A method for automatically generating FTA real-time dynamic tree based on fault knowledge base

Abstract

To automatically generate an FTA real-time dynamic tree based on a fault knowledge base.SOLUTION: In step 1, an FTA logic model is newly constructed by a method for generating an FTA real-time dynamic tree based on a fault knowledge base. In step 2, a fault knowledge base model is selected based on a logical device. In step 3, a fault library is constructed based on the FTA logic model to generate a logic rule for an FTA model, and the fault knowledge base model is processed to perform decomposition-mapping into an FTA logic fault tree model. In step 4, the fault diagnosis tool for the event node is set. In step 5, a formation of an FTA logic fault tree is completed. In step 6, the FTA logic model is published, and the FTA logic model is instantiated. In step 7, a scheduling task is configured. In step 8, the configured scheduling task performs a periodic fault diagnosis and issues a fault alarm.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention belongs to the field of industrial Internet platform, and particularly relates to a method for automatically generating FTA real-time dynamic tree based on fault knowledge base. [Background technology]

[0002] As the safe and stable operation of power grids and power plants improves, so does the standardization and refinement of management. Power plants are increasingly seeking intelligent fault diagnosis methods for electromechanical devices. Fault tree analysis (FTA) models are increasingly being used to perform reverse fault analysis and diagnosis of device failures. Fault tree analysis employs intelligent logic reasoning mechanisms and high diagnostic accuracy. However, within this process, the creation of an FTA fault tree still requires manual construction. A complete fault tree model can only be created through five steps: investigating the incident, determining the top event, determining the target value, investigating the causal event, and constructing the fault tree. Second, traditional FTA fault tree models require the constructor to determine the probability of all basic events beforehand; otherwise, quantitative analysis is not possible. Therefore, FTA fault trees are limited to static fault analysis.

[0003] Through analysis of existing technologies, the following drawbacks were found: 1. FTA's ability to analyze the causes of accidents is a strength, but to quantitatively analyze a system, the probability of all basic events occurring must be determined in advance, otherwise quantitative analysis cannot be performed. 2. The analyst must be familiar with the system being analyzed and be able to apply the analysis method accurately and skillfully. It is common for the fault tree created by different analysts to differ from the analysis results. 3. For complex systems, analysts have to go through many steps to create a fault tree, and the created fault tree is often huge and requires complex calculations, making it difficult to perform qualitative and quantitative analysis. Summary of the Invention

[0004] In view of the technical problems existing in the background art, the present invention provides a method for automatically generating a real-time dynamic FTA tree based on a fault knowledge base, which reduces the demands on personnel for creating fault trees, reduces the amount of work required for creation, and shortens time costs.By binding fault node devices and setting event node fault diagnosis tools, real-time dynamic fault diagnosis analysis is realized.

[0005] To solve the above technical problems, the present invention adopts and implements the following technical solutions. A method for automatically generating an FTA real-time dynamic tree based on a fault knowledge base, comprising: Step 1: constructing a new FTA logic model by generating a FTA real-time dynamic tree based on a fault knowledge base; Step 2: selecting a fault knowledge base model based on a logical device; Step 3: building a fault library based on the FTA logic model to generate logic rules for the FTA model, and decomposing and mapping the fault knowledge base model into an FTA logic fault tree model; Step 4: Setting up the fault diagnosis tool for the event node; Step 5 to complete the creation of the FTA logical fault tree; Step 6: publishing the FTA logic model and instantiating the FTA logic model; Step 7: Configure the scheduling task; and step 8 of periodically performing fault diagnosis and issuing a fault alarm according to the configured scheduling task.

[0006] Preferably, in step 1, the purpose of generating an FTA real-time dynamic tree based on the fault knowledge base is to eliminate the analysis process of manually identifying top events, intermediate events, basic events, and event-related logic gates, and the process of stepwise drawing a fault tree by dragging on a drawing board.

[0007] Preferably, in step 2, the logical device is an abstract description of a class of instantiated devices. For example, a water-conducting bearing is a logical device, and a water-conducting bearing of a unit of a power plant is an instantiated physical device. The sub-steps of step 2 are as follows: In step 2.1, the failure modes are filtered based on the logical device and the failure mode that the current node matches is selected. In step 2.2, the entire failure mode tree is mapped and transformed into an FTA fault tree logic model.

[0008] Preferably, in step 3, the fault library generates logic rules for the FTA model based on the fault knowledge base model selected in step 2 by mapping the top node to a top event of the FTA fault tree, the middle node to a middle event node of the FTA fault tree, and the end node to a basic event node of the FTA fault tree, where the top node, the middle node, and the end node are branch nodes in the FTA fault tree.

[0009] Preferably, when mapping to nodes, the associations of each hierarchical node are mapped synchronously, and each hierarchical node includes a top node, an intermediate node, and an end node.

[0010] Preferably, in step 4, the fault diagnosis tool includes measurement points, diagnostic rules, algorithm models and scripts.

[0011] The principle of diagnosing the measurement points is to select the measurement points of the detection device under the logical device related to the node, and diagnose the occurrence probability and fault occurrence status of the node based on the data of the measurement points.

[0012] The diagnostic principle of the diagnostic rules is to call the rule engine based on the business rules related to the logical devices of the node to diagnose the occurrence probability and fault occurrence status of the node.

[0013] The diagnostic principle of the algorithm model is to call the algorithm engine based on the hydropower-specific algorithm associated with the node's logic device to diagnose the occurrence probability and fault occurrence status of the node.

[0014] The diagnostic principle of the script is as follows: Based on the relevant cases in the fault case library related to the logical device of the node, the case library diagnostic algorithm is invoked to diagnose the occurrence probability and fault occurrence situation of the node.

[0015] Preferably, in step 4, the purpose of setting the fault diagnosis tool for the node of the fault tree is to perform dynamic analysis using the fault tree. The dynamic analysis includes the following situations: 1) For the measurement points of the detection device of the logical device, the probability of failure occurrence is predicted by the diagnostic tool of the measurement point. 2) A rule diagnosis engine is called for the business rules of the logical device to predict the probability of failure. 3) If the amount of data at the device measurement point is greater than a preset threshold, an algorithm engine associated with the logical device is invoked to predict the node failure probability.

[0016] Preferably, the sub-steps of step 6 are as follows: In step 6.1, you first publish the logic model and activate the newly created FTA fault tree logic model. In step 6.2, the next step is to instantiate, and the plant and instantiation device tree corresponding to the model to be instantiated are selected, and the model instantiation process is entered. In the instantiation process, the logical devices bound to the nodes of the FTA fault tree logical model are converted to physical devices in the plant's instantiation device tree, and the logical fault diagnosis tools bound to the nodes are converted to instantiation tools.

[0017] Preferably, the sub-steps of step 7 are as follows: In step 7.1, a timing task is created to periodically invoke the instantiated model for fault diagnosis. Step 7.2 is the diagnosis process. If a diagnostic tool is set for the event node, the diagnostic tool set for the node is called sequentially to obtain the alarm status and occurrence probability of the node, and the occurrence probability is substituted into the fault tree model. If a diagnostic tool is not set for the intermediate node, the occurrence probability of the node is estimated by the fault diagnosis method until the alarm status and occurrence probability of the top event are finally estimated.

[0018] Preferably, in step 8, the scheduling task created in step 7 is subjected to periodic fault diagnosis to generate a complete FTA fault tree, which identifies the fault conditions and occurrence probability of each event node.

[0019] The present invention can achieve the following beneficial effects: In conventional FTA fault tree diagnosis, the occurrence probability of basic events is derived solely through the experience of the fault tree creator, the static occurrence probability of the fault tree basic events is manually set, and then the fault tree diagnosis is manually performed to derive the occurrence probability of the top event.The method of the present invention binds the device and the fault diagnosis tool with the node event of the fault tree, so that real-time data of the monitoring measurement point of the current device can be obtained, and different fault diagnosis tools are adopted according to the actual situation to dynamically calculate and set the probability of the FTA fault tree event node.By using a task configuration method, FTA fault tree diagnosis is performed periodically in real time, and one FTA fault instance (diagnosis result) is generated each time. [Brief explanation of the drawings]

[0020] The present invention will be further described below in conjunction with the accompanying drawings and examples. [Figure 1] 1 is a flow chart of a method of automatically generating an FTA real-time dynamic tree based on a fault knowledge base and performing fault detection according to the present invention. [Figure 2] FIG. 2 is a diagram showing a knowledge base model of shaft system operation failure of a hydroelectric power generating unit according to the present invention. [Figure 3] FIG. 1 is a FAT fault tree diagram of a hydroelectric power generation unit shaft system operation fault of the present invention. [Figure 4] FIG. 2 is a model diagram of the generator fault knowledge base of the present invention. [Figure 5] FIG. 1 is a fault tree diagram of a generator fault FAT of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0021] The conventional fault tree drawing steps will be explained using a currently commonly used conventional fault tree drawing tool (FreeFta) as an example.

[0022] In step 1, the top event and logic gate are identified based on the results of the accident investigation and statistical analysis, referring to the frequency of accidents and the severity of accident losses. Drag the top event from the toolbar onto the canvas and set the top event number (note that the top event number is T) and name. This step is performed by the computer.

[0023] In step 2, intermediate process events are created sequentially as intermediate event fault tree nodes according to the analyzed event causes. Click the top event to add intermediate events and logic gates, and set the intermediate event number (the intermediate event number is the first part of M + a number) and name. This step is performed by a computer.

[0024] In step 3, the basic event that causes the intermediate accident is placed in the basic event node of the fault tree. Click on the intermediate event node to add a basic event, and set the basic event number (basic event number is the first letter of X + a number), name, and occurrence probability (occurrence probability is 0 to 1, including 0 and 1). This step is performed by a computer.

[0025] In step 4, the created fault tree is checked to see if it complies with the principles of logical analysis, the connections of the logic gates are checked, whether the upper-level events are necessary consequences of the lower-level events, whether the lower-level events are sufficient causes of the upper-level events, and whether all the direct cause events are present. This step is performed by a computer.

[0026] In step 5, the fault tree is drawn and the entire static tree is saved, ensuring that there are no errors. This step is performed by the computer.

[0027] The above steps require the analyst to be familiar with the system being analyzed and to apply the analysis method accurately and skillfully. Furthermore, for complex systems, the operating procedures are complicated and the amount of calculations is large, which to a certain extent poses great difficulties to qualitative and quantitative analysis. To solve this problem, the present invention generates an FTA fault tree model by directly mapping the fault tree representation of an existing fault case library, thereby reducing the demands on fault tree construction personnel, reducing the construction workload, and saving time and cost. By binding fault node devices and setting up event node fault diagnosis tools, real-time dynamic fault diagnosis analysis is realized. The specific solution is as follows:

[0028] As shown in FIGS. 1 to 5, the method for automatically generating an FTA real-time dynamic tree based on a fault knowledge base proposed in this application includes the following steps: In step 1, a new FTA logical model is created by generating an FTA real-time dynamic tree based on the fault knowledge base. This step is performed by a computer. Specifically, a new FTA logic model is created, and a real-time dynamic FTA tree is generated based on the fault knowledge base. This step eliminates the need for manual analysis of top events, intermediate events, basic events, and event-related logic gates, and the tedious process of creating a fault tree step by step using a drag-and-drop method on the drawing board.

[0029] In step 2, a fault knowledge base model is selected based on a logical device, where the logical device is an abstract description of a class of instantiated devices, such as a hydroelectric power generating unit, a rotor, an upper guide shoe, an upper frame, etc. The fault knowledge base model builds a knowledge base of power device failure modes and their effects according to specified fields (not limited to) of historical device failures (multi-level) at different levels such as system, subsystem, device, and component. This step is performed by a computer.

[0030] The relationship between logical devices and fault knowledge base models is as follows: All historical faults recorded in the fault knowledge base are managed according to logical device classification, so mapping and conversion from cases in the fault knowledge base to FTA fault trees should also be done according to logical device classification.

[0031] Specifically, in this step, the failure modes are first filtered according to the logical device, and the failure mode that the current node applies to is selected. Then, by default, the entire tree of failure modes is mapped and converted into the FTA fault tree logical model. However, this also supports cutting a part of the branch of the knowledge base model tree for conversion (the branch to be cut must include the root node and the base node).

[0032] Referring to Figure 2, if "Hydroelectric Power Generation Unit Shaft System Operation Fault" is selected as the fault knowledge base, the user can select a root node or intermediate node of the model tree, such as "Hydroelectric Power Generation Unit Shaft System Operation Fault", "Rotating Unit Fault", or "Shaft System Instability", to be cut as the top event node of the FTA fault tree. If no selection is made, the entire model tree will be converted by default.

[0033] In step 3, the fault knowledge base model is dynamically decomposed and mapped into an FTA logical fault tree model based on the FTA logical model construction rules. This step is performed by a computer.

[0034] Specifically, based on the fault knowledge base model selected in step 2, a) for example, select the "rotating unit failure" node as the fault tree top event. The "rotating unit failure" knowledge base model node is converted to an FTA fault tree top event node, and the node's logic gate (OR gate) is set to the logic gate of the fault tree node, the name is set to the failure mode of the fault tree node, and the associated logic device is set to the associated logic device of the fault tree node. b) The intermediate "shaft system instability" node of the fault knowledge base model tree is mapped to the intermediate event node of the FTA fault tree model. The node data relationship is the same as that of the top node. c) The end nodes of the fault knowledge base model tree, "loose upper frame support bolt," "loose bush support member," "large bush clearance," and "shaft system change," are mapped to the basic event nodes of the FTA fault tree model, in that order. The node data relationship is the same as that of the top node. d) The fault knowledge base model is mapped to the FTA fault tree logical model, and the mapping is complete.

[0035] In step 4, the fault diagnosis tool for the event node is set. This step is performed by a computer.

[0036] If you need to selectively perform static analysis using a fault tree, you need to manually set the occurrence probability of the fault tree's basic events. To dynamically diagnose a fault tree in real time, you need to manually set the fault diagnosis tools for the fault tree nodes. The current fault diagnosis tool supports four types of diagnosis tools: measurement points, diagnosis rules, algorithm models, and scripts. The configuration rules are as follows:

[0037] A fault diagnosis tool is selected based on the type of device currently configured in the fault node. If the probability of failure can be directly diagnosed based on the monitored data values ​​of the device measurement points, or if the probability of failure can be diagnosed by simple conditional judgment, the fault diagnosis tool can be configured directly as a measurement point. If a business rule is required to diagnose the probability of failure in this device, the fault diagnosis tool can be configured as a diagnosis rule. If the amount of data at the device measurement points is larger than a preset threshold, or if the calculation method is complex and requires the calculation of a special hydraulic algorithm to diagnose the probability of failure, the fault diagnosis tool can be configured as an algorithm model. If the failures that occur in this device can be covered by the fault case library, the fault diagnosis tool can be configured as a script. i. Measurement point: A measurement point of a detection device under a logical device related to the node is selected, and the occurrence probability and occurrence status of a failure of the node are diagnosed based on the data of the measurement point. ii. Diagnosis rules: Based on the business rules related to the logical devices of the node, the rule engine is called to diagnose the occurrence probability and fault occurrence status of the node. iii. Algorithm model: Based on the hydropower-specific algorithm associated with the node's logical device, the algorithm engine is invoked to diagnose the node's occurrence probability and fault occurrence status. iv. Script: Based on the relevant cases in the failure case library related to the logical device of the node, the case library diagnosis algorithm is called to diagnose the occurrence probability and occurrence status of the node's failure.

[0038] As shown in Figure 3, when the related logical device is "rotor," there is a matching case for the corresponding fault in the fault case library, so a script is applied as the fault diagnosis tool.When setting up a fault diagnosis tool for a node, a script-based fault detection tool is set up for that node based on the above rules, and a script for "amplitude remains almost unchanged when rotation is constant" is set up for the node "rotating unit failure," and a script for "vibrates regardless of excitation or load" is set up for the node "shaft system instability."

[0039] Step 5 completes the construction of the FTA logical fault tree. This step is performed by a computer. Specifically, the created FTA fault tree logical model is saved and later used for instantiation of another physical device.

[0040] Step 6 is to publish the logic model and instantiate the FTA logic model. This step is performed by a computer. Specifically, the logical model is first published, and the newly created FTA fault tree logical model is enabled. Only enabled models can be instantiated. Next, the instantiation operation is performed, and the plant and instantiation device tree corresponding to the model to be instantiated are selected to enter the model instantiation process. During the instantiation process, the logical devices bound to the nodes of the FTA fault tree logical model are converted to physical devices in the plant's instantiation device tree, and the logical fault diagnosis tools bound to the nodes are converted to instantiation tools.

[0041] Step 7 configures the scheduling task. This step is performed by a computer. Specifically, this step is as follows: a) A timing task is created and the instantiation model is periodically called to perform fault diagnosis. b) If a diagnostic tool is set for the event node, the diagnostic process sequentially calls the diagnostic tool set for the node to obtain the alarm state and occurrence probability of the node and assigns the occurrence probability to the fault tree model. If a diagnostic tool is not set for the intermediate node, the occurrence probability of the node is estimated using a fault diagnosis method (such as Min Cut Set or Min Path Set) until the alarm state and occurrence probability of the top event are finally estimated.

[0042] Step 8 is to issue a fault alarm. This step is performed by the computer. Periodic fault diagnosis is performed according to the scheduling task configured in step 7, and an instance diagram of the fault diagnosis results, i.e., a complete FTA fault tree, is generated. This tree identifies the fault conditions and occurrence probabilities of each event node.

[0043] 10 is a description of related concepts in this embodiment. Logical Modeling: Use logical models to define static FTA fault trees, completing the static analysis capabilities of traditional fault trees. Reuse logical models to generate instantiation models. Model Instances: Model instances are used to realize dynamic fault tree diagnosis upgrades. Scheduling Configuration: The scheduling configuration can be used to schedule and execute instantiated fault tree model diagnosis periodically or in real time. Fault Instance: By combining model instances and scheduling configurations, we realize real-time dynamic fault tree diagnosis capabilities. One fault tree instance is generated for each diagnosis, which shows the real-time fault tree occurrence probability and fault conditions.

[0044] Example 1 Here, the generator fault tree generation process will be taken as an example to describe the present invention in detail with reference to the attached drawings 4-5. In step 1, a new FTA logical model is constructed by generating an FTA real-time dynamic tree based on the fault knowledge base. This step is performed by a computer.

[0045] In step 2, the fault knowledge base is filtered by the logical path "main transformer system" in the failure mode list to find the "generator failure" model, and an FTA model is generated. This step is performed by a computer.

[0046] In step 3, the fault knowledge base model is dynamically decomposed and mapped into the FTA logical fault tree model based on the construction rules of the FTA logical model. This step is performed by a computer. Specifically, the fault knowledge base model selected in step 2 will be explained with reference to Figure 4. a) As an example, select the "Generator Failure" node as the fault tree top event. Convert the "Generator Failure" knowledge base model node to an FTA fault tree top event node, and set the node's logic gate (OR gate) to the logic gate of the fault tree node, its name to the failure mode of the fault tree node, and its associated logic device to the associated logic device of the fault tree node. b) Map the intermediate "High Temperature Overheating (Over 700) °C" node of the fault knowledge base model tree to the FTA fault tree model intermediate event node. The node data relationship is the same as that of the top node. c) Map the end node "Transformer Failure" of the fault knowledge base model tree to the basic event node of the FTA fault tree model. The node data relationship is the same as that of the top node. d) Map the fault knowledge base model to the FTA fault tree logic model, and the mapping is complete.

[0047] In step 4, the fault diagnosis tool for the event node is set. This step is performed by a computer. Associate the logical device "Main transformer" with the "High temperature overheating (over 700) °C" node, and set the fault diagnosis tool to the diagnosis rule (transformer three ratio value diagnosis method).

[0048] Step 5 completes the construction of the FTA logical fault tree. This step is performed by a computer. Specifically, the created FTA fault tree logical model can be saved and later used for instantiating other physical devices, as shown in Figure 5.

[0049] Step 6 is to publish the logic model and instantiate the FTA logic model. This step is performed by a computer. Specifically, the logical model is first published, and the newly created FTA fault tree logical model is enabled. Only enabled models can be instantiated. Next, the instantiation operation is performed, and the plant and instantiation device tree corresponding to the model to be instantiated are selected to enter the model instantiation process. During the instantiation process, the logical devices bound to the nodes in the FTA fault tree logical model are converted to physical devices in the plant's instantiation device tree, and the logical fault diagnosis tools bound to the nodes are converted to instantiation tools.

[0050] Step 7 configures the scheduling task. This step is performed by a computer. Specifically, the steps include: a) creating a timing task and periodically calling the instantiation model to perform fault diagnosis; b) if a diagnostic tool is set for an event node, the diagnostic process sequentially calls the diagnostic tool set for the node to obtain the alarm state and occurrence probability of the node and assigns the occurrence probability to the fault tree model; if a diagnostic tool is not set for an intermediate node, the occurrence probability of the node is estimated using a fault diagnosis method (minimum cut set, minimum path set, etc.) until the alarm state and occurrence probability of the top event are finally estimated.

[0051] Step 8 is to issue a fault alarm. This step is performed by the computer. Periodic fault diagnosis is performed according to the scheduling task configured in step 7, and an instance diagram of the fault diagnosis results, i.e., a complete FTA fault tree, is generated. This tree identifies the fault conditions and occurrence probabilities of each event node.

[0052] The above-mentioned embodiments are only preferred technical solutions of the present invention and should not be considered as limiting the present invention, and the protection scope of the present invention should cover the technical solutions described in the claims and equivalent alternatives including the technical features in the technical solutions described in the claims, that is, equivalent alternative improvements within this scope are also within the protection scope of the present invention.

Claims

[Claim 1] A method for automatically generating an FTA real-time dynamic tree based on a fault knowledge base, comprising: Step 1: constructing a new FTA logic model by generating a FTA real-time dynamic tree based on a fault knowledge base; Step 2: selecting a fault knowledge base model based on a logic device; Step 3: building a fault library based on the FTA logic model to generate logic rules for the FTA model, and decomposing and mapping the fault knowledge base model into the FTA logic fault tree model; Step 4: setting up a fault diagnosis tool for the event node; Step 5 of completing the construction of the FTA logical fault tree; Step 6 of publishing the FTA logic model and instantiating the FTA logic model; Step 7 of configuring the scheduling task; Step 8: periodically performing fault diagnosis and issuing a fault alarm according to the configured scheduling task; Including, In step 2, the logical device is an abstract description of a class of instantiated devices, and step 2 comprises: Step 2.1: filtering failure modes based on logic devices and selecting the failure mode that the current node matches; Step 2.2: Mapping and transforming the entire tree of failure modes into an FTA fault tree logic model; as a substep, In step 3, the fault library generates logic rules of the FTA model based on the fault knowledge base model selected in step 2, by mapping the top node to the top event of the FTA fault tree, the intermediate node to the intermediate event node of the FTA fault tree, and the end node to the basic event node of the FTA fault tree, where the top node, the intermediate node, and the end node are branch nodes in the FTA fault tree; When mapping to nodes, synchronously map the relationships of each hierarchical node, each hierarchical node including a top node, an intermediate node, and an end node; In step 4, the fault diagnosis tool includes measurement points, diagnosis rules, algorithm models and scripts; The principle of diagnosing the measurement point is to select the measurement point of the detection device under the logical device related to the node, and diagnose the occurrence probability and fault occurrence status of the node based on the data of the measurement point; The diagnostic principle of the diagnostic rule is to call the rule engine based on the business rule related to the logical device of the node to diagnose the occurrence probability and fault occurrence status of the node; The diagnostic principle of the algorithm model is to call the algorithm engine based on the hydropower-specific algorithm associated with the node's logical device to diagnose the occurrence probability and fault occurrence status of the node; The diagnostic principle of the script is to call a case library diagnostic algorithm based on the relevant cases in the fault case library related to the logical device of the node, to diagnose the occurrence probability and fault occurrence status of the node; In step 4, the purpose of configuring the fault diagnosis tools for the nodes of the fault tree is to perform dynamic analysis using the fault tree; Dynamic analysis is 1) A situation where a failure probability is predicted by a diagnostic tool at a measurement point of a detected device of a logic device. 2) A situation in which a rule diagnosis engine is called to predict the probability of a failure occurring for the business rules of a logical device; 3) If the amount of data at the measurement points of the device is greater than a preset threshold, an algorithm engine associated with the logical device is invoked to predict the node failure probability; Including some of the situations Step 6 is Step 6.1 of publishing the logic model and validating the newly created FTA fault tree logic model; Next, instantiation is performed by selecting a plant and an instantiation device tree corresponding to the model to be instantiated, and entering the model instantiation process. In the instantiation process, the logical devices bound to the nodes of the FTA fault tree logical model are converted into physical devices in the instantiation device tree of the plant, and the logical fault diagnosis tools bound to the nodes are converted into instantiation tools. Step 6.2 as a substep, Step 7 is Step 7.1 creating a timing task and periodically calling the instantiated model for fault diagnosis; Step 7.2 of the diagnosis process is to sequentially call the diagnostic tools set in the nodes when a diagnostic tool is set in the event node, obtain the alarm status and occurrence probability of the node, and substitute the occurrence probability into the fault tree model; if the diagnostic tool is not set in the intermediate node, estimate the occurrence probability of the node by the fault diagnosis method until finally estimating the alarm status and occurrence probability of the top event; as a substep, In step 8, the method for automatically generating an FTA real-time dynamic tree based on a fault knowledge base is characterized in that: the method performs periodic fault diagnosis on the scheduling task created in step 7 to generate a complete FTA fault tree, and the FTA fault tree identifies the fault conditions and occurrence probabilities of each event node.

Images