Search

CN-118644162-B - Aeroengine guarantee prediction method based on multi-agent modeling

CN118644162BCN 118644162 BCN118644162 BCN 118644162BCN-118644162-B

Abstract

The invention provides an aeroengine guarantee prediction method based on multi-agent modeling, which comprehensively considers main factors influencing the turnover backup quantity of aeroengines, provides an agent state machine model of the aeroengines aiming at the state variability characteristics of the aeroengines in the using and maintaining processes, reproduces various task scenes of the aeroengines from ordering, using, overhauling, fault returning and the like by utilizing the multi-agent simulation method, realizes the dynamic simulation of the using process of the aeroengines of the whole cluster, can predict the optimal reserve quantity of the aeroengines in a certain period in the future, namely, how many standby engines should be equipped for meeting the given combat readiness level of an airplane. Simulation results show that the invention can effectively simulate the guarantee process of the aircraft engines of the fleet and accurately acquire the number of spare engines to be equipped under the condition of given combat readiness integrity level.

Inventors

  • WU SHIHUI
  • ZHANG LIANG
  • Tang Xilang
  • Xiang Huachun
  • LI ZHENGXIN

Assignees

  • 中国人民解放军空军工程大学

Dates

Publication Date
20260505
Application Date
20240611

Claims (7)

  1. 1. The aircraft engine guarantee prediction method based on multi-agent modeling is characterized by comprising the following steps of: step 1, designing an aircraft engine intelligent body state machine model; the aeroengine intelligent body state machine model has 5 states, namely working, overhaul, fault factory return repair, backup and scrapping states; The transition model of each state of the aircraft engine agent state machine model comprises the following steps: (1) The transition of the aero-engine from the working state to the overhaul state is divided into two cases, namely, returning to the factory in advance for overhaul, and the service life of the engine reaches the stage; When the aeroengine has a large fault and needs to be subjected to advanced factory return overhaul, entering an overhaul state, wherein the stage residual life is changed to 0, and the stage life which is consumed in advance is counted; when the aero-engine reaches the stage life, entering a overhaul state; (2) Transition of the aero-engine from the working state to the fault factory-returning repair state: When the working engine has a fault requiring factory returning maintenance, factory returning maintenance is executed, the engine is returned after maintenance, and the service life of the engine at the stage after the engine is returned is not reduced; (3) The aircraft engine transitions from an operating state to a disabled state: when the service life of the engine reaches the total service life, scrapping the engine; (4) Aero-engine transitions from a overhaul state to a backup state: After a certain period of overhaul, the repaired engine is sent back to be used as a backup engine to enter a backup state; (5) The aeroengine transitions from the fault factory-return repair state to the backup state: After a certain fault repair period, the repaired engine is sent back to be used as a backup engine to enter a backup state; (6) The aeroengine transitions from the backup state to the operational state: when an event of lack of the engine occurs, the system broadcasts the lack message, and randomly selects one engine from the backup engines to accept the message so as to replace the lack engine to enter a working state; step 2, designing an aeroengine simulation driving event; Step 3, performing multi-agent-based aircraft engine guarantee process simulation according to the aircraft engine agent state machine model designed in the step 1 and the aircraft engine simulation driving event designed in the step 2, and obtaining the average satisfaction rate of the aircraft engine through simulation; And 4, adjusting the number of the standby engines, repeating the step3, and performing aircraft engine guarantee process simulation to finally obtain the number of the standby engines meeting the average meeting rate requirement of the aircraft engines.
  2. 2. The method for predicting the guarantee of the aeroengine based on multi-agent modeling as set forth in claim 1, wherein the simulation driving events of the aeroengine designed in the step 2 are a flight plan generating event, a mission receiving event and a mission distributing event, respectively.
  3. 3. The aircraft engine guarantee prediction method based on multi-agent modeling according to claim 2, wherein in step 2, the flight plan generation event is realized through the following processes: The method comprises the steps of firstly randomly generating a month flight plan of the airplane according to the predicted total annual flight hours of the airplane and a statistical rule, dividing the annual flight time into each month, and then randomly distributing the total flight time to each day in each month to form a daily flight hour schedule.
  4. 4. The method for predicting the guarantee of the aeroengine based on multi-agent modeling according to claim 2, wherein in the step 2, the mission-receiving event is realized by distributing the total time of flight of each day to each aircraft according to a statistical rule according to a daily time of flight schedule.
  5. 5. The aircraft engine guarantee prediction method based on multi-agent modeling according to claim 2, wherein in the step 2, the task allocation event is realized by adopting a triangular distribution function to represent the working hours of an engine when the aircraft is on a flight day, the working time of the engine is the set time consumption of starting inspection when the aircraft is on a maintenance day, and the working time of the engine is 0 when the aircraft is not on a flight or maintenance task or is in failure.
  6. 6. The aircraft engine guarantee prediction method based on multi-agent modeling according to claim 1, wherein in step 3, the simulation flow of the aircraft engine guarantee process based on the multi-agent is as follows: step 3.1, generating an aeroengine intelligent agent, initializing parameters according to set known conditions, setting an initial state for the aeroengine, and randomly generating daily flight hours of each aeroengine intelligent agent by utilizing the aeroengine simulation driving event designed in the step 2; step 3.2, simulating to be executed by taking a day as a time unit, wherein each aeroengine intelligent body generates state change, and counting the number of the missing fly frames in the day until a simulation ending condition is reached; And 3.3, recording and displaying a simulation result, and calculating the average satisfaction rate of the aero-engine according to a formula after the simulation is finished: 。
  7. 7. The method for predicting the guarantee of the aero-engine based on multi-agent modeling according to claim 1, wherein in step 3, through multiple simulations, the correlation between each parameter and the average satisfaction rate of the engine in the guarantee process of the aero-engine is obtained, and the guarantee parameters affecting the average satisfaction rate of the engine are determined.

Description

Aeroengine guarantee prediction method based on multi-agent modeling Technical Field The invention relates to the technical field of maintenance and guarantee of aeroengines, in particular to an aeroengine guarantee prediction method based on multi-agent modeling. Background Aero-engines are important conditions for ensuring that an aircraft can perform tasks normally. Once an aeroengine fails suddenly, the aircraft will not be able to complete the mission. Because aircraft engines are high value products that require a long time from ordering to outfitting, temporary ordering cannot be relied upon to solve the problem of lack of issue. The lack of an engine is classified into two categories, one is the foreseeable lack of an engine. If part of the engines are expected to reach the stage life in the year, the engine needs to be sent to the factory for overhaul at some time in the year, and the part of the engines are expected to reach the total life in the year, the engine needs to be eliminated in the year. The other is an unpredictable lack of hair. If the engine has a large fault (random), the engine needs to be returned to the factory for maintenance in advance. The number of these two engines together determines the number of standby engines that need to be planned in advance. Since aircraft are typically longer in life, engines are relatively shorter in life, which results in engines that are much more demanding and more costly to repair than aircraft. Enough standby aeroengines should be equipped to cope with sudden failures, supplementing the life of the engines. However, due to the complex and variable conditions of the shortage of the engine, the influencing factors include the quantity of the flight mission, the use and maintenance conditions of the airplane and the random faults of the engine, the number of spare engines with the reserved years is often excessive, the higher storage cost, the service life consumption of the spare engine and the maintenance cost are intangibly brought, and on the contrary, the number of spare engines with the reserved years is less, so that the problem of influencing the flight due to the shortage of the engine is brought. Therefore, the method adopts a scientific method, and the number of the engines which need to be prepared in a certain period in the future is analyzed and calculated from the macroscopic view, so that the method has important practical significance. Analyzing and calculating the number of engines to be prepared for a certain period in the future belongs to the problem of spare part demand prediction. For the problem of spare part demand prediction, many methods have been studied, but aeroengines belong to high-value products, and have long ordering period, and the use, consumption and maintenance flows of the aeroengines are greatly different from those of repairable spare parts, so that the traditional spare part calculation method cannot be simply used. The number of standby aircraft engines is studied at present mainly in the following aspects: The simulation method is used for analyzing, wherein the aero-engine is regarded as an important life spare part, the discrete event modeling method is adopted to describe the characteristics of an aero-engine operation system, the influence of factors such as different rejection rates, repair and transportation periods, timely replacement probabilities, average delay time and the like on the backup demand is researched, but the discrete event modeling simulation scale is smaller, the analysis is not carried out globally, and the influence of important indexes such as flight task quantity, the failure rate of the aero-engine and the like on the backup quantity is not considered. A calculation formula of the number of the standby aeroengines is given from the aspects of experience and statistics, wherein the calculation formula of the number of the standby aeroengines is obtained by analyzing main factors influencing the turnover backup number of the aeroengines, models such as the normal backup number of the engines, the initial backup number, the ordering number of the engines in the whole life cycle of the aircraft and the like are provided, but the statistical formula is too rough, and a larger error may exist in the result. From the angles of optimization and decision theory, a calculation model is constructed, namely an aeroengine backup decision method based on a scheduling plan is provided by taking the lowest maintenance and guarantee cost as an objective function, but the model construction is relatively thick, the data acquisition is problematic, and the actual application is difficult. Disclosure of Invention Aiming at the problems existing in the prior art, the invention provides an aeroengine guarantee prediction method based on multi-agent modeling, which comprehensively considers main factors influencing the turnover backup quantity of the aeroengine, and aims at the state variability