling duration.

4. The intelligent early warning method based on the degree of hydrogen leakage hazard in industrial production as described in claim 2, characterized in that, The step of extracting the target leakage concentration from the historical hydrogen leakage data based on the historical second leakage time and leakage handling duration includes: The target leakage time is calculated based on the historical second leakage time and the leakage handling time, wherein the leakage handling time is equal to the difference between the target leakage time and the historical second leakage time; The target leakage concentration is extracted from the historical hydrogen leakage data based on the target leakage time.

5. The intelligent early warning method based on the degree of hydrogen leakage hazard in industrial production as described in claim 1, characterized in that, The process of training a pre-constructed initial BP neural network using the interval leakage concentration difference, leakage ambient temperature, leakage ambient humidity, leakage ambient air pressure, and target leakage concentration to obtain a target BP neural network includes: The interval leakage concentration difference, leakage ambient temperature, leakage ambient humidity and leakage ambient air pressure are respectively input into the leakage concentration difference node, ambient temperature node, ambient humidity node and ambient air pressure node of the initial BP neural network to obtain the iteratively predicted leakage concentration. The prediction error is calculated based on the iteratively predicted leakage concentration and the target leakage concentration. Determine whether the prediction error is less than a preset error threshold; If the prediction error is not less than the error threshold, the initial BP neural network is adjusted according to the prediction error, and the above steps of sequentially extracting historical hydrogen leakage data from the pre-constructed historical hydrogen leakage dataset are returned. If the prediction error is less than the error threshold, the target BP neural network is obtained.

6. The intelligent early warning method based on the degree of hydrogen leakage hazard in industrial production as described in claim 1, characterized in that, The step of collecting the second leakage concentration according to the specified sampling interval includes: Obtain the current time, and calculate the second sampling time based on the current time and the interval sampling duration; The second leakage concentration was collected based on the second sampling time.

7. The intelligent early warning method based on the degree of hydrogen leakage hazard in industrial production as described in claim 3, characterized in that, The step of calculating the interval test concentration difference based on the first leakage concentration and the initial sampling time using the historical hydrogen leakage data includes: The initial sampling concentration is extracted from the hydrogen leak history data based on the initial sampling time. The interval test concentration difference is calculated based on the initial sampling concentration and the first leakage concentration, wherein the interval test concentration difference is equal to the difference between the initial sampling concentration and the first leakage concentration.

8. The intelligent early warning method based on the degree of hydrogen leakage hazard in industrial production as described in claim 3, characterized in that, The step of verifying the prediction accuracy of the BP neural network to be tested based on the first leakage concentration, the test interval duration, and the leakage handling duration, to obtain the prediction accuracy, includes: Randomly extract batches of hydrogen leak verification data from the historical hydrogen leak dataset, wherein the number of hydrogen leak verification data in the batches of hydrogen leak verification data is equal to the preset batch number; Based on the first leakage concentration, test interval duration, and leakage handling duration, a verification dataset is extracted from the batch of hydrogen leakage verification data. The verification data in the verification dataset includes: interval verification concentration difference, target verification concentration, verification ambient temperature, verification ambient humidity, and verification ambient pressure. The prediction accuracy of the BP neural network to be verified is obtained by using the verification dataset.

9. The intelligent early warning method based on the degree of hydrogen leakage hazard in industrial production as described in claim 8, characterized in that, The step of using the verification dataset to verify the prediction accuracy of the BP neural network to be verified, and obtaining the prediction accuracy, includes: Extract verification data sequentially from the verification dataset; The interval verification concentration difference, verification ambient temperature, verification ambient humidity, and verification ambient air pressure in the verification data are input into the BP neural network to be verified to obtain the predicted concentration to be verified. Calculate the reciprocal of the predicted difference between the predicted concentration to be verified and the target verification concentration to obtain a set of reciprocal prediction differences, wherein the reciprocal prediction difference refers to the reciprocal of the difference between the predicted concentration to be verified and the target verification concentration; Based on the inverse set of the predicted differences, the prediction accuracy is calculated using the following formula: Where z represents the prediction accuracy, c1 represents the first reciprocal of the prediction difference in the set of reciprocal prediction differences, c2 represents the second reciprocal of the prediction difference in the set of reciprocal prediction differences, and c m This represents the m-th reciprocal of the prediction difference in the set of prediction difference reciprocals.

10. An intelligent early warning system based on the degree of hydrogen leakage hazard in industrial production, characterized in that, The system includes: A neural network training module is used to sequentially extract historical hydrogen leak data from a pre-constructed historical hydrogen leak dataset. Based on a preset first leak concentration, interval sampling time, and leak processing time, it extracts interval leak concentration difference, target leak concentration, leak ambient temperature, leak ambient humidity, and leak ambient pressure from the historical hydrogen leak data. The pre-constructed initial BP neural network is trained using the interval leak concentration difference, leak ambient temperature, leak ambient humidity, leak ambient pressure, and target leak concentration to obtain a target BP neural network. The input layer of the initial BP neural network includes nodes for leak concentration difference, ambient temperature, ambient humidity, and ambient pressure, and the output layer includes a concentration prediction node. The hydrogen concentration determination module is used to monitor the current hydrogen concentration and determine whether the current hydrogen concentration is greater than the first leakage concentration; if the current hydrogen concentration is not greater than the first leakage concentration, the process returns to the above steps of monitoring the current hydrogen concentration. The leakage concentration comparison module is used to: if the current hydrogen concentration is greater than the first leakage concentration, collect a second leakage concentration according to the interval sampling time; calculate the current leakage concentration difference based on the first and second leakage concentrations, wherein the current leakage concentration difference refers to the difference between the second and first leakage concentrations; collect the current ambient temperature, current ambient humidity, and current ambient air pressure; input the current leakage concentration difference, current ambient temperature, current ambient humidity, and current ambient air pressure into the target BP neural network to obtain a predicted leakage concentration; determine whether the predicted leakage concentration is greater than a preset warning leakage concentration; if the predicted leakage concentration is not greater than the warning leakage concentration, return to the above steps of monitoring the current hydrogen concentration; The intelligent leak warning module is used to issue an intelligent warning for hydrogen leaks if the predicted leak concentration is greater than the warning leak concentration.

Images