Safety monitoring method and device for chemical enterprise, storage medium and electronic equipment
By combining recurrent neural networks based on attention mechanisms and Bayesian networks, the safety monitoring system of chemical companies can quickly and accurately obtain safety monitoring results, solving the problem that traditional systems are difficult to adapt to complex operating conditions and lack of comprehensive evaluation capabilities, achieving more comprehensive and accurate safety assessments, supporting real-time decision-making and preventive maintenance.
Patent Information
- Application Number
- CN202411817693.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-11
- Publication Date
- 2025-03-18
AI Technical Summary
Traditional chemical enterprise safety monitoring systems cannot obtain safety monitoring results quickly and accurately, and are difficult to adapt to complex and changeable operating conditions, are prone to false alarms or missed alarms, and lack the ability to comprehensively evaluate the long-term health status of the equipment, making it difficult to identify potential risks in advance.
By obtaining the equipment maintenance status data and safety monitoring timing data of the target equipment of the chemical enterprise, it is input into the pre-trained circular training neural network based on attention mechanism to obtain the monitoring data prediction results; at the same time, the equipment maintenance status data and safety monitoring data are input into the pre-constructed Bayesian network to obtain the equipment failure probability; then, based on the prediction results and failure probability, the safety evaluation results of the chemical enterprise are calculated.
It achieves faster and more accurate acquisition of safety monitoring results, enables more comprehensive and accurate assessment of the safety status of equipment, identify potential risks in advance, supports real-time decision-making and preventive maintenance, and reduces unnecessary downtime and repair costs.
Smart Images
Figure CN119646520A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data processing technology, and in particular to a safety monitoring method, device, storage medium and electronic equipment for a chemical enterprise. Background Art
[0002] Chemical companies involve a variety of complex chemical reactions and physical processes in the production process, accompanied by high-risk factors such as high temperature, high pressure, and hazardous substances. Therefore, ensuring the safe operation of production equipment is crucial to preventing accidents and protecting the lives and property of personnel. Traditionally, the safety management of chemical companies relies on regular maintenance, manual inspections, and simple threshold alarm systems. However, with the improvement of industrial automation and technological advancement, traditional management methods have been unable to meet the growing safety needs.
[0003] At present, traditional rule-based monitoring systems usually set fixed thresholds to trigger alarms, but they cannot adapt to complex and changing operating conditions and are prone to false alarms or missed alarms. They lack the ability to comprehensively evaluate the long-term health status of equipment and are difficult to identify potential risks in advance. Summary of the invention
[0004] The technical problem to be solved by this application is to provide a safety monitoring method, device, storage medium and electronic equipment for a chemical enterprise, which can quickly and accurately obtain safety monitoring results. The specific solution is as follows:
[0005] A safety monitoring method for a chemical enterprise, comprising:
[0006] Acquire equipment maintenance status data of target equipment of a chemical enterprise and safety monitoring time series data within a preset time period, wherein the preset time period is a time period including the current time point;
[0007] Input the safety monitoring time series data into a pre-trained prediction model to obtain the monitoring data prediction results output by the prediction model; the monitoring data prediction results include the predicted values of the safety monitoring data in the future;
[0008] Input the equipment maintenance status data and target safety monitoring data into a pre-built Bayesian network to obtain the failure probability of the target equipment; the target safety monitoring data is the safety monitoring data corresponding to the current time point in the safety monitoring time series data;
[0009] The safety assessment result of the chemical enterprise is obtained according to the predicted value of the safety monitoring data and the failure probability of the target equipment.
[0010] In the above method, optionally, the training process of the prediction model includes:
[0011] Obtain a training data set and an initial model; the training data set includes a plurality of training samples and a sample label of each training sample, each of the training samples includes safety monitoring time series data of a historical time period; the sample label includes actual safety monitoring data collected after the historical time period; the initial model includes a training neural network based on an attention mechanism to be trained;
[0012] The initial model is trained using the training data set until the initial model meets a preset training completion condition, and the initial model that meets the training completion condition is used as a prediction model.
[0013] The above method, optionally, inputting the equipment maintenance status data and the target safety monitoring data into a pre-built Bayesian network to obtain the failure probability of the target equipment includes:
[0014] Obtaining fault history information, environmental factor information, and personnel behavior of the target device;
[0015] The equipment maintenance status data, the target safety monitoring data, the fault history information, the environmental factor information and the personnel behavior are input into a pre-built Bayesian network to obtain the failure probability of the target equipment.
[0016] In the above method, optionally, obtaining the safety assessment result of the chemical enterprise according to the safety monitoring data prediction value and the failure probability of the target equipment includes:
[0017] Performing a weighted summation on the predicted value of the safety monitoring data and the failure probability of the target equipment to obtain a safety risk score for the chemical enterprise;
[0018] The safety risk score of the chemical enterprise is used as the safety assessment result.
[0019] The above method, optionally, after obtaining the safety risk score of the chemical enterprise, further comprises:
[0020] If the security risk score is greater than a preset score threshold, an alarm message is output.
[0021] A safety monitoring system for a chemical enterprise, comprising:
[0022] An acquisition unit, used to acquire equipment maintenance status data of a target equipment of a chemical enterprise and safety monitoring time series data within a preset time period, wherein the preset time period is a time period including the current time point;
[0023] A first execution unit is used to input the safety monitoring time series data into a pre-trained prediction model to obtain a monitoring data prediction result output by the prediction model; the monitoring data prediction result includes a safety monitoring data prediction value at a future time;
[0024] A second execution unit is used to input the equipment maintenance status data and target safety monitoring data into a pre-built Bayesian network to obtain the failure probability of the target equipment; the target safety monitoring data is the safety monitoring data corresponding to the current time point in the safety monitoring time series data;
[0025] The safety assessment unit is used to obtain the safety assessment result of the chemical enterprise according to the predicted value of the safety monitoring data and the failure probability of the target equipment.
[0026] In the above system, optionally, the first execution unit includes:
[0027] An acquisition subunit is used to acquire a training data set and an initial model; the training data set includes a plurality of training samples and a sample label of each training sample, each of the training samples includes safety monitoring time series data of a historical time period; the sample label includes actual safety monitoring data collected after the historical time period; the initial model includes a training neural network based on an attention mechanism to be trained;
[0028] The training subunit is used to train the initial model using the training data set until the initial model meets the preset training completion condition, and use the initial model that meets the training completion condition as the prediction model.
[0029] In the above system, optionally, the second execution unit includes:
[0030] An acquisition subunit, used to acquire fault history information, environmental factor information, and personnel behavior of the target device;
[0031] A subunit is constructed to input the equipment maintenance status data, the target safety monitoring data, the fault history information, the environmental factor information and the personnel behavior into a pre-constructed Bayesian network to obtain the failure probability of the target equipment.
[0032] A storage medium includes stored instructions, wherein when the instructions are executed, the device where the storage medium is located is controlled to execute a safety monitoring method for a chemical enterprise as described above.
[0033] An electronic device includes a memory and one or more instructions, wherein the one or more instructions are stored in the memory and are configured to be executed by one or more processors to implement a safety monitoring method for a chemical enterprise as described above.
[0034] Based on the above-mentioned implementation of this application, a safety monitoring method, device, storage medium and electronic device for a chemical enterprise are provided. The method includes: obtaining equipment maintenance status data of the target equipment of the chemical enterprise and safety monitoring time series data within a preset time period, wherein the preset time period is a time period including the current time point; inputting the safety monitoring time series data into a pre-trained prediction model to obtain the monitoring data prediction result output by the prediction model; the monitoring data prediction result includes the safety monitoring data prediction value of the future time; inputting the equipment maintenance status data and the target safety monitoring data into a pre-built Bayesian network to obtain the failure probability of the target equipment; the target safety monitoring data is the safety monitoring data corresponding to the current time point in the safety monitoring time series data; according to the safety monitoring data prediction value and the failure probability of the target equipment, the safety assessment result of the chemical enterprise is obtained. Through the method provided by the embodiment of the present application, an efficient data fusion mechanism is provided, which integrates the time series prediction ability of the neural network based on the attention mechanism loop training and the historical data analysis ability of the Bayesian network to achieve a more comprehensive and accurate safety assessment. Enhance the interpretability and transparency of the system so that each step of the reasoning process has a clear causal relationship and probability basis. Improve the robustness and stability of the system to ensure a high level of safety performance under various operating conditions. Quickly and accurately obtain safety monitoring results, support real-time decision-making and preventive maintenance, help enterprises plan in advance and take targeted measures to avoid unnecessary downtime and maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.
[0036] Figure 1 A method flow chart of a safety monitoring method for a chemical enterprise provided in this application;
[0037] Figure 2 A flowchart of a training process of a prediction model provided in this application;
[0038] Figure 3A schematic diagram of the structure of a safety monitoring device for a chemical enterprise provided in this application;
[0039] Figure 4 A schematic diagram of the structure of an electronic device provided in this application. DETAILED DESCRIPTION
[0040] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0041] In this application, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.
[0042] The embodiment of the present application provides a safety monitoring method for a chemical enterprise. The method can be applied to electronic devices, which may be computers, smart phones, tablet devices, smart wearable devices, server clusters or cloud computing centers, etc. The method flow chart of the method is as follows: Figure 1 As shown, specifically including:
[0043] S101: Acquire equipment maintenance status data of target equipment of a chemical enterprise and safety monitoring time series data within a preset time period, where the preset time period is a time period including the current time point.
[0044] In this embodiment, the equipment maintenance status may include information such as maintenance time, maintenance type, and maintenance results.
[0045] Optionally, the safety monitoring time series data may include a time series of safety monitoring data such as temperature, pressure, liquid level, harmful gas concentration, and vibration data collected at each time point within a preset time period.
[0046] S102: Input the safety monitoring time series data into a pre-trained prediction model to obtain a monitoring data prediction result output by the prediction model; the monitoring data prediction result includes a predicted value of safety monitoring data in the future.
[0047] In this embodiment, the prediction model can be a trained neural network based on the attention mechanism loop training. The prediction model may include an input layer, an encoder, an attention mechanism, a decoder, and an output layer. The input layer can be used to receive safety monitoring time series data, which includes data from the past T time steps, each time step has multiple features (such as temperature, pressure, liquid level, etc.), forming an input tensor of shape (T, D), where D is the number of features. The encoder uses multi-layer LSTM or GRU units as the infrastructure to process sequence data. Each unit is responsible for processing the data of one time step and passing the hidden state to the next unit. Assuming a two-layer LSTM is used, the output of each layer will be used as the input of the next layer, and the encoded hidden state H = [h1, h2, ..., hT] is finally obtained, where ht represents the hidden state of the tth time step. The attention mechanism is used to calculate the attention score of each time step based on the output of the encoder, and then normalize the attention score of each time step to obtain a normalized weight distribution, and then weighted sum the value of each time step according to the normalized weight distribution to obtain a context vector, and then the output of the previous time step and the context vector of the current time step are input into the decoder. The output layer can convert the output of the decoder into the final prediction value through one or more fully connected layers.
[0048] Optionally, the future time may be a time point after the current time, or a time period after the current time, and the time period after the current time may be used as the first future time period.
[0049] In some embodiments, after obtaining the monitoring data prediction results output by the prediction model, the prediction residual between the predicted value of the safety monitoring data and the actual observed value can also be determined; the prediction residual is input into the pre-trained isolated deep forest model to obtain the abnormal score of the prediction residual. Among them, the prediction residual can characterize the difference between the predicted value of the safety monitoring data (such as temperature, pressure, liquid level, etc.) and the actual observed value. The residual reflects the accuracy of the model prediction, and a larger residual may mean the occurrence of an abnormal situation. The calculation method of the prediction residual is prediction residual = actual observed value - safety monitoring data prediction value.
[0050] In some embodiments, if time series characteristics need to be considered, the timestamp can be input into the isolation forest model as an additional feature to help identify patterns that change over time. Depending on the specific application scenario, other features that are helpful for anomaly detection can also be introduced, such as device status, environmental conditions, etc.
[0051] Optionally, the output of the Isolation Forest model is an anomaly score or label for each input sample, indicating whether the sample is an outlier. The output can be in the form of an anomaly score or a binary label. The anomaly score is a continuous value that indicates the likelihood of the sample being an outlier. The higher the score, the more likely the sample is an outlier. Optionally, the Isolation Forest returns a score between 0 and 1, where values close to 1 indicate a high likelihood of anomaly.
[0052] Optionally, the anomaly score can be converted into a binary label (normal / abnormal) by setting a threshold. When the anomaly score exceeds the preset threshold, it is marked as abnormal; otherwise, it is marked as normal.
[0053] In this embodiment, the isolation forest isolates data points by randomly selecting features and randomly selecting split points on the selected features. Due to their uniqueness, outliers are usually easier to isolate than normal points. Therefore, outliers tend to have shorter path lengths (i.e., they can be isolated with fewer splits), while normal points require more splits to be isolated.
[0054] In some embodiments, the training process of the isolated deep forest model is as follows: collect and calculate the residuals between the predicted values of historical safety monitoring data and the actual observed values. And other relevant features (such as timestamps, equipment status, etc.). Use the residuals and other features in the historical data to train the isolated forest model. Determine the appropriate anomaly scoring threshold to ensure that both real abnormalities can be captured and too many false positives are not generated. In practical applications, new residuals are calculated in real time and input into the trained isolated forest model for anomaly detection. According to the anomaly score or label output by the model, the early warning mechanism is triggered to prompt potential safety risks. Collect feedback information after each event is processed to improve model performance. Regularly retrain the isolated forest model to ensure that it is always in the best performance state and adapts to new safety threats and changing working environments. By using the isolated forest model to monitor the predicted residuals of the predicted values of safety monitoring data, effective anomaly detection can be achieved in the safety monitoring system of chemical enterprises. This method can not only detect potential safety issues in a timely manner, but also reduce false positives and improve the reliability and practicality of the system.
[0055] S103: Input the equipment maintenance status data and target safety monitoring data into a pre-built Bayesian network to obtain the failure probability of the target equipment; the target safety monitoring data is the safety monitoring data corresponding to the current time point in the safety monitoring time series data.
[0056] In this embodiment, the failure probability of the target device may represent the probability of the target device failing within a second future time period, for example, it may represent the probability of the target device failing within a future day, and the second future time period may be longer than the first future time period.
[0057] S104: Obtaining a safety assessment result of the chemical enterprise according to the predicted value of the safety monitoring data and the failure probability of the target equipment.
[0058] In this embodiment, the safety monitoring data prediction value output by the prediction model and the failure probability of the target equipment output by the Bayesian network can be integrated to obtain the safety assessment result of the chemical enterprise.
[0059] In this embodiment, the attention mechanism-based cyclic training neural network is good at processing time series data and can capture dynamic change trends; while the Bayesian network provides a more comprehensive risk assessment through causal relationships and historical data. The combination of the two can make up for their respective limitations and improve the overall prediction accuracy. And the attention mechanism-based cyclic training neural network mainly predicts based on real-time monitoring data, while the Bayesian network integrates long-term information such as equipment maintenance status and fault history. Fusion of the two methods can utilize data from more dimensions to make more accurate judgments. The attention mechanism-based cyclic training neural network performs well in predicting under normal operating conditions, but may not be sensitive enough when facing abnormal situations. Bayesian networks can enhance the detection ability of abnormal situations by introducing additional historical and background information. Bayesian networks can better handle uncertainty and incomplete information, especially when data is sparse or noisy. The fused method can maintain high stability under different conditions.
[0060] In this embodiment, by comprehensively analyzing data from multiple sources, resources can be allocated to different risk items in a more scientific manner. For example, equipment with a high probability of failure and a significant impact can be prioritized. And the integrated evaluation results can help enterprises plan preventive maintenance activities in advance to avoid unnecessary downtime and maintenance costs.
[0061] In an embodiment provided in the present application, based on the above solution, optionally, the training process of the prediction model is as follows: Figure 2 As shown, including:
[0062] S201: Obtain a training data set and an initial model; the training data set includes multiple training samples and a sample label for each training sample, each of the training samples includes safety monitoring time series data of a historical time period; the sample label includes actual safety monitoring data collected after the historical time period; the initial model includes a training neural network to be trained based on an attention mechanism.
[0063] S202: The initial model is trained using the training data set until the initial model meets a preset training completion condition, and the initial model that meets the training completion condition is used as a prediction model.
[0064] In this embodiment, the training completion condition may include that the number of training times is greater than a preset training number threshold or the prediction accuracy of the initial model is greater than a preset accuracy threshold.
[0065] In an embodiment provided by the present application, based on the above solution, optionally, inputting the equipment maintenance status data and the target safety monitoring data into a pre-built Bayesian network to obtain the failure probability of the target equipment includes:
[0066] Obtaining fault history information, environmental factor information, and personnel behavior of the target device;
[0067] The equipment maintenance status data, the target safety monitoring data, the fault history information, the environmental factor information and the personnel behavior are input into a pre-built Bayesian network to obtain the failure probability of the target equipment.
[0068] In this embodiment, the input of the Bayesian Network (BN) is the state or observation value of each node. In the application of safety monitoring in chemical enterprises, these inputs may include but are not limited to: equipment maintenance status, fault history information, target safety monitoring data, personnel behavior, and environmental factor information.
[0069] Optionally, the equipment maintenance status may include the time of the most recent maintenance, the maintenance type, the maintenance results, etc. Fault history information may include past fault records of the equipment and their corresponding repair measures. Target safety monitoring data may include real-time monitoring data such as temperature, pressure, liquid level, and hazardous gas concentration. Personnel behavior may include compliance with operating procedures, training records, etc. Environmental factors may include external conditions such as humidity and temperature that affect equipment operation. These inputs are usually discretized into a finite number of categories or states to facilitate probability calculations by Bayesian networks.
[0070] In this embodiment, the output of the Bayesian network is the probability distribution of each target node. Specifically, in the safety monitoring of chemical enterprises, it can include the failure probability of the target equipment. The failure probability represents the possibility of the target equipment failing in a certain time period in the future.
[0071] In some embodiments, the Bayesian network construction process is as follows:
[0072] First, identify the nodes.
[0073] Optionally, all nodes in the network are first defined, each node representing a variable or event, such as temperature, pressure, liquid level, etc. For safety monitoring of chemical enterprises, the nodes can be the various input variables mentioned above and the target variables that need to be predicted (such as equipment failure).
[0074] Second, define the dependencies between nodes.
[0075] Optionally, determine which nodes have causal relationships and connect them with directed edges. For example, in a chemical company, temperature changes may affect pressure, and both of them work together to affect the health of the equipment. This causal relationship can be determined through expert knowledge, historical data analysis, or a combination of both.
[0076] Third, build a network structure.
[0077] Optionally, a structure diagram of the Bayesian network is drawn based on the nodes and their dependencies. The structure diagram may be a directed acyclic graph (DAG), in which arrows point to indicate causal directions.
[0078] Fourth, learn the conditional probability table (CPT).
[0079] In this embodiment, a conditional probability table (CPT) can be created for each node to describe the probability of the node taking different values given the parent node state. For the root node (a node without a parent node), only its marginal probability distribution needs to be defined.
[0080] Alternatively, if there is sufficient historical data, the probability values in the CPT can be learned from the data through Maximum Likelihood Estimation (MLE) or Bayesian estimation. When there is a lack of sufficient data, the knowledge of domain experts can be used to manually set or adjust the CPT.
[0081] Fifth, model verification and optimization.
[0082] The performance of the constructed Bayesian network is tested through cross-validation or other evaluation methods to ensure that it can accurately reflect the actual situation. If necessary, the network structure is adjusted or the probability value in the CPT is re-estimated to improve the model accuracy.
[0083] Sixth, reasoning and prediction.
[0084] After the construction is completed, Bayesian reasoning algorithms (such as exact reasoning or approximate reasoning methods) can be used for reasoning, that is, to infer the probability distribution of unknown target variables based on known input data. This step can be used in practical applications to predict the probability of equipment failure, assess safety risks, etc.
[0085] In this embodiment, the Bayesian network can not only integrate multi-source heterogeneous data, but also provide intuitive risk assessment and decision support. By rationally constructing and training the Bayesian network model, it can help enterprises prevent accidents more effectively and ensure production safety.
[0086] In an embodiment provided in the present application, based on the above solution, optionally, obtaining the safety assessment result of the chemical enterprise according to the safety monitoring data prediction value and the failure probability of the target equipment includes:
[0087] Performing a weighted summation on the predicted value of the safety monitoring data and the failure probability of the target equipment to obtain a safety risk score for the chemical enterprise;
[0088] The safety risk score of the chemical enterprise is used as the safety assessment result.
[0089] In this embodiment, the security risk score is calculated as follows:
[0090] Safety risk score = W1 × safety monitoring data prediction value + W2 × failure probability value
[0091] Wherein, W1 is the first weight coefficient, W2 is the second weight coefficient, and the first weight coefficient and the second weight coefficient can be determined through cross-validation or expert knowledge to optimize the final safety assessment result.
[0092] In an embodiment provided by the present application, based on the above solution, optionally, after obtaining the safety risk score of the chemical enterprise, the method further includes:
[0093] If the security risk score is greater than a preset score threshold, an alarm message is output.
[0094] In this embodiment, the alarm information may be sent to a preset terminal, or the alarm information may be output and displayed on a preset interface.
[0095] In some embodiments, another feasible way to obtain the safety assessment result of the chemical enterprise according to the safety monitoring data prediction value and the failure probability of the target equipment is:
[0096] Using Bayesian theorem, the predicted value of safety monitoring data is used as new evidence to update the probability estimate of equipment failure in the Bayesian network. The details are as follows:
[0097]
[0098] Specifically, P(failure) is the failure probability of the target device initially calculated by the Bayesian network.
[0099] P(new evidence|fault) is the probability (likelihood function) of the predicted value of the safety monitoring data under the condition of a known fault.
[0100] P(new evidence) is the marginal probability of the predicted value of safety monitoring data and can be calculated using the total probability formula.
[0101] Finally, the updated failure probability P(failure|new evidence) of the target device is obtained for more accurate risk assessment, and the updated failure probability of the target device is determined as the safety assessment result.
[0102] In some embodiments, another feasible way to obtain the safety assessment result of the chemical enterprise according to the safety monitoring data prediction value and the failure probability of the target equipment is:
[0103] The predicted value of the safety monitoring data and the failure probability of the target equipment are input into the pre-trained safety assessment model to obtain the safety assessment result.
[0104] In this embodiment, the security assessment model may be a decision tree model or a random forest model.
[0105] Optionally, the process of training a security assessment model is as follows:
[0106] Prepare a training sample set, each sample in the training sample set contains the predicted value of historical safety monitoring data and the Bayesian network probability as features, as well as the corresponding actual safety status label.
[0107] Then, the security assessment model is trained based on the training sample set so that the security assessment model can learn the security status under different feature combinations.
[0108] In practical applications, the predicted values of safety monitoring data and the Bayesian network probability are input into the trained model to obtain the final safety assessment results.
[0109] and Figure 1 Corresponding to the method described above, the embodiment of the present application also provides a safety monitoring device for a chemical enterprise, which is used to monitor Figure 1 The specific implementation of the method is shown in the following diagram: Figure 3 As shown, specifically including:
[0110] An acquisition unit 301 is used to acquire equipment maintenance status data of a target equipment of a chemical enterprise and safety monitoring time series data within a preset time period, wherein the preset time period is a time period including the current time point;
[0111] The first execution unit 302 is used to input the safety monitoring time series data into a pre-trained prediction model to obtain the monitoring data prediction result output by the prediction model; the monitoring data prediction result includes the safety monitoring data prediction value at a future time;
[0112] The second execution unit 303 is used to input the equipment maintenance status data and the target safety monitoring data into a pre-built Bayesian network to obtain the failure probability of the target equipment; the target safety monitoring data is the safety monitoring data corresponding to the current time point in the safety monitoring time series data;
[0113] The safety assessment unit 304 is used to obtain the safety assessment result of the chemical enterprise according to the predicted value of the safety monitoring data and the failure probability of the target equipment.
[0114] In an embodiment provided in the present application, based on the above solution, optionally, the first execution unit 302 includes:
[0115] An acquisition subunit is used to acquire a training data set and an initial model; the training data set includes a plurality of training samples and a sample label of each training sample, each of the training samples includes safety monitoring time series data of a historical time period; the sample label includes actual safety monitoring data collected after the historical time period; the initial model includes a training neural network based on an attention mechanism to be trained;
[0116] The training subunit is used to train the initial model using the training data set until the initial model meets the preset training completion condition, and use the initial model that meets the training completion condition as the prediction model.
[0117] In an embodiment provided in the present application, based on the above solution, optionally, the second execution unit 303 includes:
[0118] An acquisition subunit, used to acquire fault history information, environmental factor information, and personnel behavior of the target device;
[0119] A subunit is constructed to input the equipment maintenance status data, the target safety monitoring data, the fault history information, the environmental factor information and the personnel behavior into a pre-constructed Bayesian network to obtain the failure probability of the target equipment.
[0120] In an embodiment provided in the present application, based on the above solution, optionally, the security assessment unit 304 includes:
[0121] A calculation subunit, used for performing a weighted summation of the safety monitoring data prediction value and the failure probability of the target equipment to obtain a safety risk score of the chemical enterprise;
[0122] The execution subunit is used to use the safety risk score of the chemical enterprise as a safety assessment result.
[0123] In an embodiment provided in the present application, based on the above solution, optionally, it further includes:
[0124] The output unit is used to output an alarm message if the security risk score is greater than a preset score threshold.
[0125] The specific principles and execution processes of each unit and module in the safety monitoring device for a chemical enterprise disclosed in the above-mentioned embodiment of the present application are the same as those of the safety monitoring method for a chemical enterprise disclosed in the above-mentioned embodiment of the present application. Please refer to the corresponding parts of the safety monitoring method for a chemical enterprise provided in the above-mentioned embodiment of the present application, and will not be repeated here.
[0126] An embodiment of the present application further provides a storage medium, which includes stored instructions, wherein when the instructions are executed, the device where the storage medium is located is controlled to execute the above-mentioned safety monitoring method for a chemical enterprise.
[0127] The present application also provides an electronic device, whose structural diagram is as follows: Figure 4 As shown, it specifically includes a memory 401 and one or more instructions 402, wherein the one or more instructions 402 are stored in the memory 401 and are configured to be executed by one or more processors 403 to perform the following operations:
[0128] Acquire equipment maintenance status data of target equipment of a chemical enterprise and safety monitoring time series data within a preset time period, wherein the preset time period is a time period including the current time point;
[0129] Input the safety monitoring time series data into a pre-trained prediction model to obtain the monitoring data prediction results output by the prediction model; the monitoring data prediction results include the predicted values of the safety monitoring data in the future;
[0130] Input the equipment maintenance status data and target safety monitoring data into a pre-built Bayesian network to obtain the failure probability of the target equipment; the target safety monitoring data is the safety monitoring data corresponding to the current time point in the safety monitoring time series data;
[0131] The safety assessment result of the chemical enterprise is obtained according to the predicted value of the safety monitoring data and the failure probability of the target equipment.
[0132] It should be noted that each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the embodiments can be referred to each other. For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.
[0133] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations.
[0134] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and / or hardware.
[0135] It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly contributed to the prior art in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present application or certain parts of the embodiments.
[0136] The above is a detailed introduction to a safety monitoring method for a chemical enterprise provided by the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for general technical personnel in this field, according to the idea of the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.
Claims
1. A safety monitoring method for a chemical enterprise, characterized in that: include: Acquire equipment maintenance status data of target equipment of a chemical enterprise and safety monitoring time series data within a preset time period, wherein the preset time period is a time period including the current time point; Input the safety monitoring time series data into a pre-trained prediction model to obtain the monitoring data prediction results output by the prediction model; the monitoring data prediction results include the predicted values of the safety monitoring data in the future; Inputting the equipment maintenance status data and target safety monitoring data into a pre-built Bayesian network to obtain the failure probability of the target equipment; The target safety monitoring data is the safety monitoring data corresponding to the current time point in the safety monitoring time series data; The safety assessment result of the chemical enterprise is obtained according to the predicted value of the safety monitoring data and the failure probability of the target equipment.
2. The method according to claim 1, characterized in that The training process of the prediction model includes: Obtain a training data set and an initial model; the training data set includes a plurality of training samples and a sample label of each training sample, each of the training samples includes safety monitoring time series data of a historical time period; the sample label includes actual safety monitoring data collected after the historical time period; the initial model includes a training neural network based on an attention mechanism to be trained; The initial model is trained using the training data set until the initial model meets a preset training completion condition, and the initial model that meets the training completion condition is used as a prediction model.
3. The method according to claim 1, characterized in that The step of inputting the equipment maintenance status data and the target safety monitoring data into a pre-built Bayesian network to obtain the failure probability of the target equipment includes: Obtaining fault history information, environmental factor information, and personnel behavior of the target device; The equipment maintenance status data, the target safety monitoring data, the fault history information, the environmental factor information and the personnel behavior are input into a pre-built Bayesian network to obtain the failure probability of the target equipment.
4. The method according to claim 1, characterized in that: The obtaining of the safety assessment result of the chemical enterprise according to the safety monitoring data prediction value and the failure probability of the target equipment includes: Performing a weighted summation on the predicted value of the safety monitoring data and the failure probability of the target equipment to obtain a safety risk score for the chemical enterprise; The safety risk score of the chemical enterprise is used as the safety assessment result.
5. The method according to claim 4, characterized in that After obtaining the safety risk score of the chemical enterprise, it also includes: If the security risk score is greater than a preset score threshold, an alarm message is output.
6. A safety monitoring system for a chemical enterprise, characterized in that: include: An acquisition unit, used to acquire equipment maintenance status data of a target equipment of a chemical enterprise and safety monitoring time series data within a preset time period, wherein the preset time period is a time period including the current time point; A first execution unit is used to input the safety monitoring time series data into a pre-trained prediction model to obtain a monitoring data prediction result output by the prediction model; the monitoring data prediction result includes a safety monitoring data prediction value at a future time; A second execution unit is used to input the equipment maintenance status data and the target safety monitoring data into a pre-built Bayesian network to obtain the failure probability of the target equipment; The target safety monitoring data is the safety monitoring data corresponding to the current time point in the safety monitoring time series data; The safety assessment unit is used to obtain the safety assessment result of the chemical enterprise according to the predicted value of the safety monitoring data and the failure probability of the target equipment.
7. The system according to claim 6, characterized in that The first execution unit includes: An acquisition subunit is used to acquire a training data set and an initial model; the training data set includes a plurality of training samples and a sample label of each training sample, each of the training samples includes safety monitoring time series data of a historical time period; the sample label includes actual safety monitoring data collected after the historical time period; the initial model includes a training neural network based on an attention mechanism to be trained; The training subunit is used to train the initial model using the training data set until the initial model meets the preset training completion condition, and use the initial model that meets the training completion condition as the prediction model.
8. The system according to claim 6, characterized in that The second execution unit includes: An acquisition subunit, used to acquire fault history information, environmental factor information, and personnel behavior of the target device; A subunit is constructed to input the equipment maintenance status data, the target safety monitoring data, the fault history information, the environmental factor information and the personnel behavior into a pre-constructed Bayesian network to obtain the failure probability of the target equipment.
9. A storage medium, characterized in that: The storage medium includes stored instructions, wherein when the instructions are executed, the device where the storage medium is located is controlled to execute the safety monitoring method for a chemical enterprise according to any one of claims 1 to 5.
10. An electronic device, characterized in that: It comprises a memory and one or more instructions, wherein the one or more instructions are stored in the memory and are configured to be executed by one or more processors to implement a safety monitoring method for a chemical enterprise as described in any one of claims 1 to 5.