Fault tree expert system-based ship equipment abnormity diagnosis system and method

Abstract

The invention discloses a ship equipment abnormity diagnosis system based on a fault tree expert system, and the system comprises the steps: constructing an expert knowledge base module: designing an expert knowledge base, decomposing a top event into an intermediate event, and decomposing the intermediate event into a terminal event; performing qualitative analysis by using a downlink method, and determining a fault minimum combination; quantifying a fault influence weight; logically converting the fault tree model to form an expert knowledge base of a generative rule; constructing a database module: adopting a distributed architecture for an expert knowledge base, and providing data support for an inference engine; constructing a knowledge reasoning engine and an interpreter module, wherein a core mechanism of the reasoning engine adopts a forward reasoning strategy, and expert knowledge base rules are matched according to levels; a natural language interpretation is generated based on the inference path. The invention further discloses a ship equipment abnormity diagnosis method based on the fault tree expert system. According to the invention, rapid diagnosis and early warning of ship system faults are realized, the operation and maintenance reliability and intelligent level are improved, and the method can be widely applied to the technical field of ship automatic operation and maintenance.

Description

Technical Field

[0001] This invention relates to the field of automated ship maintenance technology, and in particular to a ship equipment anomaly diagnosis system and method based on a fault tree expert system. Background Technology

[0002] With the development of the global shipping industry, modern ship systems are characterized by high integration and automation. The coordinated operation of subsystems such as propulsion, power, and navigation relies on complex electromechanical coupling and information interaction. In long-term, high-load operating environments, potential failures in any subsystem can trigger a chain reaction, leading to interruptions in maintenance tasks or even major safety accidents.

[0003] Fault Tree Analysis (FTA), a core method in safety systems engineering, achieves causal analysis of faults through logical deduction from top events to terminal events. However, it faces two major bottlenecks in marine applications: first, marine systems are highly susceptible to environmental interference and have limited sensor deployment, leading to insufficient accuracy in fault tree modeling; second, traditional FTA can only perform qualitative or simple quantitative analysis, failing to adapt to the scenarios of multiple coupled faults and dynamic evolution in ship operation and maintenance. To address these issues, existing technologies attempt to combine expert systems with fault trees, such as marine engine diagnostic systems based on confidence rule sets and fuzzy neural network remote diagnostic architectures. While these approaches improve intelligence, they still suffer from drawbacks such as reliance on expert experience for knowledge acquisition, lack of data-driven reasoning processes, and insufficient prioritization of critical faults.

[0004] Furthermore, existing ship maintenance expert systems often rely on a single rule base for knowledge representation, making it difficult to comprehensively map the logical relationships of "AND gates" and "OR gates" in fault trees. Their reasoning strategies also lack quantitative analysis of fault importance, resulting in low diagnostic efficiency. Therefore, how to deeply integrate the structured analysis of fault trees with the intelligent reasoning of expert systems to achieve rapid fault location and priority diagnosis has become a pressing technical challenge in the field of automated ship maintenance. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the above-mentioned background technology and provide a ship equipment anomaly diagnosis system and method based on fault tree expert system. It aims to achieve rapid location, priority diagnosis and early warning of ship system faults through deep integration of fault tree model and expert system, thereby improving operation and maintenance reliability and intelligence level.

[0006] This invention provides a ship equipment anomaly diagnosis system based on a fault tree expert system, comprising the following components: An expert knowledge base construction module: The expert knowledge base is designed with the integrity of the ship's "detection-control-actuator" task resources as the top event. It is decomposed into intermediate events composed of equipment-level and system-level faults using AND and OR gates, and further decomposed into several terminal events. A descending method is used to complete qualitative analysis, and the minimum combination of faults is determined by expanding the logic gates layer by layer. The impact weight of faults is quantified based on the probability of terminal event occurrence. The fault tree model logic is transformed into production rules, forming an expert knowledge base of production rules. A database construction module: The expert knowledge base is designed with a distributed architecture, storing ship data in real time, supporting database queries and data visualization analysis, and providing dynamic data support for the knowledge reasoning engine. A knowledge reasoning engine and interpreter module: The core mechanism of the reasoning engine adopts a forward reasoning strategy, triggered from the top event, matching expert knowledge base rules according to the "top event → intermediate event → terminal event" hierarchy. Natural language explanations are generated based on the reasoning path, representing the causal relationship of the fault.

[0007] In the above technical solution, in the module for constructing an expert knowledge base, the intermediate events are further decomposed into 11 terminal events, which are the 11 minimum cut sets in the fault tree model. The quantification elements of the fault impact weight include the formula-calculated top event probability, structural importance, probabilistic importance, and critical importance.

[0008] In the above technical solution, the real-time stored ship data in the constructed database module includes ship operation data and fault history records. The fault history records include timestamps, fault types, processing results, and model parameters. The model parameters include terminal event probability thresholds.

[0009] In the above technical solution, the specific process of the forward reasoning strategy in the knowledge reasoning engine and interpreter module includes four stages: feature extraction, rule matching, iterative derivation, and result verification.

[0010] In the above technical solution, the reasoning process of the forward reasoning strategy includes the feature extraction stage, which is the fault data preprocessing process; the rule matching stage, which is the condition and fact comparison process; the iterative derivation stage, which is the reasoning process if the result is not the minimum cut set; and the result verification stage, which is the comparison process with historical fault data.

[0011] In the above technical solution, during the rule matching stage of constructing the knowledge reasoning engine and interpreter module, i.e. the process of comparing conditions with facts, when a fault signal is received, the terminal events with high critical importance are checked first.

[0012] In the above technical solution, the process of generating natural language explanations based on reasoning paths in the knowledge reasoning engine and interpreter module is located in the iterative derivation stage. The process of generating natural language explanations based on reasoning paths is based on rule matching work, first deriving device-level faults or system-level faults as intermediate events, and finally locating the top event that occurred.

[0013] The above technical solution also includes a human-machine interface module and a communication interface module. The human-machine interface module uses a graphical design and integrates a fault data input unit and a diagnostic result visualization unit. The communication interface module uses TCP / IP (Transmission Control Protocol / Internet Protocol) to interface with the ship's local area network, supports OPCUA (OPC Unified Architecture, a vendor-independent industrial automation communication protocol where OPC uses OLE / COM as the communication standard for applications) and Modbus (an industrial communication protocol), and enables data interaction with sensors, PLCs (Programmable Logic Controllers), and remote monitoring centers. The interface has data encryption and breakpoint resume functions, supports external system call requests, and enables cross-platform sharing of fault diagnosis results.

[0014] This invention also provides a method for diagnosing ship equipment anomalies based on a fault tree expert system, comprising the following steps: S1, the communication interface acquires real-time data from sensors via the Modbus protocol, stores it in a database, and then the inference engine calls rules from the expert knowledge base for matching, retrieving the corresponding sensor data for analysis; S2, a fault tree of the top event T is constructed, connecting device-level faults or system-level faults through AND gates or OR gates, and M1 is then connected through AND gates or OR gates to one or more of the following: control device faults, actuator faults, detection device faults, communication faults, or execution accuracy faults, using a downlink method for analysis. Eleven minimal cut sets are generated by analyzing and generating them. The priority of fault events is obtained by checking the critical importance of each event, and the events that the inference engine should prioritize for diagnosis are determined. S3: The inference engine identifies fault events based on the identified data and selects rules from the expert knowledge base. If the selected rule matches the fault event, it further determines whether it is a terminal event. Otherwise, it reselects rules. If it is determined to be one of the terminal events in the fault tree after the judgment, the result is output and the inference ends. If no rule is matched, no diagnosis result is output, the rules in the expert knowledge base are supplemented, and the inference engine ends.

[0015] In the above technical solution, in steps S1 to S3, the information interaction process in each step is conducted through the communication interface of the ship's local area network accessed by the TCP / IP protocol. The communication interface integrates OPCUA and Modbus industrial protocols to realize data interaction with sensors, PLCs and remote monitoring centers. It has data encryption and breakpoint resume functions, supports external system calls, and realizes cross-platform sharing of fault diagnosis results. In step S2, the downlink method analysis is implemented by Python script.

[0016] The present invention relates to a ship equipment anomaly diagnosis system and method based on a fault tree expert system, which has the following beneficial effects: 1. This invention shortens the fault location time compared to traditional methods by prioritizing critical importance reasoning. Combined with quantitative fault tree analysis, it achieves extremely high diagnostic accuracy and can handle complex faults, thus improving both diagnostic efficiency and accuracy.

[0017] 2. The fault tree model of the present invention automatically generates production rules, reduces reliance on a large amount of expert experience, supports real-time data iterative updates of the knowledge base, adapts to the needs of ship system upgrades, and improves the ability to acquire knowledge automatically.

[0018] 3. The distributed database and open communication interface designed in this invention support access from multiple ship types and multiple systems (such as propulsion systems and power systems). The inference engine can dynamically adjust the diagnostic strategy according to the ship's operating conditions, enhancing dynamic adaptability and scalability.

[0019] 4. This invention provides full-process visual decision support. The human-computer interaction interface displays the reasoning path and fault probability in real time, and the interpreter provides logical tracing, which reduces the technical threshold for operation and maintenance personnel and greatly improves the efficiency of maintenance guidance.

[0020] 5. Deeply integrate the structured logic of fault tree analysis with the intelligent reasoning of expert systems to break through the application bottleneck of traditional diagnostic methods in complex systems and form a closed-loop technology system of "modeling-analysis-reasoning-verification". Attached Figure Description

[0021] Figure 1 This is an overall architecture diagram of the ship equipment anomaly diagnosis system based on a fault tree expert system according to the present invention; Figure 2 This is a model diagram of the fault tree in the ship equipment anomaly diagnosis system based on the fault tree expert system of the present invention; Figure 3 This is a functional diagram illustrating the knowledge reasoning engine and interpreter module in the ship equipment anomaly diagnosis system based on the fault tree expert system of the present invention. Figure 4This is a schematic diagram of the architecture of the communication interface module in the ship equipment anomaly diagnosis system based on the fault tree expert system of the present invention. Figure 5 This is a flowchart illustrating the ship equipment anomaly diagnosis method based on a fault tree expert system according to the present invention. Detailed Implementation

[0022] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments, but these embodiments should not be construed as limiting the present invention.

[0023] To address the problems of low diagnostic efficiency, poor dynamic adaptability, and lack of intelligent reasoning in traditional ship fault analysis methods, the technical solution of this invention involves fault modeling, knowledge base construction, intelligent reasoning, and fault location technology for ship mission resource systems.

[0024] Example 1 See Figure 1 The present invention relates to a ship equipment anomaly diagnosis system based on a fault tree expert system, comprising the following parts: (1) Constructing a human-computer interaction interface module. The human-computer interaction interface adopts a graphical design and integrates a fault data input module and a diagnostic result visualization module.

[0025] See Figure 2 (2) Construct an expert knowledge base and database. Design an expert knowledge base, taking the integrity of the ship's "detection-control-actuator" task resources as the top event (T), and decompose it into intermediate events such as equipment-level failure (M1) and system-level failure (M2) through "AND gate" or "OR gate", and further decompose it into 11 terminal events (X1-X11, forming a fault tree model containing 11 minimal cut sets, such as {X1, X2}, {X3}, etc.; use the down-step method to complete the qualitative analysis, and determine the minimum combination of failures by expanding layer by layer through logic gates; based on the probability of the occurrence of terminal events, use formulas to calculate the probability of the top event, structural importance, probability importance and key Importance is quantified by assigning weights to the impact of faults; the fault tree logic is transformed into production rules, such as "IF actuator equipment failure THEEN excessive vibration" or "IF task resource integrity failure THEEN equipment-level failure OR system-level failure," forming a knowledge base containing 50+ rules. Simultaneously, the knowledge base needs to be designed with a distributed architecture to store real-time ship operation data, fault history records (including timestamps, fault types, and processing results), and model parameters such as terminal event probability thresholds. It should support SQL queries and data visualization analysis, providing dynamic data support for the inference engine.

[0026] (3) Construct a knowledge reasoning engine and interpreter. The core mechanism of the reasoning engine adopts a forward reasoning strategy, starting from the top event and matching knowledge base rules according to the hierarchy of "top event → intermediate event → terminal event". For example, when a "task resource failure" signal is received, terminal events with high critical importance are checked first. The reasoning process includes four stages: feature extraction (fault data preprocessing), rule matching (condition and fact comparison), iterative deduction (if the result is not the minimum cut set, continue reasoning), and result verification (comparison with historical fault data). Natural language explanations are generated based on the reasoning path, such as "due to the detection of abnormal vibration, combined with the rule 'IF actuator equipment failure THEN abnormal vibration', the actuator equipment failure is deduced to be an intermediate event, and finally located to the equipment-level failure that caused the top event to occur", which helps maintenance personnel understand the causal relationship of the fault.

[0027] Example 2 This embodiment is basically the same as Embodiment 1, except that: (4) Design of the communication interface. The communication interface uses the TCP / IP protocol to interface with the ship's local area network and supports industrial communication protocols such as OPCUA and Modbus to realize data interaction with sensors, PLC controllers, and remote monitoring centers. In this embodiment, the sensor is one or more of a temperature sensor, pressure sensor, liquid level sensor, or vibration sensor. The interface has data encryption and breakpoint resume functions to ensure the reliability of real-time data transmission, and also supports external system call requests to realize cross-platform sharing of fault diagnosis results.

[0028] Example 3 The purpose of this invention is to propose a method for diagnosing ship equipment anomalies based on a fault tree expert system, comprising the following steps, such as... Figure 5 As shown: S1. The communication interface acquires real-time data from the vibration sensor via the Modbus protocol. After being stored in the database, the inference engine calls rules from the expert knowledge base for matching. For example, when a "task resource failure" signal is received, the inference engine... Figure 2 The process retrieves and analyzes the corresponding sensor data.

[0029] S2, Press Figure 2 A fault tree is constructed: the top event T is "fault in mission resource integrity," connected to M1 (device-level fault) and M2 (system-level fault) via an OR gate. M1 is then connected to control device fault (M3), actuator fault (M4), etc., via an OR gate. A descending analysis method is implemented using Python scripts, generating 11 minimal cut sets. By examining the critical importance of each event, the priority of fault events can be obtained, determining the events that the inference engine should prioritize for diagnosis.

[0030] S3. When the sensor receives fault data, the inference engine needs to identify the fault event based on the data and select rules from the knowledge base. If the selected rule matches the fault event, it further determines whether it is a terminal event. Otherwise, it reselects a rule. If the judgment determines that it is one of the terminal events in the fault tree, it outputs the result and ends the inference. Otherwise, it reselects a rule. If none of the rules match, it outputs no diagnostic result, needs to supplement the rule knowledge base, and ends the inference engine.

[0031] Figure 4 The diagram shows the communication interface architecture. This communication interface connects to the ship's local area network via TCP / IP protocol and integrates industrial protocols such as OPCUA and Modbus to achieve data interaction with sensors, PLCs, and remote monitoring centers. It features data encryption and breakpoint resume capabilities, ensuring reliable real-time transmission while supporting external system calls and enabling cross-platform sharing of fault diagnosis results.

[0032] The advantages of this invention lie in its intelligent diagnosis of ship maintenance faults, enabling rapid fault location and improving diagnostic efficiency and accuracy. It provides a systematic solution for enhancing ship maintenance reliability and optimizing efficiency. Traditional fault analysis methods, such as Fault Mode and Effects Analysis (FMEA) and Event Tree Analysis (ETA), rely on human experience and lack dynamic reasoning capabilities, making them insufficient for the rapid location and early warning needs of complex systems.

[0033] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

[0034] The contents not described in detail in this specification are existing technologies known to those skilled in the art.

Claims

1. A ship equipment anomaly diagnosis system based on a fault tree expert system, characterized in that: Includes the following parts: Construct an expert knowledge base module: Design an expert knowledge base with the integrity of ship "detection-control-actuator" mission resources as the top event. Using AND and OR gates, decompose the knowledge base into intermediate events consisting of equipment-level and system-level faults, further decomposing them into several terminal events. A descending method is used to complete qualitative analysis, determining the minimum combination of faults through layer-by-layer expansion using logic gates. Based on the probability of terminal event occurrence, quantify the impact weight of faults. Transform the fault tree model logic into production rules, forming an expert knowledge base of production rules. Database module construction: The expert knowledge base is designed with a distributed architecture to store ship data in real time, support database queries and data visualization analysis, and provide dynamic data support for the knowledge reasoning engine; Construct a knowledge reasoning engine and an interpreter module: The core mechanism of the reasoning engine adopts a forward reasoning strategy, which is triggered from the top event and matches the expert knowledge base rules in the hierarchy of "top event → intermediate event → terminal event"; it generates natural language explanations based on the reasoning path to represent the causal relationship of the fault.

2. The ship equipment anomaly diagnosis system based on fault tree expert system according to claim 1, characterized in that: In the module for constructing the expert knowledge base, the intermediate events are further decomposed into 11 terminal events, which are the 11 minimum cut sets in the fault tree model. The quantification elements of the fault impact weight include the formula-calculated top event probability, structural importance, probabilistic importance, and critical importance.

3. The ship equipment anomaly diagnosis system based on fault tree expert system according to claim 2, characterized in that: The constructed database module stores real-time ship data including ship operation data and fault history records. The fault history records include timestamps, fault types, processing results, and model parameters. The model parameters include terminal event probability thresholds.

4. The ship equipment anomaly diagnosis system based on fault tree expert system according to claim 3, characterized in that: The specific process of the forward reasoning strategy in the knowledge reasoning engine and interpreter module includes four stages: feature extraction, rule matching, iterative derivation, and result verification.

5. The ship equipment anomaly diagnosis system based on fault tree expert system according to claim 4, characterized in that: In the reasoning process of the forward reasoning strategy, the feature extraction stage is the fault data preprocessing process, the rule matching stage is the condition and fact comparison process, the iterative derivation stage is the reasoning process if the result is not the minimum cut set, and the result verification stage is the comparison process with historical fault data.

6. The ship equipment anomaly diagnosis system based on fault tree expert system according to claim 5, characterized in that: In the rule matching phase of constructing the knowledge reasoning engine and interpreter module, i.e. the process of comparing conditions with facts, when a fault signal is received, the terminal events with high critical importance are checked first.

7. The ship equipment anomaly diagnosis system based on fault tree expert system according to claim 6, characterized in that: In the construction of the knowledge reasoning engine and interpreter module, the process of generating natural language explanations based on reasoning paths is located in the iterative derivation stage. The process of generating natural language explanations based on reasoning paths is based on rule matching work, first deriving device-level faults or system-level faults as intermediate events, and finally locating the top event that occurred.

8. The ship equipment anomaly diagnosis system based on fault tree expert system according to claim 7, characterized in that: It also includes building a human-computer interaction interface module and a communication interface module. Human-computer interaction interface module: The human-computer interaction interface adopts a graphical design and integrates a fault data input unit and a diagnostic result visualization unit. Communication Interface Module: The communication interface uses the TCP / IP protocol to connect with the ship's local area network, supports OPCUA and Modbus industrial communication protocols, and enables data interaction with sensors, PLC controllers and remote monitoring centers. The interface has data encryption and breakpoint resume functions, supports external system call requests, and realizes cross-platform sharing of fault diagnosis results.

9. A method for diagnosing ship equipment anomalies based on a fault tree expert system, characterized in that: Includes the following steps: S1. The communication interface acquires real-time data from the sensor via the Modbus protocol. After being stored in the database, the inference engine calls the rules in the expert knowledge base for matching and retrieves the corresponding sensor data for analysis. S2. Construct the fault tree of the top event T, and connect the device-level faults or system-level faults through "AND gate" or "OR gate". M1 is then connected to one or more of the following faults through "AND gate" or "OR gate": control device faults, actuator faults, detection device faults, communication faults or execution accuracy faults. Analyze using the downlink method and generate 11 minimal cut sets. By checking the critical importance of each event, the priority of the fault events is obtained, and the events that the inference engine should prioritize to check during diagnosis are determined. S3. Based on the identified data, the inference engine identifies fault events and selects rules from the expert knowledge base. If the selected rule matches the fault event, it further determines whether it is a terminal event. Otherwise, it reselects rules. If the judgment determines that it is one of the terminal events in the fault tree, it outputs the result and ends the inference. If none of the rules match, it outputs no diagnostic result, supplements the rules in the expert knowledge base, and ends the inference engine.

10. The method for diagnosing ship equipment anomalies based on a fault tree expert system according to claim 9, characterized in that: In steps S1 to S3, the information exchange process in each step is conducted through the communication interface of the ship's local area network accessed by the TCP / IP protocol. The communication interface integrates OPCUA and Modbus industrial protocols to realize data exchange with sensors, PLCs and remote monitoring centers. It has data encryption and breakpoint resume functions, supports external system calls, and realizes cross-platform sharing of fault diagnosis results. In step S2, the downlink analysis is implemented using a Python script.

Images