An intelligent safety monitoring and early warning system and method for hydrogen energy processing equipment

Abstract

The present invention discloses an intelligent safety monitoring and early warning system and method for hydrogen energy processing equipment, relating to the field of hydrogen energy safety technology, and is used to solve the problem of imperfect safety monitoring technology for current hydrogen energy processing equipment. The method installs a variety of sensors at key parts of the equipment to collect the operation data of the hydrogen energy processing equipment, and uses data processing technologies such as Kalman filtering to ensure the accuracy and timeliness of the data. The spectral characteristics of the equipment operation data are extracted through fast Fourier transform, and the time-domain characteristics of the data are analyzed in combination with a sliding time window to further optimize the characteristic data. The analysis of the hydrogen concentration change rate and diffusion rate is introduced to detect the fluctuation of the hydrogen concentration in real time. Cluster analysis and the deviation degree of characteristic data are used to detect abnormal conditions of the hydrogen energy processing equipment and compensate for abnormal data. Finally, by predicting future trends, a scientific basis is provided for equipment operation and maintenance to ensure the safe and efficient operation of the equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of hydrogen energy safety, and more specifically, the present invention relates to an intelligent safety monitoring and early warning system and method for hydrogen energy processing equipment. Background Art

[0002] As a clean energy, hydrogen energy has gradually become an important direction for global energy transformation. Hydrogen energy processing equipment plays a key role in the energy conversion and storage process, and its safety is directly related to production efficiency and personnel safety. However, due to the high flammability of hydrogen and the complexity of the equipment operating environment, different types of failures may occur during the long-term operation of hydrogen energy processing equipment, such as equipment aging, sensor failure, hydrogen leakage, etc. These failures may lead to serious safety accidents. Therefore, real-time monitoring and fault early warning of hydrogen energy processing equipment have become urgent problems to be solved.

[0003] Currently, traditional safety monitoring methods for hydrogen energy processing equipment mostly rely on manual inspections or simple alarm devices, and cannot achieve intelligent and real-time fault early warning. This method not only has a lag in response, but also cannot comprehensively monitor the operating status of the equipment, and it is difficult to detect potential safety hazards in a timely manner. At the same time, with the progress of intelligent technology, equipment monitoring systems based on big data and artificial intelligence have become an important direction to improve equipment safety. Through the real-time collection and analysis of sensor data, and the application of machine learning algorithms, equipment failures can be predicted more accurately, early warnings can be given, and a scientific basis can be provided for equipment management and maintenance.

[0004] In view of the above problems, the present invention proposes a solution. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides an intelligent safety monitoring and early warning system and method for hydrogen energy processing equipment to solve the problems raised in the above background art.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] In a preferred embodiment, it includes:

[0008] Step 1: Monitor and collect the operating data and hydrogen concentration data of the hydrogen energy processing equipment;

[0009] Step 2: Extract the characteristic data of the collected operating data and hydrogen concentration data of the hydrogen energy processing equipment and optimize the characteristic data in combination with abnormal hydrogen concentration conditions;

[0010] Step 3: Detect abnormal conditions of the hydrogen energy processing equipment based on clustering analysis technology combined with the characteristic data deviation degree;

[0011] Step 4: Classification Alarm for Abnormal Points and Prediction of Future Trends.

[0012] In a preferred embodiment, in Step 1, a variety of sensors are installed on the hydrogen energy processing equipment to collect the operation data and hydrogen concentration data of the hydrogen energy processing equipment, and perform standardization and normalization processing.

[0013] In a preferred embodiment, in Step 2, the characteristic data of the operation data and hydrogen concentration data of the hydrogen energy processing equipment are extracted, and the characteristic data is dimensionally reduced; the hydrogen concentration change rate Rh and the hydrogen diffusion rate Dh are calculated, and the weighted sum of the hydrogen concentration change rate Rh and the hydrogen diffusion rate Dh is used to determine the current hydrogen concentration change coefficient Qs. A hydrogen concentration change threshold Yq is set. When the hydrogen concentration change coefficient Qs is greater than or equal to the hydrogen concentration change threshold Yq, it indicates that the hydrogen concentration data of the hydrogen energy processing equipment fluctuates violently at this time.

[0014] In a preferred embodiment, in Step 2, the error and change rate of the operation data of the hydrogen energy processing equipment before and after the abnormality are calculated, and the characteristic data of the operation data of the hydrogen energy processing equipment after dimensional reduction is adjusted and compensated through the error and change rate of the operation data of the hydrogen energy processing equipment before and after the abnormality; and the characteristic data of the hydrogen concentration data after dimensional reduction is adjusted through the characteristic data of the operation data of the hydrogen energy processing equipment after compensation.

[0015] In a preferred embodiment, in Step 3, the Euclidean distance from the data point to the cluster center is calculated based on the clustering analysis technique. When the Euclidean distance from the data point to the cluster center is greater than or equal to the safety distance threshold Ya, the data point is marked as an abnormal data point.

[0016] In a preferred embodiment, in Step 3, the main body condition data of each main body of the hydrogen energy processing equipment is collected, and the characteristic data deviation degree ΔPy of each main body is calculated; when the characteristic data deviation degree ΔPy of the main body exceeds the deviation degree threshold Ypy, it indicates that the abnormal data point is an abnormal point, and the number of abnormal points under each main body is calculated and the abnormal point type is divided.

[0017] In a preferred embodiment, in Step 4, the abnormal point deviation value of each main body is calculated, and the weighted sum of the abnormal point deviation value of each main body and the number of abnormal points under each main body is used to determine the abnormal priority of each main body.

[0018] In a preferred embodiment, the method for obtaining the deviation degree threshold Ypy is as follows:

[0019] Calculate the mean value Pj and standard deviation Pb of the characteristic data deviation degree in the historical data; use the anomaly detection method based on the mean value and standard deviation to analyze the mean value Pj and standard deviation Pb of the characteristic data deviation degree, and determine the deviation degree threshold Ypy.

[0020] In a preferred embodiment, the method for obtaining the safety distance threshold Ya is as follows:

[0021] Calculate the average distance davg from all data points in the cluster to the center; calculate the standard deviation of all data points in the cluster to the center ; Combine the average distance davg and the standard deviation According to 3 Set the safety distance threshold Ya according to the rule.

[0022] In a preferred embodiment, it includes: a data acquisition module, a feature data extraction module, an anomaly detection module, a future prediction module, and a data storage module, with signal connections between the modules;

[0023] The data acquisition module is mainly used to collect the operation data and hydrogen concentration data of the hydrogen energy processing equipment:

[0024] The feature data extraction module is mainly used to extract the feature data of the operation data and hydrogen concentration data of the hydrogen energy processing equipment, and further optimize the feature data in combination with the change of hydrogen concentration:

[0025] The anomaly detection module is mainly used to detect the anomalies of the hydrogen energy processing equipment based on the clustering analysis technology combined with the feature data deviation degree;

[0026] The future prediction module mainly predicts the future operation status and anomaly point trend of the hydrogen energy processing equipment based on historical data and existing data;

[0027] The data storage module is mainly used to store all data during the processing.

[0028] The technical effects and advantages of an intelligent safety monitoring and warning system and method for hydrogen energy processing equipment of the present invention:

[0029] The present invention collects the operation status and hydrogen concentration data of the equipment through multi-sensor fusion, and realizes the in-depth mining of multi-dimensional data by combining time-frequency domain feature extraction and dynamic threshold analysis. Based on the clustering algorithm, abnormal data points are identified, a feature deviation degree evaluation mechanism is introduced synchronously, the sensitivity of anomaly detection is dynamically optimized, and intelligent classification of anomaly types is realized according to the operation characteristics of the equipment main body. The system adopts a hierarchical warning strategy, combines real-time status and historical trends to predict potential risks, and provides dynamic decision-making support for equipment maintenance. This method breaks through the limitations of traditional static threshold monitoring, significantly improves the real-time response ability and warning accuracy of safety hazards such as hydrogen leakage and equipment failures, and is applicable to the safety management of hydrogen energy equipment under complex working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a schematic structural diagram of an intelligent safety monitoring and warning system for hydrogen energy processing equipment of the present invention.

[0031] Figure 2 This is the operation flow chart of an intelligent safety monitoring and warning method for a hydrogen energy processing device of the present invention. Specific implementation manners

[0032] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0033] Embodiment

[0034] The present invention discloses an intelligent safety monitoring and warning method for a hydrogen energy processing device, as Figure 2 shown, including:

[0035] Step 1: Monitor and collect the operation data and hydrogen concentration data of the hydrogen energy processing device;

[0036] Adopt a distributed sensor network, install a variety of sensors on the hydrogen energy processing device body, key components and connecting pipelines, collect the operation data such as the pressure, temperature, vibration frequency and flow rate data of the hydrogen energy processing device and the hydrogen concentration data. At the same time, set the data sampling frequency fs and the number of sampling points N1 of each sensor to ensure that the data has sufficient time series resolution;

[0037] Specific sensor types and installation and deployment locations:

[0038] Install a pressure sensor at the gas storage tank and pipeline interface of the hydrogen energy processing device to detect and obtain the pressure data of the hydrogen energy processing device;

[0039] Install a temperature sensor at the heating unit and cooling system of the hydrogen energy processing device to detect and obtain the temperature data of the hydrogen energy processing device;

[0040] Install a vibration sensor at the hydrogen compressor and pump body of the hydrogen energy processing device to detect and obtain the vibration frequency data of the hydrogen energy processing device;

[0041] Install a flow sensor in the pipeline of the hydrogen energy processing device to detect and obtain the hydrogen flow rate data in the pipeline of the hydrogen energy processing device;

[0042] Install a hydrogen concentration sensor around the hydrogen energy processing device and in the key sealing areas to detect and obtain the hydrogen concentration data of the hydrogen energy processing device.

[0043] After that, the Kalman filter is used to process the operation data of the hydrogen energy processing equipment and the hydrogen concentration data collected by the various sensors, and the state estimation algorithm is used to reduce random noise; for impact data such as vibration frequency, median filtering is used to remove instantaneous impact interference; the operation data of the hydrogen energy processing equipment and the hydrogen concentration data are standardized by Z-score standardization, and Min-Max normalization is used to map the data to between [0, 1]:

[0044] Step 2: Extract the characteristic data of the operation data of the hydrogen energy processing equipment and the hydrogen concentration data collected, and optimize the characteristic data in combination with the abnormal hydrogen concentration situation;

[0045] Extract the characteristic data of the operation data of the hydrogen energy processing equipment and the hydrogen concentration data:

[0046] The fast Fourier transform (FFT) is used to calculate the spectral characteristics of the operation data of the hydrogen energy processing equipment and the hydrogen concentration data collected in Step 1, according to the formula:

[0047]

[0048] where, X(k) represents the spectral data, and x(n) represents the original data;

[0049] After that, calculate the spectral energy distribution in the operation data of the hydrogen energy processing equipment and the hydrogen concentration data, and extract the main frequency component and its amplitude change;

[0050] Set the sliding time window T, and extract the time-domain characteristics of the operation data of the hydrogen energy processing equipment and the hydrogen concentration data within the window, specifically as follows:

[0051] Calculate the mean value Tz1 of the operation data of the hydrogen energy processing equipment and the hydrogen concentration data within the sliding time window T to reflect the overall trend of the operation data of the hydrogen energy processing equipment and the hydrogen concentration data:

[0052]

[0053] Calculate the variance Tz2 of the operation data of the hydrogen energy processing equipment and the hydrogen concentration data within the sliding time window T to measure the volatility of the operation data of the hydrogen energy processing equipment and the hydrogen concentration data:

[0054]

[0055] Calculate the skewness Tz3 of the operation data of the hydrogen energy processing equipment and the hydrogen concentration data within the sliding time window T to describe the symmetry of the distribution of the operation data of the hydrogen energy processing equipment and the hydrogen concentration data:

[0056]

[0057] Calculate the kurtosis Tz4 of the operation data and hydrogen concentration data of the hydrogen energy processing equipment within the sliding time window T to measure the sharpness of the distribution of the operation data and hydrogen concentration data of the hydrogen energy processing equipment:

[0058]

[0059] Furthermore, form the feature data of the operation data and hydrogen concentration data of the hydrogen energy processing equipment into a matrix X, and define each row of the matrix as a sample and each column as a feature. Calculate the covariance matrix C of the feature matrix X, specifically according to the formula:

[0060]

[0061] where m represents the number of samples and Xt represents the transpose of matrix X;

[0062] After that, calculate the eigenvalues λi of the covariance matrix C and their corresponding eigenvectors vi:

[0063]

[0064] Sort the eigenvalues and select the principal components whose cumulative contribution rate exceeds 95%:

[0065]

[0066] where k represents the number of selected principal components and n represents the number of original features;

[0067] Select the first k eigenvectors to form the projection matrix W: W = [v1, v2,..., vk], and map the original feature matrix X to the new low-dimensional feature space: X′ = XW, where X′ represents the dimension-reduced feature data.

[0068] Considering that when the hydrogen concentration of the hydrogen energy processing equipment reaches a certain level, it will affect the operation data of the hydrogen energy processing equipment. Therefore, after each extraction of the feature data of the operation data and hydrogen concentration data of the hydrogen energy processing equipment, set the sliding time window T1 and calculate the hydrogen concentration change rate Rh within the sliding time window T1, specifically according to the formula:

[0069]

[0070] where Ct represents the hydrogen concentration at the current time, Ct-1 represents the hydrogen concentration at the previous time, and S represents the time interval of the sliding window; then, calculate the hydrogen diffusion rate Dh within the sliding time window T1, specifically according to the formula:

[0071]

[0072] Among them, k1 represents the diffusion coefficient, ΔC represents the concentration gradient, and Δx represents the diffusion distance;

[0073] Furthermore, the weighted sum of the hydrogen concentration change rate Rh and the hydrogen diffusion rate Dh is used to determine the current hydrogen concentration change coefficient Qs, and a hydrogen concentration change threshold Yq is set. The hydrogen concentration change coefficient Qs is compared with the hydrogen concentration change threshold Yq. When the hydrogen concentration change coefficient Qs is greater than or equal to the hydrogen concentration change threshold Yq, it indicates that the hydrogen concentration data fluctuates violently when the hydrogen energy processing equipment runs to the time window T1;

[0074] After detecting that the hydrogen concentration of the hydrogen energy processing equipment reaches the threshold and fluctuates violently, the operation data of the hydrogen energy processing equipment is recollected through the arranged sensors, and the error and change rate of the operation data of the hydrogen energy processing equipment before and after the abnormality are calculated. Specifically:

[0075] Calculation of the error Eerr of the operation data of the hydrogen energy processing equipment before and after the abnormality:

[0076] Eerr = |Xpost - Xpre|

[0077] Among them, Xpost represents the operation data of the hydrogen energy processing equipment before the abnormality, and Xpre represents the operation data of the hydrogen energy processing equipment after the abnormality;

[0078] Calculation of the change rate Rchange of the operation data of the hydrogen energy processing equipment before and after the abnormality:

[0079] Rchange = (Xpost - Xpre) / Xpre

[0080] Based on the error Eerr and the change rate Rchange of the operation data of the hydrogen energy processing equipment before and after the abnormality, the weighted sum algorithm is used to adjust and compensate the characteristic data of the operation data of the hydrogen energy processing equipment after dimensionality reduction: Xcomp = Xp + γ × Eerr × Rchange, where γ represents the compensation coefficient, Xp represents the characteristic data of the operation data of the hydrogen energy processing equipment before the abnormality, and Xcomp represents the characteristic data of the operation data of the hydrogen energy processing equipment after compensation after the abnormality.

[0081] After that, the current hydrogen concentration is adjusted through the characteristic data Xcomp of the operation data of the hydrogen energy processing equipment after compensation: Cadj = λ × Xcomp + (1 - λ) × Ccurrent, where Cadj represents the characteristic data of the adjusted hydrogen concentration data, Ccurrent represents the characteristic data of the hydrogen concentration data before the abnormality, and λ represents the adjustment coefficient, and its value range is [0, 1].

[0082] Step 3: Detect the abnormal situation of the hydrogen energy processing equipment based on the clustering analysis technology combined with the feature data deviation degree;

[0083] Based on the feature data extracted in step 2, the optimal number of clusters k2 is determined using the elbow method based on K-Means clustering:

[0084] Calculate the within-cluster sum of squares (WSS) for different values of k2:

[0085]

[0086] Among them, Ci represents the i-th cluster, μi is the cluster center. Then, draw the elbow method curve, find the obvious inflection point of the within-cluster sum of squares (WSS), and determine the value of the optimal number of clusters k;

[0087] Furthermore, use the K-Means++ initialization strategy to randomly initialize k2 random clustering centers, and calculate the Euclidean distance from each data point to the clustering center:

[0088]

[0089] Classify the data points into the nearest cluster according to the minimum distance. Furthermore, calculate the new mean of each cluster:

[0090]

[0091] Update the clustering centers and iterate until the clustering centers no longer change or reach the maximum number of iterations.

[0092] After that, calculate the Euclidean distance from the data points to the clustering center, denoted as D1, and calculate the average distance davg from all data points in the cluster to the center:

[0093]

[0094] Calculate the standard deviation of all data points in the cluster to the center :

[0095]

[0096] And according to 3 rules, set the safety distance threshold Ya. When the Euclidean distance from the data point to the clustering center is greater than or equal to the safety distance threshold Ya, mark the data point as an abnormal data point;

[0097] Furthermore, deploy sensors on each main body of the hydrogen energy processing equipment to collect the data of the main body status. For example, collect pressure data from the hydrogen storage tank and the pipeline; collect temperature data from the booster and the cooling device, collect vibration frequency data from the hydrogen compressor; collect hydrogen flow data of each pipeline;

[0098] Based on the historical operation data of each main body, calculate the feature data deviation degree ΔPy of each main body:

[0099]

[0100] Among them, P represents the characteristic data of each current subject in real time, and Pbenchmark represents the mean value of the characteristic data of each subject; and the mean value Pj and the standard deviation Pb of the characteristic data deviation degree in the historical data are calculated, and an anomaly detection method based on the mean value and the standard deviation is used to analyze the mean value Pj and the standard deviation Pb of the characteristic data deviation degree to determine the deviation threshold Ypy; when the characteristic data deviation degree ΔPy of the subject exceeds the deviation threshold Ypy, it indicates that this non-conventional data point is an anomaly point, and combined with the different operating characteristics of different subjects, the anomaly point types are classified. For example, the anomaly points in the pipeline are leakage anomaly points or blockage anomaly points.

[0101] Furthermore, calculate the number of anomaly points under each subject:

[0102]

[0103] Among them, Nanomaly represents the number of anomaly points under this subject, 1(ΔPyi>Ypy) means that if ΔPy exceeds Ypy, it is recorded as an anomaly, and n represents the number of devices;

[0104] Step 4: Anomaly point grading and warning and future trend prediction;

[0105] According to the characteristic data deviation degree ΔPy of each subject obtained in Step 3, use the outlier detection method Z-score to calculate the anomaly point deviation value of each subject, and determine the anomaly priority of each subject by weighted summing the anomaly point deviation values of each subject and the number of anomaly points under each subject. Use MQTT / Kafka to push the anomaly data in descending order from the subject with the largest anomaly priority to notify the maintenance personnel.

[0106] At the same time, set the historical data window length Td and the prediction time Th, collect the data of the nearest N time windows: X = [Py(t−N), Py(t−N+1),..., Py(t)], and perform normalization processing; construct an LSTM model, and input the data of N time windows into the LSTM model to predict the anomaly points within the future Th time.

[0107] The present invention also proposes an intelligent safety monitoring and warning system for hydrogen energy processing equipment, as Figure 1 shown, including: a data acquisition module, a characteristic data extraction module, an anomaly detection module, a future prediction module, and a data storage module, and the modules are signal-connected to each other;

[0108] The data acquisition module is mainly used to collect the operation data and hydrogen concentration data of the hydrogen energy processing equipment:

[0109] The feature data extraction module is mainly used to extract the feature data of the operation data and hydrogen concentration data of the hydrogen energy processing equipment, and further optimize the feature data in combination with the change of hydrogen concentration:

[0110] The anomaly detection module is mainly used to detect the anomalies of the hydrogen energy processing equipment based on the clustering analysis technology combined with the feature data deviation degree;

[0111] The future prediction module mainly predicts the future operation state and anomaly point trend of the hydrogen energy processing equipment based on historical data and existing data;

[0112] The data storage module is mainly used to store all data during the processing.

[0113] The above formulas are all dimensionless and take their numerical values for calculation. The formula is a formula obtained by collecting a large amount of data for software simulation to get the closest to the real situation. The preset parameters in the formula are set by those skilled in the art according to the actual situation.

[0114] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product.

[0115] Those of ordinary skill in the art can realize that the modules and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application of the technical solution and the invention constraints. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0116] In addition, in each embodiment of this application, the functional modules can be integrated in a processing module, or each module can exist physically alone, or two or more modules can be integrated in one module.

[0117] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0118] Finally: The above is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An intelligent safety monitoring and early warning method for hydrogen energy processing equipment It is characterized in that it includes: Step 1: Monitor and collect the operation data of the hydrogen energy processing equipment and the hydrogen concentration data; Step 2: Extract the characteristic data of the collected operation data of the hydrogen energy processing equipment and the hydrogen concentration data, and optimize the characteristic data in combination with the abnormal hydrogen concentration situation; Step 3: Detect the abnormal situation of the hydrogen energy processing equipment based on the clustering analysis technology combined with the characteristic data deviation degree; Step 4: Abnormal point classification warning and future trend prediction; In Step 3, based on the clustering analysis technology, calculate the Euclidean distance from the data point to the cluster center. When the Euclidean distance from the data point to the cluster center is greater than or equal to the safety distance threshold Ya, mark this data point as an abnormal data point; In Step 3, collect the main body status data of each main body of the hydrogen energy processing equipment, and calculate the characteristic data deviation degree ΔPy of each main body; when the characteristic data deviation degree ΔPy of the main body exceeds the deviation degree threshold Ypy, it indicates that this abnormal data point is an abnormal point, calculate the number of abnormal points under each main body and classify the abnormal point types; The method for obtaining the deviation degree threshold Ypy is as follows: Calculate the mean value Pj and the standard deviation Pb of the characteristic data deviation degree in the historical data; use the abnormal detection method based on the mean value and the standard deviation to analyze the mean value Pj and the standard deviation Pb of the characteristic data deviation degree, and determine the deviation degree threshold Ypy.

2. The intelligent safety monitoring and warning method for a hydrogen energy processing device according to claim 1, characterized in that: In Step 1, install a variety of sensors on the hydrogen energy processing equipment, collect the operation data and the hydrogen concentration data of the hydrogen energy processing equipment, and perform standardization and normalization processing.

3. An intelligent safety monitoring and early warning method for a hydrogen energy processing device according to claim 2, characterized in that: In Step 2, extract the characteristic data of the operation data and the hydrogen concentration data of the hydrogen energy processing equipment, and perform dimensionality reduction processing on the characteristic data; calculate the hydrogen concentration change rate Rh and the hydrogen diffusion rate Dh, and determine the current hydrogen concentration change coefficient Qs by weighted summing the hydrogen concentration change rate Rh and the hydrogen diffusion rate Dh. Set the hydrogen concentration change threshold Yq. When the hydrogen concentration change coefficient Qs is greater than or equal to the hydrogen concentration change threshold Yq, it indicates that the hydrogen concentration data of the hydrogen energy processing equipment fluctuates violently at this time.

4. An intelligent safety monitoring and early warning method for a hydrogen energy processing device according to claim 3, characterized in that ; In Step 2, calculate the error and the change rate of the operation data of the hydrogen energy processing equipment before and after the abnormality, and adjust and compensate the characteristic data of the operation data of the hydrogen energy processing equipment after dimensionality reduction through the error and the change rate of the operation data of the hydrogen energy processing equipment before and after the abnormality; and adjust the characteristic data of the hydrogen concentration data after dimensionality reduction through the characteristic data of the operation data of the hydrogen energy processing equipment after compensation.

5. An intelligent safety monitoring and warning method for a hydrogen energy processing device according to claim 1, characterized in that: In Step 4, calculate the abnormal point deviation value of each main body, and determine the abnormal priority of each main body by weighted summing the abnormal point deviation value of each main body and the number of abnormal points under each main body.

6. The intelligent safety monitoring and warning method for a hydrogen energy processing device according to claim 1, characterized in that: The method for obtaining the safety distance threshold Ya is as follows: Based on the characteristic data extracted in Step 2, use the elbow method based on K-Means clustering to determine the optimal number of clusters k2; Calculate the average distance davg from all data points within the optimal number of clusters k2 to the center; calculate the standard deviation of all data points within the optimal number of clusters k2 to the center ; Combine the average distance davg and the standard deviation According to 3 Set the safety distance threshold Ya according to the rule.

7. An intelligent safety monitoring and warning system for hydrogen energy processing equipment, which is used to implement an intelligent safety monitoring and warning method for hydrogen energy processing equipment described in any one of the above claims 1-6. It is characterized in that; It includes: a data acquisition module, a characteristic data extraction module, an abnormal detection module, a future prediction module, and a data storage module, and the signals of each module are connected; The data acquisition module is mainly used to collect the operation data and the hydrogen concentration data of the hydrogen energy processing equipment: The feature data extraction module is mainly used to extract the feature data of the operation data and hydrogen concentration data of the hydrogen energy processing equipment, and further optimize the feature data in combination with the change of hydrogen concentration: The anomaly detection module is mainly used to detect the anomalies of the hydrogen energy processing equipment based on the clustering analysis technology combined with the feature data deviation degree; The future prediction module mainly predicts the future operation status and anomaly point trend of the hydrogen energy processing equipment based on historical data and existing data; The data storage module is mainly used to store all data during the processing process.

Images