CN-116305567-B - Reliability simulation method and system for ship transmission shaft based on deep reinforcement learning

CN116305567BCN 116305567 BCN116305567 BCN 116305567BCN-116305567-B

Abstract

The invention discloses a reliability simulation method and a system of a ship transmission shaft based on deep reinforcement learning, and belongs to the field of ship transmission shaft design. The method organically combines a deep reinforcement learning algorithm with fault characteristics and reliability characteristic quantities, substitutes the reliability characteristic quantities of transmission shaft parts into a fault tree for quantitative analysis, calculates the state transition probability of each node, utilizes a Markov chain model to induce the state transition rule of each node fault mode to establish a system fault state transition probability formula, establishes a fault state space through fault mode analysis, qualitatively analyzes probability association degrees of bottom events and top events through the fault tree to form an action space, generates reward function weight factors of the Markov decision model through the mathematical relationship of the reliability characteristic quantities of each level unit, solves and generates a reliability simulation event list, timely reflects the reliability change condition of the transmission shaft, greatly improves the comprehensive prediction precision of the reliability of the transmission shaft, and ensures the reliability and the availability of the transmission shaft in operation.

Inventors

XIONG YAO
XU WEI
ZHANG WENJUN
BAI YAHE
ZHOU HAIBO
SUN YUPING
CUI XIAOLONG

Assignees

中国船舶集团有限公司第七一九研究所

Dates

Publication Date: 20260505
Application Date: 20230315

Claims (5)

1. The reliability simulation method of the ship transmission shaft based on deep reinforcement learning is characterized by comprising the following steps of: s1, determining a part set of a ship transmission shaft, a fault mode set of each part and a reliability characteristic quantity; S2, analyzing the reliability characteristic quantity of each part of the ship transmission shaft by utilizing FMEA, and determining the risk, task and time priority coefficient of each part fault mode; s3, constructing a transmission shaft fault mode fault tree taking a ship transmission shaft fault as a top event, and determining all minimum cut sets of the fault tree top event, wherein the minimum cut sets are all possible bottom events which cause the fault tree top event to occur, and any set of one bottom event is removed; S4, forming a state space by each node fault mode of the fault tree, forming an action space by changing each base event of all minimum cutsets from a fault state to a normal state, inducing the state change rule of each node fault mode by using a Markov chain model, and determining a system fault state transition probability formula; The fault mode set and the reliability characteristic quantity of each part are obtained by the following modes: (1) Carrying out dynamic simulation through the transmission load to obtain a fault mode of each part; (2) Inputting loads of different magnitudes, and determining the tensile stress and the shear stress born by each part in each failure mode according to the four-intensity theory of material mechanics; (3) Calculating the reliability characteristic quantity of the parts according to the failure resistance of the parts and the stress born by the parts : Wherein, the The degree of reliability of the time of day, The failure rate of the moment of time is calculated, For the average lifetime of the device, The reliability coefficient is determined by the failure resistance of the parts and the stress born by each part; The calculation formula of the reliability coefficient is as follows: Wherein, the Mathematical expectations of the resistance of the parts themselves to failure and the stresses to which the parts are subjected, respectively; The capacity of the parts against failure and the standard deviation of stress bearing of the parts are respectively; the step S2 comprises the following steps: s21, constructing an FMEA worksheet according to the reliability characteristic quantity of each part; s22, determining the severity level and the priority detection difficulty level of each part fault mode according to the fault mode and the FEMA worksheet; s23, calculating risk priority coefficients of fault modes of all parts Task priority coefficient And a time priority coefficient : Wherein, the As a probability of occurrence of the failure mode, In order to be of a level of severity, In order to prioritize the detection of the difficulty level, ; The value of the weight parameter is smaller than 0.01; The fault tree is built by 1) writing out a top event, wherein the top event is indicated as the least desirable fault event, the top event is indicated as a first row of the fault tree, 2) using a fault reason causing the top event as a second row of the fault tree to be indicated by a corresponding symbol and adopting a proper logic gate to be connected with the top event, 3) using the fault reason causing the second row of the fault event as a third row and using the fault gate to be connected with the second row, 4) tracing all bottom events of the system fault according to the layer-by-layer index of the reasons of the fault occurrence, thus forming an inverted fault tree taking the top event as a root, the middle event as a section, and the bottom event as a leaf.
2. The method of claim 1, wherein the system failure state transition probabilities are determined using a markov chain model and a probability of change of failure mode state for each node: Wherein, the Is the first The probability significance of an individual cell, As a probability of failure of the system, Is the first Probability of failure of individual cells.
3. The method of claim 1, wherein the reward function is: Wherein, the In order to maintain the amount of resources, To ensure the resource quantity.
4. A method according to any one of claims 1 to 3, wherein the set of parts of the ship's drive shaft comprises a thrust shaft, a tail shaft, a propeller shaft and an intermediate shaft.
5. The reliability simulation system of the ship transmission shaft based on deep reinforcement learning is characterized by comprising a processor and a memory; The memory is used for storing computer execution instructions; the processor configured to execute the computer-executable instructions such that the method of any one of claims 1 to 4 is performed.

Description

Reliability simulation method and system for ship transmission shaft based on deep reinforcement learning Technical Field The invention belongs to the field of ship transmission shaft design, and in particular relates to a reliability simulation method and system of a ship transmission shaft based on deep reinforcement learning. Background With the vigorous development of high and new technologies such as computer aided manufacturing, intelligent manufacturing and the like, the ship manufacturing technology and manufacturing process are rapidly developed, electronic, digital and intelligent equipment is continuously applied to ships, and the performance of ship power equipment is correspondingly required to meet higher requirements. The ship transmission shaft is used as a multi-task repairable complex system in terms of structure, function and performance, the relation among product design, structure, process and faults is determined, a reliability simulation analysis model is established, a mechanism or cause for causing failure is found out, defects caused by unreliable factors are detected to the greatest extent, and the reliability information of the system is more complete. The method is not only an important component of technical indexes of equipment tactics, but also a basis for equipment development, production and experiments of contractors. Therefore, in order to improve the operational efficiency and the guarantee capability of the transmission shaft of the ship power system, accurate and effective reliability modeling analysis should be synchronously carried out on the ship transmission shaft in the design and development processes. The ship transmission shaft has the function of transmitting power generated by the engine to the propeller shaft through the intermediate shaft, and the propeller shaft drives the propeller to generate thrust so as to push the ship body to advance. When the ship is operated, the transmission shaft system can generate three types of vibration, namely torsional vibration, longitudinal vibration and rotary vibration, the three types of vibration not only can seriously influence the operation of the transmission shaft system, but also can cause abnormal abrasion and fatigue damage of the transmission shaft, and the underwater low-frequency and multi-frequency strong line spectrum radiation noise is caused, so that the sound recessive energy is weakened sharply in the ship operation process, and further, fault events with different degrees are generated. On the other hand, under the influence of external environment, the reverse impact force on the ship transmission shaft system caused by the operation of the weapon system of the ship can greatly damage the reliability of the ship transmission shaft and cause the occurrence of fault events due to the impact load caused by underwater contact explosion and non-contact explosion. Therefore, it is important to analyze other process factors such as physicochemical processes, design defects, process defects and the like which directly cause faults or cause performance degradation and further develop into faults, and to qualitatively analyze and summarize the fault mechanism of the ship transmission shaft. Through analysis of fault mechanism, fault mode and fault influence can be accurately described, and fault periodic characteristic support is provided for reliability modeling analysis of ship transmission shafts. The transmission shafting of the ship power system is a complex mechanism system with the characteristics of complex structure, large power density span, complex and harsh working environment, variable operation working conditions, long power transmission path and the like. The random factors and fuzzy factors in the matching motion process of the units are affected, and the transmission impact caused by the motion of the units, the vibration impact caused by the severe change of external environment and the friction abrasion caused by the relative motion of the matched units can all bring different driving effects on the reliability change of the transmission system, so that fault events are frequently generated, and the resource occupation conflict is guaranteed. The method establishes a complete and reliable ship power system transmission shaft reliability simulation model, which has very important significance for improving the system operation reliability and researching the whole transmission system reliability simulation modeling technology. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a reliability simulation method and a system of a ship transmission shaft based on deep reinforcement learning, which aim at establishing a reliability analysis and evaluation model for the ship transmission shaft, improving the task success rate and the combat readiness integrity of transmission shaft equipment and solving the problem that the reliability of the transmission shaft ca