Maintenance decision method and device for multi-component system based on third-order state

Abstract

The present application discloses a maintenance decision method for a multi-component system based on a third-order state, constructs a state transition mechanism between the three states of a component, obtains the first state transition information, the first spare parts inventory, and the first maintenance requirement quantity of each component in the current detection system; determines the second state transition and state transition probability that may occur in the component during the next detection, determines the probability of the maintenance requirement of the component during the next detection, the second maintenance requirement quantity of the system during the next detection, and the probability of the system maintenance requirement; determines the second spare parts inventory during the next detection; determines the average cost rate within a unit detection cycle based on the system maintenance requirement probability and the second maintenance requirement quantity; uses the average cost rate as a fitness function, and uses the detection cycle and safety inventory as decision variables to determine the optimal detection cycle and optimal safety inventory of the system according to a genetic algorithm. The present application determines a maintenance strategy for joint optimization of system maintenance and spare parts inventory.

Description

Technical Field

[0001] The present application relates to the technical field of equipment maintenance decision-making, and in particular to a maintenance decision-making method and device for a multi-component system based on a third-order state. Background Art

[0002] In order to adapt to increasingly complex application scenarios and achieve richer functions, equipment is gradually developing towards integration. The various components are interconnected and influence each other to form a system to jointly complete complex and diverse tasks. In actual engineering practice, the performance of various equipment is gradually degraded due to the influence of materials, structural strength or surrounding environment. When the equipment degrades to a certain extent, it will malfunction or fail. For example, generators, stamping machines, CNC machine tools, etc. will malfunction due to performance degradation. The harm caused by equipment failure to enterprises is multi-faceted, multi-level and even long-term.

[0003] Condition-based maintenance is used in various occasions where maintenance is required. It emphasizes arranging preventive maintenance strategies based on the operating conditions of system components, and can make timely adjustments to the operating status of components. In the related art, for general multi-component systems, condition-based maintenance can achieve satisfactory results in system maintenance decisions composed of two-state components (including normal working state and functional failure state). However, in most cases, during the equipment degradation process, it exhibits three-state characteristics, such as normal working state, potential failure state, and functional failure state. When a component is in a potential failure state, it will incur higher operating costs than in a normal working state, and there is a higher risk of failure. Therefore, in engineering practice, the component should be avoided as much as possible in this state. The technical solutions related to the optimization of the condition-based maintenance strategy for three-state components mainly consider the situation on a single-component system, without considering the situation on a multi-component system.

[0004] Maintenance parts require spare parts, but spare parts inventory is not available instantly due to inventory cost, delivery time and other constraints. Maintenance and spare parts inventory affect and restrict each other. When making optimal maintenance decisions in a three-state component system, spare parts inventory is rarely considered. Summary of the invention

[0005] The purpose of this application is to provide a maintenance decision method for a multi-component system based on a third-order state, comprehensively consider the maintenance characteristics and spare parts characteristics of the components in the multi-component system, determine a strategy for the joint optimization of system maintenance and spare parts inventory, and effectively reduce production costs.

[0006] In a first aspect, the present application provides a maintenance decision method for a multi-component system based on a third-order state, wherein the multi-component system includes a plurality of identical components, and the method includes:

[0007] The states of components are defined as normal working state, potential failure state and functional failure state, and the cumulative distribution function of the components when they are transferred between various states is constructed, wherein components in potential failure state have preventive maintenance requirements, and components in functional failure state have functional maintenance requirements;

[0008] Acquire first state transition information, first spare parts inventory, and first maintenance requirement quantity of each component in the current detection system, wherein the first maintenance requirement quantity is the number of components with preventive maintenance requirements and the number of components with corrective maintenance requirements in the system;

[0009] For any component, based on the first state transition information and the cumulative distribution function of the component when it transitions between various states, determine the second state transition information and state transition probability of the component that may appear in the next detection, and determine the maintenance requirement probability of the component in the next detection according to the state transition probability, and then determine the second maintenance requirement quantity and system maintenance requirement probability of the multi-component system in the next detection;

[0010] Determine the second spare parts inventory quantity for the next inspection based on the first maintenance demand quantity, the first spare parts inventory quantity and the preset safety inventory quantity;

[0011] Determine the average total maintenance cost based on the system maintenance demand probability and the second maintenance demand quantity, determine the average total spare parts cost based on the system maintenance demand probability and the second spare parts inventory quantity, determine the average total operation cost based on the state transition probability, and determine the average expense rate within a unit inspection cycle based on the average total maintenance cost, the average total spare parts cost, the average total operation cost and the single inspection cost;

[0012] Taking the average cost rate as the fitness function, the inspection cycle and the safety stock as the decision variables, the optimal inspection cycle and the optimal safety stock of the multi-component system are determined according to the genetic algorithm.

[0013] In a second aspect, the present application provides a maintenance decision-making device for a multi-component system based on a third-order state, the device comprising:

[0014] A state transition module is used to define the states of components, including normal working state, potential fault state and functional fault state, and to construct a cumulative distribution function when components transition between various states, wherein components in a potential fault state have a preventive maintenance requirement, and components in a functional fault state have a functional maintenance requirement;

[0015] An acquisition module is used to acquire first state transition information, first spare parts inventory, and first maintenance requirement quantity of each component in the current detection system, wherein the first maintenance requirement quantity is the number of components with preventive maintenance requirements and the number of components with corrective maintenance requirements in the system;

[0016] A maintenance requirement probability module is used to determine, for any component, the second state transition information and the state transition probability of the component that may appear at the next detection based on the first state transition information and the cumulative distribution function of the component when it transitions between various states, and determine the maintenance requirement probability of the component at the next detection based on the state transition probability, and further determine the second maintenance requirement quantity and system maintenance requirement probability of the multi-component system at the next detection;

[0017] A spare parts inventory module, used to determine the second spare parts inventory at the next inspection based on the first maintenance demand quantity, the first spare parts inventory and the preset safety inventory;

[0018] A cost rate module, used to determine an average total maintenance cost based on the system maintenance demand probability and the second maintenance demand quantity, determine an average total spare parts cost based on the system maintenance demand probability and the second spare parts inventory quantity, determine an average total operating cost based on the state transition probability, and determine an average cost rate within a unit inspection cycle based on the average total maintenance cost, the average total spare parts cost, the average total operating cost and the single inspection cost;

[0019] The optimal model module is used to determine the optimal inspection cycle and the optimal safety stock of the multi-component system according to the genetic algorithm, taking the average cost rate as the fitness function, the inspection cycle and the safety stock as the decision variables.

[0020] This application considers the three-state characteristics of the components in the system for maintenance decisions on the same multi-component system, comprehensively considers the maintenance characteristics and spare parts characteristics of the components in the multi-component system, and determines a strategy for joint optimization of system maintenance and spare parts inventory, thereby improving the operating reliability of the multi-component system and effectively reducing production costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A first flow chart of a maintenance decision method for a multi-component system based on a third-order state provided in an embodiment of the present application;

[0022] Figure 2 A second flow chart of the maintenance decision-making method for a multi-component system based on a third-order state provided in an embodiment of the present application;

[0023] Figure 3 A third flow chart of the maintenance decision method for a multi-component system based on a third-order state provided in an embodiment of the present application;

[0024] Figure 4 A fourth flow chart of the maintenance decision method for a multi-component system based on a third-order state provided in an embodiment of the present application;

[0025] Figure 5A fifth flow chart of the maintenance decision method for a multi-component system based on a third-order state provided in an embodiment of the present application;

[0026] Figure 6 A sixth flow chart of a maintenance decision method for a multi-component system based on a third-order state provided in an embodiment of the present application;

[0027] Figure 7 A system block diagram of a maintenance decision-making device for a multi-component system based on a third-order state provided in an embodiment of the present application;

[0028] Figure 8 A system block diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0029] The present application will be described in detail below in conjunction with the specific implementation modes shown in the accompanying drawings, but these implementation modes do not limit the present application. Structural, methodological, or functional changes made by ordinary technicians in the field based on these implementation modes are included in the protection scope of the present application.

[0030] Please refer to Figure 1 The present application provides a maintenance decision method for a multi-component system based on a third-order state, wherein the multi-component system includes a plurality of identical components, and the method includes the steps of:

[0031] S101. Define the states of components, including normal working state, potential fault state and functional fault state, and construct a cumulative distribution function of the components when they are transferred between various states, wherein a component in a potential fault state has a preventive maintenance requirement, and a component in a functional fault state has a functional maintenance requirement;

[0032] S102, obtaining first state transition information, first spare parts inventory, and first maintenance requirement quantity of each component in the current detection system, wherein the first maintenance requirement quantity is the number of components with preventive maintenance requirements and the number of components with corrective maintenance requirements in the system;

[0033] S103, for any component, based on the first state transition information and the cumulative distribution function of the component when it transitions between various states, determine the second state transition information and the state transition probability of the component that may appear in the next detection, and determine the maintenance requirement probability of the component in the next detection according to the state transition probability, and then determine the second maintenance requirement quantity and system maintenance requirement probability of the multi-component system in the next detection;

[0034] S104, determining the second spare parts inventory for the next inspection based on the first maintenance demand quantity, the first spare parts inventory and the preset safety inventory;

[0035] S105, determining an average total maintenance cost based on the system maintenance demand probability and the second maintenance demand quantity, determining an average total spare parts cost based on the system maintenance demand probability and the second spare parts inventory, determining an average total operating cost based on the state transition probability, and determining an average expense rate within a unit inspection cycle based on the average total maintenance cost, the average total spare parts cost, the average total operating cost, and the single inspection cost;

[0036] S106. Taking the average cost rate as the fitness function, and the inspection period and the safety stock as decision variables, determine the optimal inspection period and the optimal safety stock of the multi-component system according to the genetic algorithm.

[0037] This implementation method targets a multi-component system containing multiple identical components, considers the three-state characteristics of the components, namely, normal working state, potential fault state and functional fault state, determines the transition mechanism of the components between various states, obtains the state transition of each component in the current detection system, determines the possible state transition of each component and its state transition probability at the next detection, determines the probability of maintenance requirement of a single component at the next detection, and then determines the number of maintenance requirements of the system at the next detection and its system maintenance requirement probability, determines the spare parts inventory at the next detection, and thus determines the average cost rate within a single detection cycle, and constructs an optimization model for the average cost rate, thereby determining the optimal detection cycle and optimal safety inventory of the system, comprehensively considering the maintenance characteristics and spare parts characteristics of the components in the multi-component system, and determining a strategy for the joint optimization of system maintenance and spare parts inventory, thereby improving the operating reliability of the multi-component system and effectively reducing production costs.

[0038] In the present application, a multi-component system includes a system with multiple identical components. For example, a wind farm includes multiple wind turbines with similar structures, and the wind turbines operate independently of each other. Multi-component systems share spare parts inventory, and any spare parts can be used to repair the components. The spare parts inventory only needs to consider the number of all spare parts in the system. Since the components are similar in structure and the same spare parts are used, when solving the maintenance needs of the system, only the number of components that require different types of repairs needs to be considered. The components of different component systems have different structures, and spare parts can only correspond one to one. The spare parts inventory of each component needs to be considered separately, and the maintenance of each component needs to be considered separately.

[0039] In some embodiments, the status of a component in a multi-component system is defined to include a normal working state, a potential fault state, and a functional fault state, that is, the normal working state is s0, the potential fault state is s1, and the functional fault state is s2. The potential fault state is a transition state between the normal working state and the functional fault state. When a component is in a potential fault state, the failure rate is significantly higher than that in a normal working state. A component can continue to operate in a potential fault state, but the operating cost is significantly higher than that in a normal working state.

[0040] In some embodiments, the time point at which the state transition of the component from s0 to s2 occurs is defined to obey an exponential distribution with a parameter λ0, and a first probability distribution density function is determined at the time when the state transition of the component from s0 to s2 occurs. , based on the first probability distribution density function Determine the first cumulative distribution function , where the first cumulative distribution function is the cumulative distribution function of the moment when the state transition of the component from s0 to s2 occurs, and the first probability distribution density function and the first cumulative distribution function Respectively expressed as:

[0041] ;

[0042] ;

[0043] in, is the moment when the state transition of the component from s0 to s2 occurs, is the probability of a component transferring from s0 to s2 at time t, λ0 is the exponential distribution parameter, For The probability of a component transferring from s0 to s2 within a time period;

[0044] Define the time point of the state transition from s1 to s2 of the component to obey the exponential distribution with parameter λ1, and determine the second probability distribution density function of the time when the state transition from s1 to s2 of the component occurs , based on the second probability distribution density function Determine the first cumulative distribution function , where the second cumulative distribution function is the cumulative distribution function of the moment when the state transition of the component from s1 to s2 occurs, and the second probability distribution density function and the second cumulative distribution function Respectively expressed as:

[0045] ;

[0046] ;

[0047] in, is the moment when the state of the component changes from s1 to s2, is the probability of a component transferring from s1 to s2 at time t, λ1 is the exponential distribution parameter, For The probability of a component transferring from s1 to s2 within a time period;

[0048] Define the time point of the state transition of the component from s0 to s1 to obey the Weibull distribution with α as the scale parameter and β as the shape parameter, and determine the third probability distribution density function of the moment when the state transition of the component from s0 to s1 occurs , based on the third probability distribution density function Determine the third cumulative distribution function , where the third cumulative distribution function is the cumulative distribution function of the moment when the state transition of the component from s0 to s1 occurs, and the third probability distribution density function and the third cumulative distribution function Respectively expressed as:

[0049] ;

[0050] ;

[0051] in, is the moment when the state transition of the component from s0 to s1 occurs, is the probability of a component transferring from s0 to s1 at time t, α is the scale parameter, β is the shape parameter, For The probability of a component transferring from s0 to s1 during the time period.

[0052] When a component is detected to be in a potential failure state, preventive maintenance is required, and when a component is detected to be in a functional failure state, corrective maintenance is required. In this embodiment, the state transition of a component or the state of a component is determined by periodic system detection. For example, if a component is found to be in a normal working state in the last detection, and it is found to be in a potential failure state in this detection, we can know that its state transition has occurred during this period. It is assumed that the repaired component is restored to be like new, that is, it is in a normal working state again, and each maintenance consumes a spare part. In this embodiment, based on the operating characteristics of the equipment and considering the three-state characteristics of the component, a three-stage state model of the component is established that conforms to the actual situation, and the state transition of the component is determined to be concentrated, which can fit the component degradation process and conform to the degradation mechanism of the component, thereby providing a basis for the joint decision-making of subsequent system maintenance and spare parts inventory.

[0053] Perform periodic nondestructive testing on all components in the system. Each testing cycle is dynamically adjusted according to the operating status of all components and the spare parts inventory. Based on the technical solution of joint decision-making on system maintenance needs and spare parts inventory in this application, the optimal testing cycle is determined as the next testing cycle. Therefore, the testing cycle is dynamically adjusted, which can better reduce production costs and has good engineering value for equipment maintenance decisions.

[0054] In some embodiments, at each inspection, the maintenance activities and spare parts ordering activities required are determined based on the status of each component and the spare parts inventory. Assume that at the kth inspection, the system maintenance demand is , the spare parts inventory is , Indicates that there is Parts require corrective repairs. Parts require preventive maintenance, and each maintenance activity incurs maintenance preparation costs , which is independent of the number of parts to be repaired. Assume that the system's safety stock is s. When the spare parts inventory cannot meet the maintenance demand or the remaining spare parts quantity is less than the safety stock, new spare parts must be ordered. The spare parts arrival time is The ordering cost of a single spare part is Spare parts inventory will generate corresponding spare parts holding costs. The spare parts holding cost per unit time for a single inventory spare part is When the spare parts inventory cannot fully meet the maintenance needs, spare parts are supplied first to the parts in a functional failure state. If the parts are not repaired in time, it may be necessary to shut down. The downtime loss cost per unit time when the parts perform the downtime operation is defined as The strategy for setting system maintenance and safety spare parts inventory is as follows:

[0055] like , Stock spare parts quantity If the corrective maintenance requirements are not met, spare parts are ordered immediately at the current inspection time. Faulty parts that have not undergone corrective maintenance will stop operating until the ordered spare parts arrive. Parts that require preventive maintenance will be scheduled for subsequent operation by assessing the risk: If the spare parts arrive at the time of If the expected cost of continuing to operate such parts is less than the expected cost of stopping operation within the specified period, such parts will continue to operate with degradation; otherwise, such parts will stop operating before the ordered spare parts arrive. All deferred maintenance will be completed immediately when the spare parts arrive;

[0056] like , all parts requiring corrective repair and remaining The parts that need preventive maintenance are repaired immediately. The spare parts are ordered immediately at the current inspection time. The arrangements for the remaining parts that need preventive maintenance are the same as above. All deferred maintenance will be completed immediately when the spare parts arrive;

[0057] like , the number of spare parts in stock can meet all maintenance needs at the same time, and the corresponding maintenance can be arranged immediately for all parts that need maintenance activities. If the remaining spare parts inventory after meeting the maintenance needs is less than the safety stock s, in order to reduce the spare parts holding cost, after this inspection point Order spare parts at all times, and do not order spare parts in other cases.

[0058] Based on the current detection, the first state transfer information, the first spare parts inventory and the first maintenance requirement quantity of each component in the system are obtained. The first maintenance requirement quantity is the number of components with preventive maintenance requirements and the number of components with corrective maintenance requirements in the system.

[0059] In some embodiments, Figure 2 As shown, based on the first state transition information and the cumulative distribution function of the component when it transitions between various states, determining the second state transition information and the state transition probability of the component that may appear in the next detection, includes:

[0060] If the component is in a normal state or fails and undergoes maintenance during the previous inspection, the possible second state transition includes the following four situations, where maintenance activities include preventive maintenance and functional maintenance:

[0061] S201, defining a situation where a component is in a normal state in the current detection and is in a normal state in the next detection as a first situation, and determining a first state transition probability when the first situation occurs according to a first cumulative distribution function and a third cumulative distribution function;

[0062] S202, defining the situation where the component is transferred from the normal state to the potential fault state in the current detection and transfers from the normal state to the potential fault state in the next detection as the second situation, and determining the second state transition probability when the second situation occurs according to the first cumulative distribution function and the second cumulative distribution function;

[0063] S203, defining the situation where the component is transferred from the normal state to the functional failure state in the current detection and is repaired, and is transferred from the normal state to the functional failure state in the next detection as a third situation, and determining the third state transition probability when the third situation occurs according to the first cumulative distribution function and the third cumulative distribution function;

[0064] S204. The situation where the component is transferred from a normal state to a potential fault state and then to a functional fault state during the current detection and repair is performed, and the situation where the component is transferred from a normal state to a potential fault state and then to a functional fault state during the next detection is defined as a fourth situation, and the fourth state transition probability when the fourth situation occurs is determined based on the first cumulative distribution function and the second cumulative distribution function.

[0065] A specific embodiment of the present application is described in The time point of the current test , get the status of each component in the system , the spare parts inventory is The maintenance requirement of the system is , which means the current detection time point have Parts require corrective repairs. A component needs preventive maintenance. Analyze the possible state transitions.

[0066] If the component is at time In normal state or maintenance activities are carried out, the possible second state transition includes the above four situations. Maintenance activities refer to preventive maintenance or corrective maintenance of the parts:

[0067] For the first case, the component is in normal state during the current detection cycle and does not need to be repaired in the next detection. The first state transition probability of the first case is expressed as for:

[0068] ;

[0069] in, For The probability of a component transferring from s0 to s2 within a time period; For The probability of a component transferring from s0 to s1 within a time period;

[0070] For the second case, the component changes state from s0 to s1 during the current detection cycle, but no failure occurs. Preventive maintenance is required for the next detection. The second state transition probability of the second case is expressed as for:

[0071] ;

[0072] in, Indicates the current detection time point. Indicates the next detection time point, Indicates the moment when the component transitions from state s0 to state s1. For components in The probability of transition from s0 to s1 at time instant, for The probability that no transfer from s0 to s2 occurs during the time period, for The probability that no transfer from s1 to s2 occurs during the time period;

[0073] For the third case, the component changes state from s0 to s2 during the current detection cycle, and the next detection requires corrective maintenance. The third state transition probability of the third case is expressed as for:

[0074] ;

[0075] in, Indicates the time when the state transition from s0 to s2 occurs. express The probability that no transfer from s0 to s1 occurs during the time period, express The probability of transition from s0 to s2 occurring at a given moment;

[0076] For the fourth case, the component changes state from s0 to s1 during the current detection cycle, and then changes state from s1 to s2. The next detection requires corrective maintenance. The fourth state transition probability of the fourth case is expressed as for:

[0077] ;

[0078] in, Indicates the time when the state transition from s1 to s2 occurs. express The probability of transition from s1 to s2 occurring at time instant.

[0079] In some embodiments, Figure 3 As shown, based on the first state transition information and the cumulative distribution function of the component when it transitions between various states, determining the second state transition information and the state transition probability of the component that may appear in the next detection, includes:

[0080] If a component fails during the previous inspection and needs to be repaired later due to a shortage of spare parts, and is repaired after the spare parts arrive, the second state transition that may occur during the interval between the spare parts arrival time and the next inspection of the component includes the following four situations;

[0081] S301, defining a situation where the component is in a normal state during a time period as a fifth situation, and determining a fifth state transition probability in the fifth situation according to the first cumulative distribution function and the third cumulative distribution function;

[0082] S302, defining a situation where a component transitions from a normal state to a potential fault state in a time period as a sixth situation, and determining a sixth state transition probability in the sixth situation according to the first cumulative distribution function and the second cumulative distribution function;

[0083] S303, defining a situation where a component transitions from a normal state to a functional failure state in a time period as a seventh situation, and determining a seventh state transition probability in the seventh situation according to the first cumulative distribution function and the third cumulative distribution function;

[0084] S304, defining the situation where the component transitions from a normal state to a potential fault state and then to a functional fault state in a time period as an eighth situation, and determining the eighth state transition probability in the eighth situation according to the first cumulative distribution function and the second cumulative distribution function.

[0085] If the component is When the system cannot be repaired in time due to a shortage of spare parts, it is necessary to decide whether to continue operation before the spare parts arrive based on the status of the parts. If the part is in a potential failure state, it is necessary to assess the risk and choose If the machine is in a functional failure state, it must be shut down within the time period or continue to operate in a potential failure state. The spare parts ordered are The part was delivered on time, and after delayed repair, it was restored to normal working condition. The fifth to eighth situations mentioned above may occur during the time period.

[0086] For the fifth case, the components are It is in normal working condition during the time period, and no maintenance is required for the next inspection. The fifth state transition probability of the fifth situation is expressed as for:

[0087] ;

[0088] in, Indicates the current detection time point. Indicates the next detection time point, T indicates the preset detection cycle, Indicates spare parts arrival time. express The probability that the component does not transfer from s0 to s2 during the time period, express The probability that the component does not transfer from s0 to s1 during the time period;

[0089] For the sixth case, the components are The state transition from s0 to s1 occurs during the time period, but no failure occurs. The next inspection requires preventive maintenance. The probability of the sixth state transition in the sixth case is expressed as for:

[0090] ;

[0091] in, Indicates the moment when the component transitions from state s0 to state s1. For components in The probability of transition from s0 to s1 at time instant, express The probability that the component does not transfer from s0 to s2 during the time period, for The probability that a component does not transfer from s1 to s2 during the time period.

[0092] For the seventh case, the components are The state transition from s0 to s2 occurs during the time period. The next inspection requires preventive maintenance. The probability of the seventh state transition in the seventh case is expressed as for:

[0093] ;

[0094] in, is the moment when the state transition of the component from s0 to s2 occurs, express The probability that a component does not transfer from s0 to s1 during the time period, express The probability of a component transferring from s0 to s2 at a given moment;

[0095] For the seventh case, the components are The state transition from s0 to s1 occurs during the time period, and then the state transition from s1 to s2 occurs. The next inspection requires corrective maintenance. The probability of the eighth state transition of the eighth situation is expressed as for:

[0096] ;

[0097] in, Indicates the moment when the component transitions from state s1 to state s2. Indicates the moment when the component transitions from state s0 to state s1. For components in The probability of transition from s0 to s1 at time instant, express The probability of a component transferring from s1 to s2 at a given moment, express The probability that a component does not transfer from s0 to s2 during the time period.

[0098] In some embodiments, based on the state transition probability of a single component obtained above, the maintenance requirement probability of the single component can be determined, that is, the maintenance requirement probability of the component at the next inspection is determined according to the state transition probability, including:

[0099] If the component is in normal working condition at the last inspection or fails and maintenance activities are performed, the first maintenance requirement probability that the component does not need maintenance at the next inspection is determined based on the first state transition probability, the second maintenance requirement probability that only one preventive maintenance is required for the component at the next inspection is determined based on the second state probability, and the third maintenance requirement probability that only one corrective maintenance is required for the component at the next inspection is determined based on the third state transition probability and the fourth state transition probability.

[0100] In a specific embodiment of the present application, if the component is in normal working state or fails and undergoes maintenance in the current detection, that is, the first to fourth cases mentioned above, the maintenance demand probability of a single component is analyzed for these four cases respectively. The components in a multi-component system with the same number of components M are numbered and represented as , if the component In the current inspection cycle, the component is in normal working condition or has undergone maintenance activities. The probability of the first repair need that may not require repair It is expressed as:

[0101] ;

[0102] If the parts In the current detection cycle, the state transition from s0 to s1 occurs, but no fault occurs. The next detection requires preventive maintenance. The component is at the next detection time. The probability of a second maintenance request requiring only one preventive maintenance It is expressed as;

[0103] ;

[0104] If the parts In the current detection cycle, a state transition from s0 to s2 occurs, or a state transition from s0 to s1 occurs, and then a state transition from s1 to s2 occurs. The next detection requires corrective maintenance, and the component is at the next detection time point. The probability of the third repair need requiring only one corrective repair It is expressed as:

[0105] .

[0106] In some embodiments, based on the state transition probability of a single component obtained above, the maintenance requirement probability of the single component can be determined, that is, the maintenance requirement probability of the component at the next inspection is determined according to the state transition probability, including:

[0107] If the component fails during the last inspection and maintenance needs to be delayed due to insufficient spare parts, and the repair is performed after the spare parts arrive, the fourth maintenance requirement probability that the component does not need maintenance during the next inspection is determined based on the fifth state transition probability, the fifth maintenance requirement probability that only preventive maintenance is required when the component is next inspected is determined based on the sixth state probability, and the sixth maintenance requirement probability that only corrective maintenance is required when the component is next inspected is determined based on the seventh state transition probability and the eighth state transition probability.

[0108] In a specific embodiment of the present application, if the component is When repairs cannot be performed due to lack of spare parts, the ordered spare parts will be The part was delivered on time, and after delayed repair, it was restored to normal working condition. The fifth to eighth situations mentioned above may occur in the time period. The maintenance demand probability of a single component is analyzed for these four situations respectively. The components in a multi-component system with the same number of components M are numbered and represented as If the component is The parts are in normal working condition during the period. At the next testing time The fourth repair demand probability that may not require repair It is expressed as:

[0109] ;

[0110] If the component is The state transition from s0 to s1 occurs during the time period, but no failure occurs. The next inspection requires preventive maintenance. Next detection time point The fifth maintenance requirement probability of only one preventive maintenance It is expressed as:

[0111] ;

[0112] If the parts If a state transition from s0 to s2 occurs during the current detection cycle, or a state transition from s0 to s1 occurs, and then a state transition from s1 to s2 occurs, the next detection requires corrective maintenance, and the component At the next testing time The probability of the sixth repair need requiring only one corrective repair It is expressed as:

[0113] .

[0114] In some embodiments, when calculating the maintenance requirement probability of a multi-component system, the multi-component system is divided into two subsystems, one of the components is used as one subsystem, and the remaining components are used as another subsystem. The maintenance requirement probability of the multi-component system is the superposition of the maintenance requirement probabilities of the subsystems, that is:

[0115] A multi-component system with the same number of components M is divided into a first subsystem composed of any component and a second subsystem composed of the remaining M-1 components, and the maintenance requirement probability of the first subsystem is calculated using the first maintenance requirement probability, the second maintenance requirement probability, the third maintenance requirement probability, the fourth maintenance requirement probability, the fifth maintenance requirement probability and the sixth maintenance requirement probability;

[0116] Divide the second subsystem into a third subsystem composed of any component and a fourth subsystem composed of the remaining M-2 components, and determine the maintenance demand probability of the third subsystem;

[0117] And so on, the second maintenance requirement quantity of the multi-component system at the next inspection is recursively obtained, as well as the first system maintenance requirement probability of not requiring maintenance, and the system maintenance requirement probability of needing m preventive maintenance and l corrective maintenance, wherein the second maintenance requirement quantity includes the number m of components for preventive maintenance and the number l of components for corrective maintenance.

[0118] For a multi-component system with M components, the maintenance demand is Repair needs and parts The combination of the maintenance requirements of a multi-component system with M-1 components is obtained, and the maintenance requirement probability recursive formula of the multi-component system with M components is obtained. The maintenance requirement probability of a single component can be determined by the above description. The maintenance requirement of a multi-component system with M-1 components can be decomposed by the same method as follows: Maintenance requirements and components for a multi-component system consisting of M-2 components The maintenance demand of M is obtained by recursion, and all possible maintenance demands and their maintenance demand probabilities of a multi-component system with M components are obtained.

[0119] A specific embodiment of the present application defines the next detection time of the multi-component system The maintenance demand is , that is, the number of parts requiring preventive maintenance is , the number of parts requiring corrective repair is , define a multi-component system consisting of M identical components composition,

[0120] when When the multi-component system The probability of system maintenance needing no maintenance at all times for:

[0121] ;

[0122] in, For components The multi-component system The probability of not needing maintenance at all times, For components exist The probability that maintenance is not required at all times;

[0123] when When the multi-component system is The probability of maintenance requirement for preventive maintenance for:

[0124] ;

[0125] in, For components The multi-component system Always needed The probability of preventive maintenance, For components A multi-component system requires The probability of preventive maintenance, For components exist The probability that only one preventive maintenance is required at any time, represents an exponential function. If condition A is met, then ,otherwise ;

[0126] when When the multi-component system Always The probability of needing repairs for corrective repairs for:

[0127] ;

[0128] in, For components The multi-component system Always needed The probability of corrective repairs, For components The multi-component system Always needed The probability of corrective repairs, For components exist The probability that only one corrective repair is required at any time;

[0129] when When the multi-component system Always Corrective repairs and The probability of maintenance requirement for preventive maintenance for:

[0130] ;

[0131] in, For components The multi-component system Always needed Corrective repairs and The probability of preventive maintenance, For components The multi-component system Always needed Corrective repairs and The probability of preventive maintenance, For components The multi-component system Always needed Corrective repairs and The probability of a preventive maintenance visit.

[0132] When the number of system components is known, the probability of state transition of a single component in a detection cycle is analyzed, and the obtained probability is substituted into the above expression of maintenance demand of the multi-component system to obtain the probability of system maintenance demand at the next detection.

[0133] In some embodiments, based on the status of each component in the system detected at the current time, the first maintenance demand quantity, the first spare parts inventory quantity, and the preset safety inventory quantity, determining the second spare parts inventory quantity for the next detection includes:

[0134] Define the first spare part inventory as , the first maintenance demand quantity The number of parts required for preventive maintenance in the system and the number of parts required for corrective repairs , the safety spare parts inventory is s;

[0135] like After the maintenance activity, all parts are in normal working condition. Since the spare parts inventory after meeting the maintenance demand is not less than the safety spare parts inventory, the spare parts ordering activity is not triggered. Therefore, the second spare parts inventory at the next inspection is ;

[0136] like After the maintenance activity, all parts are in normal working condition. Since the spare parts inventory is less than the safety spare parts inventory after meeting the maintenance demand, the spare parts ordering activity is triggered. After the spare parts arrive, delayed maintenance is performed. After the maintenance, the spare parts inventory is restored to the safety spare parts inventory, so the second spare parts inventory at the next inspection ;

[0137] like ,have A component cannot be repaired immediately due to insufficient spare parts, triggering a spare parts ordering activity. After the spare parts arrive, the repair is delayed. After the repair, the spare parts inventory is restored to the safe spare parts inventory, so the second spare parts inventory at the next inspection .

[0138] The total cost incurred during the inspection cycle is determined based on the average total maintenance cost, average total spare parts cost, average total operating cost and single inspection cost, and then the average cost rate per unit inspection cycle is determined.

[0139] In some embodiments, Figure 4 As shown, the average total maintenance cost is determined based on the system maintenance demand probability and the second maintenance demand quantity, including:

[0140] S401, determining an average maintenance preparation cost in the current detection cycle based on the system maintenance demand probability and the maintenance preparation cost generated by a single maintenance;

[0141] S402, determining an average maintenance activity cost in the current detection cycle based on the system maintenance demand probability, the cost of a preventive maintenance of a component, the cost of a corrective maintenance of a component, the second spare parts inventory, and the second maintenance demand quantity;

[0142] S403, determining an average downtime loss cost in the current detection cycle based on the system maintenance demand probability, the downtime cost of a single component per unit time, the spare parts arrival time, the second spare parts inventory, and the second maintenance demand quantity;

[0143] S404: Determine the average total maintenance cost in the current detection cycle based on the average maintenance preparation cost, the average maintenance activity cost, and the average downtime loss cost.

[0144] In a specific embodiment of the present application, the average maintenance preparation cost It represents the expected value of the maintenance preparation cost within a detection cycle. If there is no maintenance requirement, If all maintenance needs are completed in time, then no maintenance preparation cost is incurred. and When the maintenance cost is less than 10%, only the parts that need maintenance need to be repaired at the inspection point at the same time, and only one maintenance preparation cost is incurred. If the maintenance is delayed due to insufficient spare parts within the cycle, and , the delayed maintenance will increase the maintenance preparation cost, so the average maintenance preparation cost in the current inspection cycle is It is expressed as:

[0145] ;

[0146] in, Represents a multi-component system in Always Corrective repairs and The probability of maintenance requirement for preventive maintenance, Represents the repair preparation cost incurred for a single repair, which is independent of the number of parts repaired.

[0147] The average maintenance activity cost is the total cost of timely maintenance and delayed maintenance at the inspection point. The cost of a preventive maintenance of a component is The cost of a corrective repair is Generally speaking, the difficulty of repairing a component with a functional failure is much greater than that of repairing a component in a potential failure state. If there is no maintenance requirement at the time of inspection, If all maintenance needs can be completed in time, then no maintenance activity costs will be incurred. and When the maintenance cost is If the maintenance is delayed due to insufficient spare parts during the cycle, the delayed preventive maintenance is considered as corrective maintenance if it continues to operate. Therefore, the average maintenance activity cost during the current inspection cycle is expressed as:

[0148] ;

[0149] in, represents the cost of a preventive maintenance of a component, represents the cost of a corrective repair of a component, Indicates the second spare parts inventory, represents the expected cost of keeping the component in state s1 running due to insufficient spare parts inventory, including the running cost of the component and the cost of corrective repair when the spare parts arrive. The expected cost of parts downtime due to insufficient spare parts inventory, including the cost of downtime for the parts and the cost of preventive maintenance when the spare parts arrive, Indicates the number of parts that require preventive maintenance, Indicates the number of parts requiring corrective repair.

[0150] The average downtime loss cost refers to the cost of a component that needs to be repaired and is waiting for repair due to insufficient spare parts. The downtime cost of a single component per unit time is .when hour, A component in a malfunction state is stopped and waiting. The component in potential failure state is evaluated for risk and the operator is asked to decide whether to shut down and wait. The downtime is the delivery time of the spare parts. ;when hour, The component in potential failure state is evaluated for risk and the operator is asked to decide whether to shut down and wait. The downtime is the delivery time of the spare parts. , so the average downtime loss cost in the current detection cycle is It is expressed as:

[0151] ;

[0152] in, It represents the downtime cost of a single component per unit time. Indicates spare parts arrival time;

[0153] The average total maintenance cost includes three parts: average maintenance preparation cost, average maintenance activity cost, average downtime loss cost, and average total maintenance cost within the current inspection cycle. It is expressed as:

[0154] .

[0155] In some embodiments, Figure 5 As shown, the average total spare parts cost is determined based on the maintenance demand probability and the second spare parts inventory, including:

[0156] S501, determining an average spare parts ordering cost in the current detection cycle based on the system maintenance demand probability, the ordering cost of a single spare part, the second spare parts inventory, the preset safety inventory, and the second maintenance demand quantity;

[0157] S502, determining the average spare parts holding cost in the current detection cycle based on the system maintenance demand probability, spare parts arrival time, spare parts holding cost generated per unit time for a single inventory spare part, the second spare parts inventory, the preset safety inventory, and the second maintenance demand quantity;

[0158] S503: Determine the average total spare parts cost in the current inspection cycle based on the average spare parts ordering cost and the average spare parts holding cost.

[0159] In a specific embodiment of the present application, the average spare parts ordering cost is the expected value of the spare parts ordering cost within a detection cycle. When a spare part is ordered, it will incur ordering costs. The ordering cost of a single spare part is If spare parts cannot meet all demand, the order quantity should include the backorder quantity for delayed repairs. and newly added safety stock. If the spare parts can meet all the needs, the remaining quantity should be On the basis of Therefore, the average spare parts ordering cost in the current inspection cycle is It is expressed as:

[0160] ;

[0161] in, represents the ordering cost of a single spare part, represents the safety stock, Indicates the second spare part inventory quantity.

[0162] Average spare parts holding cost is the expected value of spare parts holding cost within a testing cycle. The spare parts holding cost per unit time of a single inventory spare part is .like , no need to order spare parts, the spare parts inventory is , holding time is T; if , the inventory of spare parts after maintenance is , holding time is T, due to Always order spare parts, spare parts are The spare parts ordered will not incur holding costs during this inspection period. , after the spare parts are repaired, the spare parts inventory is 0. Order spare parts at any time, the quantity ordered is , spare parts in When the time arrives, the spare parts inventory after delayed maintenance is s, and the holding time is , so the average spare parts holding cost during the current inspection cycle is It is expressed as:

[0163] ;

[0164] in, represents the ordering cost of a single spare part, T represents the inspection cycle, Represents the spare parts holding cost per unit time of a single inventory spare part;

[0165] The average total spare parts cost is the expected value of the total cost related to spare parts in a testing cycle, including the average spare parts ordering cost and the average spare parts holding cost. The average total spare parts cost in the current testing cycle It is expressed as:

[0166] .

[0167] In some embodiments, Figure 6 As shown, the average total operating cost is determined based on the state transition probability, including:

[0168] S601, define the cost of a single component operating in a normal state per unit time as a first cost, the cost of a single component operating in a potential failure state per unit time as a second cost, and the cost of a single component operating in a functional failure state per unit time as a third cost;

[0169] S602, if the component is in a normal state or fails and maintenance activities are performed in the next detection, determine a first expected operating time of the component in a normal state in the next detection based on the first cumulative distribution function, the third cumulative distribution function and the first state transition probability, determine a second expected operating time of the component in a potential failure state in the next detection based on the first cumulative distribution function, and determine a third expected operating time of the component in a functional failure state in the next detection based on the first cumulative distribution function;

[0170] S603. If the component fails to meet the maintenance requirement in the next inspection and is shut down within the spare parts arrival time, determine a fourth expected operating time of the component in a normal state in the next inspection based on the first cumulative distribution function, the second cumulative distribution function, the third cumulative distribution function and the fifth state transition probability, determine a fourth fifth expected operating time of the component in a potential failure state in the next inspection based on the first cumulative distribution function and the second cumulative distribution function, and determine a sixth expected operating time of the component in a functional failure state in the next inspection based on the first cumulative distribution function and the third cumulative distribution function;

[0171] S604. If the component fails to meet the maintenance requirement in the next detection and operates in a potential fault state within the spare parts arrival time, determine a seventh expected operation time for the component to be in a normal state in the next next detection based on the first cumulative distribution function, the second cumulative distribution function, and the third cumulative distribution function; determine an eighth expected operation time for the component to be in a potential fault state in the next next detection based on the first cumulative distribution function, the second cumulative distribution function, and the spare parts arrival time; and determine a ninth expected operation time for the component to be in a functional fault state in the next next detection based on the first cumulative distribution function and the second cumulative distribution function;

[0172] S605. Determine an average total operating cost based on the first cost, the second cost, the third cost, the first expected operating time, the second expected operating time, the third expected operating time, the fourth expected operating time, the fifth expected operating time, the sixth expected operating time, the seventh expected operating time, the eighth expected operating time and the ninth expected operating time.

[0173] In a specific embodiment of the present application, the average total operating cost is the expected value of the total operating cost within a detection cycle, that is, the operating cost generated when the component is in the three states s0, s1, and s2. The cost generated by a single component operating in a normal state per unit time is defined as the first cost The cost of a single component running in a potential failure state per unit time is the second cost. The cost of a single component running in a functional failure state per unit time is the third cost. , . Because in the next detection time period The state transfer mechanism within has not changed, and the same method as above can be used to analyze Situations that may occur within a time period and their probability.

[0174] Specifically, if the component is in a normal state or fails and undergoes maintenance in the next inspection, the component is determined to be in the next inspection time period based on the first cumulative distribution function, the third cumulative distribution function and the first state transition probability. The first expected running time in normal state It is expressed as:

[0175] ;

[0176] Where T represents the detection period, Indicates the next detection time point. Indicates the next detection time point, Indicates the moment when the component transitions from state s0 to state s1. represents the first state transition probability, For components in The probability of transition from s0 to s1 at time instant, for The probability that the component does not transfer from s0 to s2 during the time period, for The probability that the component does not transfer from s1 to s2 during the time period, Indicates the moment when the component transitions from state s0 to state s2. express The probability that a component does not transfer from s0 to s1 during the time period, express The probability of a component transferring from s0 to s2 at time instant, Indicates the moment when the component transitions from state s1 to state s2. express The probability of a component transitioning from s1 to s2 at a given moment.

[0177] A second expected operating time of the component in a potential failure state during the next detection is determined based on the first cumulative distribution function. The second expected operating time It is expressed as:

[0178] ;

[0179] A third expected operating time of the component being in a functional failure state in the next detection is determined based on the first cumulative distribution function. The third expected operating time It is expressed as:

[0180] ;

[0181] If the component fails to meet the maintenance requirements in the next inspection and is shut down within the spare parts arrival time, the fourth expected operating time of the component in the next inspection in a normal state is determined based on the first cumulative distribution function, the second cumulative distribution function, the third cumulative distribution function and the fifth state transition probability. for:

[0182] ;

[0183] in, Indicates the moment when the component transitions from state s0 to state s1. represents the fifth state transition probability, for The probability that the component does not transfer from s0 to s2 during the time period, for The probability that the component does not transfer from s1 to s2 during the time period, Indicates the moment when the component transitions from state s0 to state s2. express The probability that a component does not transfer from s0 to s1 during the time period.

[0184] Determine a fifth expected operating time of the component in a potential failure state at a next detection based on the first cumulative distribution function and the second cumulative distribution function. for:

[0185] ;

[0186] in, for The probability that the component does not transfer from s0 to s2 during the time period, for The probability that a component does not transfer from s1 to s2 during the time period.

[0187] Determine a sixth expected operating time of the component in a functional failure state at the next next detection based on the first cumulative distribution function and the third cumulative distribution function. for:

[0188] .

[0189] If the component fails to meet the maintenance requirements in the next inspection and operates in a potential fault state within the spare parts arrival time, the seventh expected operating time of the component in a normal state in the next inspection is determined based on the first cumulative distribution function, the second cumulative distribution function and the third cumulative distribution function. for:

[0190] ;

[0191] If the component does not meet the maintenance requirements in the next inspection and operates in a potential failure state within the spare parts arrival time, determine the eighth expected operating time of the component in the potential failure state in the next inspection based on the first cumulative distribution function, the second cumulative distribution function and the spare parts arrival time. for:

[0192] ;

[0193] If the component fails to meet the maintenance requirements in the next inspection and operates in a potential failure state within the spare parts arrival time, the ninth expected operation time of the component in the functional failure state in the next inspection is determined based on the first cumulative distribution function and the second cumulative distribution function. for:

[0194] ;

[0195] Average total operating cost It is expressed as:

[0196] ;

[0197] in, represents the first cost, represents the second cost, Represents the third cost.

[0198] Average total cost during the current inspection cycle for:

[0199] ;

[0200] in, Represents the cost of a single test.

[0201] Above The solution is based on the calculation of various possible maintenance requirements and their probabilities, and the calculation of maintenance requirements and their probabilities requires solving the state transitions and probabilities that may occur in a single component within a detection cycle.

[0202] An optimization model is established with the minimum average cost rate as the goal and the detection cycle and safety inventory as decision variables. It can be expressed as:

[0203] ;

[0204] in, is the relationship function between the cost rate and the decision variables.

[0205] In some embodiments, the maintenance requirements and spare parts inventory of the system at the next inspection are obtained, and a maintenance decision is made for the next inspection. Time or Always order spare parts to The lowest average cost rate within the time period is taken as the goal, and the optimal joint strategy is obtained by using the genetic algorithm.

[0206] Exemplarily, real number coding is used to encode individuals in the population, and each individual represents a joint strategy, that is, a combination of a detection cycle and a safety stock. Different individuals represent different maintenance strategies as their genotypes. After this, an initial population is produced, and an initial population is constructed by randomly producing a group of individuals (i.e., different joint strategies). The population size is set according to the complexity of the specific problem. In this embodiment, the population size is set to 50 to ensure population diversity. The average cost rate is used as the fitness function. Each individual represents a different joint strategy. When calculating the fitness function of an individual, the corresponding joint strategy is used, that is, a combination of a detection cycle and a safety stock. The detection cycle and safety stock corresponding to the individual are used as decision variables and substituted into the above-mentioned average cost rate model to obtain the specific value of the average cost rate. The smaller the average cost rate, the higher the fitness. After the fitness function is determined, the selection link is entered. Individuals are selected to enter the next generation according to the fitness value. The probability of each individual being selected is proportional to its fitness, and individuals with high fitness are more likely to be selected. New individuals are produced by combining partial genetic information of two individuals. The goal of the crossover operation is to produce new individuals with higher fitness. The mutation operation is to introduce new genes and increase the diversity of the population, thereby avoiding premature convergence. Mutation will randomly change the genome of an individual. The value of the detection cycle and safety stock corresponding to the individual can be randomly increased or decreased. The probability of mutation is set low, at 0.02. The iterative process is the core of the genetic algorithm. Each generation will undergo selection, crossover and mutation operations to produce a new population. In each generation, individuals with higher fitness have a greater chance of reproduction, allowing better joint strategies to continue. As the iteration proceeds, the average fitness of individuals in the population will gradually increase and approach the optimal solution. When the number of iterations reaches 20, the algorithm iteration ends. The genotype of the optimal individual in the last generation of the population is the optimal joint decision sought, that is, the combination of the optimal detection cycle and safety stock, which minimizes the average cost rate.

[0207] Based on the same inventive concept, the embodiment of the present application also provides a maintenance decision device for a multi-component system based on a third-order state for implementing the maintenance decision method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more device embodiments provided below can refer to the limitations on the maintenance decision method above, and will not be repeated here.

[0208] Please refer to Figure 7 The present application provides a maintenance decision device for a multi-component system based on a third-order state, comprising:

[0209] The state transition module 701 is used to define the states of the components, including normal working state, potential fault state and functional fault state, and construct the cumulative distribution function of the components when they transition between various states, wherein the components in the potential fault state have preventive maintenance requirements, and the components in the functional fault state have functional maintenance requirements;

[0210] The acquisition module 702 is used to acquire the first state transition information, the first spare parts inventory and the first maintenance requirement quantity of each component in the current detection system, wherein the first maintenance requirement quantity is the number of components with preventive maintenance requirements and the number of components with corrective maintenance requirements in the system;

[0211] The maintenance requirement probability module 703 is used to determine, for any component, the second state transition information and the state transition probability of the component that may appear at the next detection based on the first state transition information and the cumulative distribution function of the component when it transitions between various states, and determine the maintenance requirement probability of the component at the next detection based on the state transition probability, and further determine the second maintenance requirement quantity and system maintenance requirement probability of the multi-component system at the next detection;

[0212] The spare parts inventory module 704 is used to determine the second spare parts inventory at the next inspection based on the first maintenance demand quantity, the first spare parts inventory and the preset safety inventory;

[0213] Cost rate module 705, used to determine the average total maintenance cost based on the system maintenance demand probability and the second maintenance demand quantity, determine the average total spare parts cost based on the system maintenance demand probability and the second spare parts inventory, determine the average total operation cost based on the state transition probability, and determine the average cost rate within the unit inspection cycle based on the average total maintenance cost, the average total spare parts cost, the average total operation cost and the single inspection cost;

[0214] The optimal model module 706 is used to determine the optimal inspection cycle and the optimal safety stock of the multi-component system according to the genetic algorithm, using the average cost rate as the fitness function, the inspection cycle and the safety stock as decision variables.

[0215] The present application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned maintenance decision-making method when executing the computer program. Figure 8803 is a hardware structure diagram of a computer device provided in an embodiment of the present application, including: a processor 801, a communication interface 802, a memory 803 and a communication bus 804, wherein the processor 801, the communication interface 802 and the memory 803 complete mutual communication through the communication bus 804. The memory 803 can be used to store a computer program, which may include instructions and data to implement the steps of any of the above methods. The memory 803 may be a random access memory, a read-only memory, a non-volatile, a programmable ROM, an erasable PROM, an electrically erasable, a flash memory, an optical memory and a register, etc. The processor 801 may be a general-purpose processor, which may be a processor that performs specific steps and / or operations by reading and executing a computer program stored in the memory, and the general-purpose processor may use data stored in the memory during the execution of the steps and / or operations. The general-purpose processor may be a central processing unit, an ASIC and an FPGA, etc. The communication interface 802 may include an input / output interface, a physical interface and a logical interface, etc., which are used to implement the interfaces for interconnecting devices within the network device. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware or software instructions in the processor 801. The method disclosed in the embodiment of the present application can be directly embodied as a hardware processor, or a combination of hardware and software modules in the processor.

Claims

1. A maintenance decision method for a multi-component system based on a third-order state, wherein the multi-component system comprises a plurality of identical components, characterized in that: The method comprises the steps of: The states of the components are defined to include a normal working state, a potential fault state and a functional fault state, and a cumulative distribution function of the components when they are transferred between various states is constructed, wherein the components in the potential fault state have a preventive maintenance requirement, and the components in the functional fault state have a functional maintenance requirement; Acquire first state transition information, first spare parts inventory, and first maintenance requirement quantity of each component in the current detection system, wherein the first maintenance requirement quantity is the number of components with preventive maintenance requirements and the number of components with corrective maintenance requirements in the system; For any component, based on the first state transition information and the cumulative distribution function of the component when it transitions between various states, determine the second state transition information and state transition probability of the component that may appear in the next detection, and determine the maintenance requirement probability of the component in the next detection according to the state transition probability, and then determine the second maintenance requirement quantity and system maintenance requirement probability of the multi-component system in the next detection; Determine the second spare parts inventory for the next inspection based on the first maintenance demand quantity, the first spare parts inventory and the preset safety inventory; Determine an average total maintenance cost based on the system maintenance demand probability and the second maintenance demand quantity, determine an average total spare parts cost based on the system maintenance demand probability and the second spare parts inventory, determine an average total operating cost based on the state transition probability, and determine an average expense rate within a unit inspection cycle based on the average total maintenance cost, the average total spare parts cost, the average total operating cost and the single inspection cost; Taking the average cost rate as the fitness function, taking the inspection period and the safety stock as decision variables, determining the optimal inspection period and the optimal safety stock of the multi-component system according to the genetic algorithm; Wherein, based on the first maintenance demand quantity, the first spare parts inventory quantity and the preset safety inventory quantity, determining the second spare parts inventory quantity for the next inspection includes: Define the first spare part inventory as , the first maintenance demand quantity The number of parts required for preventive maintenance in the system and the number of parts required for corrective repairs , the safety spare parts inventory is s; like , the second spare parts inventory ; like , the second spare parts inventory ; like , the second spare parts inventory .

2. The method according to claim 1, characterized in that The determining, based on the first state transition information and the cumulative distribution function of the component when it transitions between various states, the second state transition information and the state transition probability of the component that may appear during the next detection includes: Determine a first cumulative distribution function of the time when a state transition of a component from a normal working state to a functional failure state occurs, a second cumulative distribution function of the time when a state transition from a potential failure state to a functional failure state occurs, and a third cumulative distribution function of the time when a state transition from a normal working state to a potential failure state occurs; If the component is in normal working state or fails and undergoes maintenance during the current inspection, the possible second state transition is determined to include the following four situations, wherein the maintenance activities include preventive maintenance and functional maintenance: A situation where the component is in a normal working state in the current detection and is in a normal working state in the next detection is defined as a first situation, and a first state transition probability when the first situation occurs is determined according to the first cumulative distribution function and the third cumulative distribution function; The situation where the component is transferred from the normal working state to the potential fault state in the current detection and is transferred from the normal working state to the potential fault state in the next detection is defined as a second situation, and the second state transition probability when the second situation occurs is determined according to the first cumulative distribution function and the second cumulative distribution function; The situation where the component is transferred from the normal working state to the functional failure state in the current detection and is repaired, and is transferred from the normal working state to the functional failure state in the next detection is defined as a third situation, and the third state transition probability when the third situation occurs is determined according to the first cumulative distribution function and the third cumulative distribution function; A situation where the component is transferred from a normal working state to a potential fault state and then to a functional fault state during the current detection and is repaired, and is transferred from a normal working state to a potential fault state and then to a functional fault state during the next detection is defined as a fourth situation, and a fourth state transition probability when the fourth situation occurs is determined according to the first cumulative distribution function and the second cumulative distribution function; and / or, If a component fails during the previous inspection and needs to be repaired later due to a shortage of spare parts, and is repaired after the spare parts arrive, the second state transition that may occur during the interval between the spare parts arrival time and the next inspection of the component includes the following four situations; A situation where the component is in a normal working state during the time period is defined as a fifth situation, and a fifth state transition probability when the fifth situation occurs is determined according to the first cumulative distribution function and the third cumulative distribution function; A situation where the component changes from a normal working state to a potential fault state during the time period is defined as a sixth situation, and a sixth state transition probability when the sixth situation occurs is determined according to the first cumulative distribution function and the second cumulative distribution function; A situation where the component changes from a normal working state to a functional failure state during the time period is defined as a seventh situation, and a seventh state transition probability when the seventh situation occurs is determined according to the first cumulative distribution function and the third cumulative distribution function; The situation where the component transfers from a normal working state to a potential fault state and then to a functional fault state during the time period is defined as the eighth situation, and the eighth state transition probability when the eighth situation occurs is determined based on the first cumulative distribution function and the second cumulative distribution function.

3. The method according to claim 2, characterized in that Construct the cumulative distribution function when the component transitions between various states, including: Define the normal working state of the component as s0, the potential fault state as s1, and the functional fault state as s2; Define the time point of the state transition of the component from s0 to s2 to obey the exponential distribution with parameter λ0, and determine the first probability distribution density function of the time when the state transition of the component from s0 to s2 occurs , based on the first probability distribution density function Determine the first cumulative distribution function The first cumulative distribution function is the cumulative distribution function of the moment when the state transition of the component from s0 to s2 occurs, and the first probability distribution density function and the first cumulative distribution function Respectively expressed as: ； ； Define the time point of the state transition from s1 to s2 of the component to obey the exponential distribution with parameter λ1, and determine the second probability distribution density function of the time when the state transition from s1 to s2 of the component occurs , based on the second probability distribution density function Determine the first cumulative distribution function , wherein the second cumulative distribution function is the cumulative distribution function of the moment when the state transition of the component from s1 to s2 occurs, and the second probability distribution density function and the second cumulative distribution function Respectively expressed as: ； ； Define the time point of the state transition of the component from s0 to s1 to obey the Weibull distribution with α as the scale parameter and β as the shape parameter, and determine the third probability distribution density function of the moment when the state transition of the component from s0 to s1 occurs , the third probability distribution density function Determine the third cumulative distribution function , wherein the third cumulative distribution function is the cumulative distribution function of the moment when the state transition of the component from s0 to s1 occurs, and the third probability distribution density function and the third cumulative distribution function Respectively expressed as: ； ； Determine the second state transition information and state transition probability of the component that may appear during the next detection, including: The first state transition probability representation for: ； in, For The probability of a component transferring from s0 to s2 within a time period; For The probability of a component transferring from s0 to s1 within a time period; The second state transition probability is expressed as for; ； in, Indicates the current detection time point. Indicates the next detection time point, Indicates the moment when the component transitions from state s0 to state s1. For components in The probability of transition from s0 to s1 at time instant, For The probability that the component does not transfer from s0 to s2 during the time period, For The probability that the component does not transfer from s1 to s2 during the time period; The third state transition probability representation for: ； in, Indicates the time when the state transition from s0 to s2 occurs. express The probability that a component does not transfer from s0 to s1 during the time period, express The probability of a component transferring from s0 to s2 at a given moment; The fourth state transition probability representation for: ； in, Indicates the moment when the component transitions from state s1 to state s2. express The probability of a component transferring from s1 to s2 at a given moment; Fifth state transition probability representation for: ； Wherein, T represents the preset detection period, Indicates spare parts arrival time. express The probability that the component does not transfer from s0 to s2 during the time period, express The probability that the component does not transfer from s0 to s1 during the time period; The sixth state transition probability representation for: ； in, express The probability that the component does not transfer from s0 to s2 during the time period, for The probability that the component does not transfer from s1 to s2 during the time period; The seventh state transition probability representation for: ； in, express The probability that the component does not transfer from s0 to s1 during the time period; Eighth state transition probability representation for: ； in, express The probability that a component does not transfer from s0 to s2 during the time period.

4. The method according to claim 2, characterized in that: The determining, according to the state transition probability, the probability of maintenance requirement of the component at the next detection includes: If the component is in a normal working state or fails and undergoes maintenance activities at the time of the current inspection, a first maintenance requirement probability that the component does not need maintenance at the next inspection is determined based on the first state transition probability, a second maintenance requirement probability that only one preventive maintenance is required at the next inspection of the component is determined based on the second state probability, and a third maintenance requirement probability that only one corrective maintenance is required at the next inspection of the component is determined based on the third state transition probability and the fourth state transition probability; and / or, If a component fails during the current inspection and needs to be repaired later due to a shortage of spare parts, and the repair is performed after the spare parts arrive, a fourth repair requirement probability that the component does not need to be repaired during the next inspection is determined based on the fifth state transition probability, a fifth repair requirement probability that only one preventive maintenance is required during the next inspection of the component is determined based on the sixth state probability, and a sixth repair requirement probability that only one corrective maintenance is required during the next inspection of the component is determined based on the seventh state transition probability and the eighth state transition probability; The determining of the second maintenance requirement quantity and maintenance requirement probability of the multi-component system during the next inspection includes: A multi-component system with the same number of components M is divided into a first subsystem composed of any component and a second subsystem composed of the remaining M-1 components, and the maintenance requirement probability of the first subsystem is calculated using the first maintenance requirement probability, the second maintenance requirement probability, the third maintenance requirement probability, the fourth maintenance requirement probability, the fifth maintenance requirement probability and the sixth maintenance requirement probability; Divide the second subsystem into a third subsystem composed of any component and a fourth subsystem composed of the remaining M-2 components, and determine the maintenance demand probability of the third subsystem; By analogy, the second maintenance requirement of the multi-component system at the next inspection is recursively obtained, as well as the number of maintenance operations that need to be performed. Preventive maintenance and The probability of a system maintenance requirement of a corrective maintenance, wherein the second maintenance requirement quantity includes the number of parts for preventive maintenance and the number of parts required for corrective repairs .

5. The method according to claim 3, characterized in that: part The probability of first maintenance demand It is expressed as: ； part The second maintenance demand probability It is expressed as; ； part The third maintenance demand probability It is expressed as: ； part The fourth maintenance demand probability It is expressed as: ； part The fifth maintenance demand probability It is expressed as: ； part The sixth maintenance demand probability It is expressed as: ； Define the next detection time The number of parts requiring preventive maintenance is , the number of parts requiring corrective repair is , define a multi-component system consisting of M identical components composition; when When the multi-component system The probability of system maintenance needing no maintenance at all times for: ； in, For components The multi-component system The probability of not needing maintenance at all times, For components exist The probability that maintenance is not required at any time; when When the multi-component system is The probability of maintenance requirement for preventive maintenance for: ； in, For components The multi-component system Always needed The probability of preventive maintenance, For components A multi-component system requires The probability of preventive maintenance, For components exist The probability that only one preventive maintenance is required at any time, represents an exponential function. If condition A is met, then ,otherwise ; when When the multi-component system Always The probability of needing repairs for corrective repairs for: ； in, For components The multi-component system Always needed The probability of corrective repairs, For components The multi-component system Always needed The probability of corrective repairs, For components exist The probability that only one corrective repair is required at any time; when When the multi-component system Always Corrective repairs and The probability of maintenance requirement for preventive maintenance for: ； in, For components The multi-component system Always needed Corrective repairs and The probability of preventive maintenance, For components The multi-component system Always needed Corrective repairs and The probability of preventive maintenance, For components The multi-component system Always needed Corrective repairs and The probability of a preventive maintenance visit.

6. The method according to claim 5, characterized in that The determining of the average total maintenance cost based on the maintenance requirement probability and the second maintenance requirement quantity includes: Determine an average maintenance preparation cost based on the system maintenance demand probability and the maintenance preparation cost generated by a single maintenance; Determine an average maintenance activity cost based on the system maintenance demand probability, the cost of performing a preventive maintenance on the component, the cost of performing a corrective maintenance on the component, the second spare parts inventory, and the second maintenance demand quantity; Determine an average downtime loss cost based on the system maintenance demand probability, the downtime cost of a single component per unit time, the spare parts arrival time, the second spare parts inventory, and the second maintenance demand quantity; Determine an average total maintenance cost based on the average maintenance preparation cost, the average maintenance activity cost and the average downtime loss cost; The determining the average total spare parts cost based on the maintenance demand probability and the second spare parts inventory comprises: Determine an average spare parts ordering cost based on the system maintenance demand probability, the ordering cost of a single spare part, the second spare parts inventory, the preset safety inventory, and the second maintenance demand quantity; Determine the average spare parts holding cost based on the system maintenance demand probability, spare parts arrival time, spare parts holding cost generated per unit time for a single inventory spare part, the second spare parts inventory, the preset safety inventory and the second maintenance demand quantity; determining an average total spare parts cost based on the average spare parts ordering cost and the average spare parts holding cost; The determining of the average total operating cost based on the state transition probability includes: Define the cost of a single component operating in a normal working state per unit time as the first cost, the cost of a single component operating in a potential failure state per unit time as the second cost, and the cost of a single component operating in a functional failure state per unit time as the third cost; If the component is in normal working condition at the next detection or fails and maintenance activities are performed, determine a first expected operating time for the component to be in normal working condition at the next next detection based on the first cumulative distribution function and the third cumulative distribution function, determine a second expected operating time for the component to be in potential failure condition at the next next detection based on the first cumulative distribution function, and determine a third expected operating time for the component to be in functional failure condition at the next next detection based on the first cumulative distribution function; If the component fails to meet the maintenance requirements in the next inspection and is shut down within the spare parts arrival time, determine a fourth expected operating time for the component to be in a normal working state in the next inspection based on the first cumulative distribution function, the second cumulative distribution function, and the third cumulative distribution function, determine a fourth-fifth expected operating time for the component to be in a potential failure state in the next inspection based on the first cumulative distribution function and the second cumulative distribution function, and determine a sixth expected operating time for the component to be in a functional failure state in the next inspection based on the first cumulative distribution function and the third cumulative distribution function; If the component fails to meet the maintenance requirements in the next inspection and operates in a potential fault state within the spare parts arrival time, determine a seventh expected operating time for the component to be in a normal working state in the next inspection based on the first cumulative distribution function, the second cumulative distribution function, and the third cumulative distribution function; determine an eighth expected operating time for the component to be in a potential fault state in the next inspection based on the first cumulative distribution function, the second cumulative distribution function, and the spare parts arrival time; and determine a ninth expected operating time for the component to be in a functional fault state in the next inspection based on the first cumulative distribution function and the second cumulative distribution function; An average total operating cost is determined based on the first cost, the second cost, the third cost, the first expected operating time, the second expected operating time, the third expected operating time, the fourth expected operating time, the fifth expected operating time, the sixth expected operating time, the seventh expected operating time, the eighth expected operating time and the ninth expected operating time.

7. The method according to claim 6, characterized in that Average repair preparation cost It is expressed as: ； in, Represents a multi-component system in Always Corrective repairs and The probability of maintenance requirement for preventive maintenance, It represents the maintenance preparation cost incurred by a single maintenance; The average maintenance activity cost is expressed as: ； in, represents the cost of a preventive maintenance of a component, represents the cost of a corrective repair of a component, Indicates the second spare parts inventory, represents the expected cost of keeping the component in state s1 running due to insufficient spare parts inventory, represents the expected cost of parts downtime due to insufficient spare parts inventory, Indicates the number of parts that require preventive maintenance, Indicates the number of parts requiring corrective repair; Average cost of downtime It is expressed as: ； in, It represents the downtime cost of a single component per unit time. Indicates spare parts arrival time; Average total repair cost It is expressed as: ； Average spare parts ordering cost It is expressed as: ； in, represents the ordering cost of a single spare part, Indicates the safety stock amount; Average spare parts holding cost It is expressed as: ； in, represents the ordering cost of a single spare part, T represents the inspection cycle, Represents the spare parts holding cost per unit time of a single inventory spare part; Average total spare parts cost It is expressed as: 。 8. The method according to claim 7, characterized in that First expected running time It is expressed as: ； in, represents the first state transition probability, for The probability that the component does not transfer from s0 to s2 during the time period, for The probability that the component does not transfer from s1 to s2 during the time period; Second expected running time It is expressed as: ； Third expected running time It is expressed as: ； Fourth expected running time It is expressed as: ； in, represents the fifth state transition probability, for The probability that the component does not transfer from s0 to s2 during the time period, for The probability that the component does not transfer from s1 to s2 during the time period, express The probability that the component does not transfer from s0 to s1 during the time period; Fifth expected running time It is expressed as: ； in, for The probability that the component does not transfer from s0 to s2 during the time period, for The probability that the component does not transfer from s1 to s2 during the time period; Sixth expected running time It is expressed as: ； Expected running time It is expressed as: ； 8th Expected Run Time It is expressed as: ； 9. Expected Run Time It is expressed as: ； The average total operating cost is expressed as: ； in, represents the first cost, represents the second cost, represents the third cost; Average total cost during the current inspection cycle for: ； in, Represents the cost of a single test.

9. A maintenance decision device for a multi-component system based on a third-order state, characterized in that: The device comprises: A state transfer module is used to define the states of components, including normal working state, potential fault state and functional fault state, and to construct a cumulative distribution function when the components transfer between various states, wherein the components in the potential fault state have preventive maintenance requirements, and the components in the functional fault state have functional maintenance requirements; An acquisition module, used to acquire first state transition information, first spare parts inventory, and first maintenance requirement quantity of each component in the current detection system, wherein the first maintenance requirement quantity is the number of components with preventive maintenance requirements and the number of components with corrective maintenance requirements in the system; A maintenance requirement probability module is used to determine, for any component, the second state transition information and the state transition probability of the component that may appear at the next detection based on the first state transition information and the cumulative distribution function of the component when it transitions between various states, and determine the maintenance requirement probability of the component at the next detection based on the state transition probability, and further determine the second maintenance requirement quantity and system maintenance requirement probability of the multi-component system at the next detection; A spare parts inventory module, used to determine the second spare parts inventory at the next inspection based on the first maintenance demand quantity, the first spare parts inventory and the preset safety inventory; A cost rate module, used to determine an average total maintenance cost based on the system maintenance demand probability and the second maintenance demand quantity, determine an average total spare parts cost based on the system maintenance demand probability and the second spare parts inventory, determine an average total operating cost based on the state transition probability, and determine an average cost rate within a unit inspection cycle based on the average total maintenance cost, the average total spare parts cost, the average total operating cost and the single inspection cost; The optimal model module is used to determine the optimal inspection cycle and the optimal safety stock of the multi-component system according to the genetic algorithm, taking the average cost rate as the fitness function and the inspection cycle and the safety stock as the decision variables; Wherein, the spare parts inventory module is specifically used for: Define the first spare part inventory as , the first maintenance demand quantity The number of parts required for preventive maintenance in the system and the number of parts required for corrective repairs , the safety spare parts inventory is s; like , the second spare parts inventory ; like , the second spare parts inventory ; like , the second spare parts inventory .

Images