Aircraft maintenance scheme optimization analysis method and system

Abstract

The invention discloses an aircraft maintenance scheme optimization analysis method and system, and relates to the technical field of maintenance optimization, and the method comprises the following steps: building a maintenance demand prediction model of aircraft parts based on a random forest regression model; carrying out fault mode and effect analysis on key parts of the aircraft, and calculating a fault risk priority number to quantify the risk of the fault; analyzing the maintenance historical data to obtain a fault maintenance period and maintenance cost data; constructing a maintenance scheme evaluation model based on a machine learning algorithm, and analyzing the optimization effect of the maintenance scheme; and carrying out comparative analysis on the maintenance scheme evaluation data and the maintenance demand prediction value, and adjusting decision tree parameters in the random forest regression model. According to the method, the problem that in the prior art, the requirements of airline companies for high efficiency, accuracy and cost effectiveness of aircraft maintenance are difficult to meet is solved, the maintenance process is optimized in a scientific and reasonable mode, and the reliability and operation and maintenance benefits of aircrafts are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of maintenance optimization, and more specifically, to an aircraft maintenance plan optimization analysis method and system. Background Art

[0002] With the continuous development of the aviation industry, the number of aircraft has been increasing, and the requirements for their safety and availability are also getting higher and higher. The aviation industry has always been regarded by the aviation industry as an industry with high risks, high investments, and high debt ratios. For Chinese airlines, the cost of aviation fuel is the main part of the airline's costs and also the most uncontrollable cost. Therefore, the aircraft maintenance cost, which is another major part of the airline's costs, has become another major entry point for airlines to reduce costs. In today's increasingly competitive aviation market, operations emphasize intensification and scientific management. How to optimize the airline's maintenance and production planning and control system to reduce costs is an important issue faced by all Chinese airlines at present and also the only way to remain invincible in the competition.

[0003] Deficiencies of the prior art:

[0004] Most of the existing aircraft maintenance plan optimization methods focus on the optimization of maintenance scheduling, and less involve the comprehensive optimization of multiple aspects such as aircraft fault prediction, maintenance resource scheduling, and maintenance task priority ranking. This makes the aircraft maintenance work fail to fully utilize the advantages of informatization and intelligence, and it is difficult to meet the airline's requirements for the efficiency, accuracy, and cost-effectiveness of aircraft maintenance. Therefore, it is of great research and practical application significance to propose an aircraft maintenance plan optimization analysis method and system to optimize the maintenance process in a scientific and reasonable manner and improve the reliability and operation and maintenance benefits of aircraft.

[0005] In view of the above problems, the present invention proposes a solution. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide an aircraft maintenance plan optimization analysis method and system, by proposing an aircraft maintenance plan optimization analysis method and system to solve the problems raised in the above background art.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] An aircraft maintenance plan optimization analysis method and system, comprising the following steps: collecting historical maintenance data of the aircraft, preprocessing and feature extracting the historical maintenance data, and establishing a maintenance demand prediction model for aircraft components based on the historical maintenance data using a random forest regression model; performing a failure mode and effects analysis on the critical components of the aircraft according to the historical maintenance data, predicting the possible failure types and occurrence probabilities, and calculating the failure risk priority number according to the failure types and occurrence probabilities to quantify the risk of the failure; analyzing the failure maintenance cycle and maintenance cost data based on the maintenance historical data, and obtaining the maintenance time of the current failure according to the failure maintenance cycle; constructing a maintenance plan evaluation model based on a machine learning algorithm by combining the failure risk priority number, the maintenance interval of the failure, and the maintenance cost data, obtaining maintenance plan evaluation data, and analyzing the optimization effect of the maintenance plan according to the maintenance plan evaluation data;

[0009] Comparing and analyzing the maintenance plan evaluation data with the maintenance demand prediction values, and adjusting the decision tree parameters in the random forest regression model according to the comparison and analysis results.

[0010] In a preferred embodiment, the historical maintenance data includes maintenance records, component usage conditions, failure modes, environmental factors, and component replacement conditions.

[0011] In a preferred embodiment, the establishing a maintenance demand prediction model for aircraft components based on the random forest regression model includes the following steps: inputting the quantitatively analyzable historical maintenance data extracted by feature extraction, and dividing the historical maintenance data into a training set and a test set; using the training set data to train the random forest regression model, and the model fits the data by constructing a number of decision trees to obtain a model capable of predicting maintenance demand; using the test set data to test the obtained random forest regression model for predicting maintenance demand, and adjusting the parameters of the decision tree.

[0012] In a preferred embodiment, the steps of performing a failure mode and effects analysis on the critical components of the aircraft to predict the possible failure types and occurrence probabilities are as follows: identifying the possible failure modes of each critical component, and analyzing the potential consequences of each failure mode; and obtaining the past failure historical data according to the historical maintenance data; obtaining the failure types according to the failure historical data, and the failure types include the severity of the failure and the difficulty of failure detection; and calculating the failure occurrence frequency based on the Bayesian network Sunav.

[0013] In a preferred embodiment, the maintenance cost data includes material cost, labor cost, and operating losses caused by aircraft grounding; the steps of obtaining the maintenance time of the current fault according to the fault maintenance cycle are as follows: calculate the average value of all historical maintenance cycles according to the time interval between every two faults; and add the maintenance time of the previous fault of the current fault to the average value of the historical maintenance cycles to obtain the maintenance time of the current fault;

[0014] In a preferred embodiment, the specific process of evaluating the optimization effect of the maintenance plan according to the maintenance plan evaluation data is as follows: compare and analyze the maintenance plan evaluation data with a preset maintenance plan evaluation threshold; if the maintenance plan evaluation data is greater than the preset maintenance plan evaluation threshold, the optimization effect of the maintenance plan is good, and there is no need to adjust the maintenance plan temporarily; if the maintenance plan evaluation data is less than the preset maintenance plan evaluation threshold, the optimization effect of the maintenance plan is poor, and it is necessary to shorten the maintenance grounding time and reduce the frequency of maintenance grounding by improving the fault maintenance efficiency, and extend the maintenance interval to optimize the maintenance plan.

[0015] In a preferred embodiment, the specific steps of comparing and analyzing the maintenance plan evaluation data with the maintenance demand prediction value and adjusting the decision tree parameters in the random forest regression model according to the comparison and analysis results are as follows: compare and analyze the maintenance plan evaluation data with the maintenance demand prediction value; calculate the absolute error between the maintenance plan evaluation data and the maintenance demand prediction value; and compare the absolute error with a preset error threshold. If the absolute error is greater than the preset error threshold, the model prediction is inaccurate; it is necessary to adjust the decision tree parameters in the random forest regression model to make the absolute error less than the preset error threshold.

[0016] A system for an aircraft maintenance plan optimization analysis method includes a prediction demand module, a fault evaluation module, a maintenance analysis module, a plan evaluation module, and a model optimization module, and there are connections between the modules:

[0017] The prediction demand module is used to collect the historical maintenance data of the aircraft, preprocess and extract features from the historical maintenance data, and establish a maintenance demand prediction model for aircraft components based on the random forest regression model according to the historical maintenance data;

[0018] The fault evaluation module is used to perform fault mode and effect analysis on the key components of the aircraft according to the historical maintenance data, predict the possible fault types and occurrence probabilities, and calculate the fault risk priority number according to the fault types and occurrence probabilities to quantify the risk of the fault;

[0019] The maintenance analysis module is used to analyze the fault maintenance cycle and maintenance cost data according to the analysis of the maintenance historical data, and obtain the maintenance time of the current fault according to the fault maintenance cycle;

[0020] A solution evaluation module, which is used to construct a maintenance plan evaluation model based on a machine learning algorithm by combining the failure risk priority number with the maintenance interval and maintenance cost data of the failure, obtain maintenance plan evaluation data, and analyze the optimization effect of the maintenance plan according to the maintenance plan evaluation data;

[0021] A model optimization module, which is used to compare and analyze the maintenance plan evaluation data with the predicted value of maintenance requirements, and adjust the decision tree parameters in the random forest regression model according to the comparison and analysis results.

[0022] The technical effects and advantages of a method and system for optimizing and analyzing an aircraft maintenance plan according to the present invention:

[0023] 1. By optimizing and analyzing the aircraft maintenance plan, the present invention can significantly improve the efficiency of the maintenance process; through an intelligent analysis method, it can accurately judge and arrange the priority of each maintenance step, reduce unnecessary steps and repetitive labor, thereby greatly saving time and human resources; reduce the aircraft grounding time, ensure the efficient development of maintenance activities, and thus improve the operation efficiency of the airline and reduce the unplanned downtime of the aircraft; through a systematic and data-driven optimization analysis method, each link of the maintenance plan can be determined more accurately, ensuring that the maintenance work meets the standard requirements, reducing human errors and omissions, and improving the maintenance quality; at the same time, the analysis process can timely identify potential failure hazards or wear trends during long-term use, and perform preventive maintenance in advance to reduce safety risks.

[0024] 2. By optimizing the maintenance plan, the present invention can allocate resources reasonably and avoid unnecessary maintenance operations and waste of parts; through accurate prediction of the maintenance cycle, workload and required resources, it can optimize the procurement and management of spare parts, thereby reducing material waste and capital expenditure during the maintenance process; reduce the maintenance cost, increase the service life of the equipment, enabling the airline to reduce the operation cost and improve the economic benefit; the system uses big data analysis and machine learning technology, and can formulate a more accurate and intelligent maintenance plan through historical data and real-time data analysis; the system can automatically optimize the scheduling of maintenance tasks, predict maintenance requirements, and even automatically adjust the maintenance plan according to environmental changes and the status of the aircraft; by introducing advanced data analysis, artificial intelligence and Internet of Things technologies, the present invention promotes the digital transformation of the aircraft maintenance industry; traditional maintenance plans often rely on experience and manual judgment, while this system provides more scientific and automated support for maintenance decisions through data analysis and model prediction. Brief Description of the Drawings

[0025] Figure 1 It is a schematic structural diagram of a method for optimizing and analyzing an aircraft maintenance plan according to the present invention.

[0026] Figure 2 This is a schematic diagram of the system structure of an optimization analysis method for an aircraft maintenance plan according to the present invention. Specific embodiments

[0027] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0028] Embodiment 1 Figure 1 An optimization analysis method for an aircraft maintenance plan according to the present invention is given.

[0029] S10. Collect the historical maintenance data of the aircraft, preprocess and extract features from the historical maintenance data, and establish a maintenance requirement prediction model for aircraft components based on the random forest regression model according to the historical maintenance data;

[0030] The historical maintenance data includes maintenance records, component usage, failure modes, environmental factors,

[0031] Component replacement situations.

[0032] Maintenance records: including the time, maintenance items, maintenance types, maintenance components, fault descriptions, etc. of each maintenance;

[0033] Component usage: including flight time, flight cycles, flight times, etc. of the aircraft;

[0034] Failure modes: record the failure history of each component, including the frequency, cause, and scope of influence of failures, etc.;

[0035] Environmental factors: external factors that may affect component life and failure rate, such as environmental variables such as flight altitude, flight temperature, humidity, etc.;

[0036] Component replacement situations: record the historical information of component replacements, including replacement time, component model, service life, etc.

[0037] Before inputting the data into the machine learning model, certain preprocessing and feature engineering extractions are required, including:

[0038] Delete duplicate or invalid data, fill in missing values, and handle outliers;

[0039] Extract features from the quantitatively analyzable variables in the historical maintenance data, and perform aggregation processing on the maintenance history of components;

[0040] To avoid the dimensionality differences between features, it is usually necessary to normalize the data.

[0041] The establishment of a maintenance requirement prediction model for aircraft components based on a random forest regression model includes the following steps:

[0042] Extract the quantifiable historical maintenance data of the input features, and divide the historical maintenance data into a training set and a test set;

[0043] Use the training set data to train the random forest regression model. The model fits the data by constructing several decision trees to obtain a model that can predict maintenance requirements;

[0044] The decision tree training formula is as follows: In the formula, is the predicted maintenance requirement value of the i-th decision tree, x i is the input feature of this decision tree, and θ is the parameter of the tree;

[0045] Use the test set data to test the obtained random forest regression model for predicting maintenance requirements, and adjust the parameters of the decision tree;

[0046] The final predicted maintenance requirement value of the random forest is the average of the predicted values of all trees: In the formula, y is the predicted maintenance requirement value, is the predicted maintenance requirement value of the i-th decision tree, and N is the number of trees.

[0047] The predicted maintenance requirement value includes the time when the component needs maintenance in the future, the future failure risk priority number, and the future maintenance cost.

[0048] S20. Conduct a failure mode and effect analysis on the key components of the aircraft based on historical maintenance data, predict the possible failure types and occurrence probabilities, and calculate the failure risk priority number according to the failure types and occurrence probabilities to quantify the risk of failure;

[0049] The steps of conducting a failure mode and effect analysis on the key components of the aircraft to predict the possible failure types and occurrence probabilities are as follows:

[0050] Identify the possible failure modes of each key component. For example, the possible failure modes of an engine include:

[0051] Fuel pump failure; Turbine blade wear; Engine oil leakage, etc.;

[0052] Analyze the potential consequences of each failure mode. For example: Fuel pump failure may cause the engine to stop;

[0053] Turbine blade wear may cause a decrease in thrust and affect flight safety;

[0054] Obtain past fault history data based on historical maintenance data, obtain the fault type according to the fault history data, and calculate the fault occurrence frequency based on the Bayesian network;

[0055] The fault type includes the severity of the fault and the difficulty of fault detection;

[0056] The steps to calculate the fault risk priority number to quantify the fault risk according to the fault type and occurrence probability are as follows:

[0057] Generally, the fault risk priority number (RPN, Risk Priority Number) can be used to quantify the fault risk, and the calculation formula is: RPN = Severity × Occurrence × Detection;

[0058] Where: Severity (severity): The severity of the fault, usually rated on a scale of 1 to 10, with 10 representing the most severe fault;

[0059] Occurrence (occurrence probability): The frequency of fault occurrence, rated on a scale of 1 to 10, with 10 representing the highest occurrence probability;

[0060] Detection (detectability): The difficulty of fault detection, rated on a scale of 1 to 10, with 10 representing the most difficult to detect.

[0061] S30. Analyze the maintenance history data to obtain the fault repair cycle and maintenance cost data, and obtain the repair time of the current fault according to the fault repair cycle;

[0062] The fault repair cycle includes the time interval between two faults, and the specific calculation formula is as follows: Z j = T j - T j-1 ; In the formula, Z j is the time interval between the jth fault repair and the previous fault repair, T j is the jth fault repair time, and T j-1 is the (j - 1)th fault repair time;

[0063] The steps to obtain the repair time of the current fault according to the fault repair cycle are as follows:

[0064] Calculate the average value of all historical repair cycles according to the time interval between every two faults, and add the previous fault repair time of the current fault to the average value of the historical repair cycles to obtain the repair time of the current fault. The specific calculation formula of the repair time of the current fault is as follows: In the formula, T kis the repair time of the current failure, T K-1 is the previous repair time of the current failure; is the average value of the historical repair cycle, k - 1 is the number of failure repairs, is the sum of all failure cycles, Z j is the time interval between the jth failure repair and the previous failure repair.

[0065] The said repair cost data includes material cost, labor cost, and operating loss caused by the aircraft being out of service; the specific calculation formula for calculating the repair cost data is as follows: C = C 1 + C 2 + C 3 ; where C is the repair cost data, C 1 is the material cost, C 2 is the labor cost, C 3 is the operating loss caused by the aircraft being out of service;

[0066] S40. Combine the failure risk priority number with the repair interval and repair cost data of the failure to construct a repair plan evaluation model based on a machine learning algorithm, obtain the repair plan evaluation data, and analyze the optimization effect of the repair plan according to the repair plan evaluation data;

[0067] Combine the failure risk priority number with the repair interval and repair cost data of the failure to construct a repair plan evaluation model based on a machine learning algorithm, obtain the repair plan evaluation data, and the specific calculation formula for the said repair plan evaluation data is as follows: P = μ 1 T k - μ 2 C - μ 3 RPN; where P is the repair plan evaluation data, T k is the repair time of the current failure, C is the repair cost data, RPN is the failure risk priority number, μ 1 is the weight factor of the repair time of the current failure, μ 2 is the weight factor of the repair cost data, μ 3 is the failure risk priority number, where μ 1 , μ 2 , μ 3 are all greater than 0.

[0068] The process of analyzing the optimization effect of the repair plan according to the repair plan evaluation data is as follows:

[0069] Compare and analyze the repair plan evaluation data with the preset repair plan evaluation threshold;

[0070] If the repair plan evaluation data is greater than the preset repair plan evaluation threshold, the optimization effect of the repair plan is good, and there is no need to adjust the repair plan temporarily;

[0071] If the maintenance plan evaluation data is less than the preset maintenance plan evaluation threshold, the optimization effect of the maintenance plan is not good. It is necessary to shorten the maintenance downtime and reduce the frequency of maintenance downtime by improving the fault maintenance efficiency, and extend the maintenance interval to optimize the maintenance plan.

[0072] S50, compare and analyze the maintenance plan evaluation data with the predicted value of maintenance demand, and adjust the decision tree parameters in the random forest regression model according to the comparison and analysis results.

[0073] Compare and analyze the maintenance plan evaluation data with the predicted value of maintenance demand, calculate the absolute error between the maintenance plan evaluation data and the predicted value of maintenance demand, and compare the absolute error with the preset error threshold. If the absolute error is greater than the preset error threshold, the model prediction is inaccurate, and it is necessary to adjust the decision tree parameters in the random forest regression model to make the absolute error less than the preset error threshold.

[0074] Embodiment 2 Figure 2 The system structure diagram of a method for optimizing and analyzing an aircraft maintenance plan according to the present invention is given.

[0075] A system for a method for optimizing and analyzing an aircraft maintenance plan includes a predicted demand module, a fault evaluation module, a maintenance analysis module, a plan evaluation module, and a model optimization module, and there are connections between the modules:

[0076] The predicted demand module is used to collect the historical maintenance data of the aircraft, preprocess and extract features from the historical maintenance data, and establish a maintenance demand prediction model for aircraft components based on the random forest regression model according to the historical maintenance data;

[0077] The fault evaluation module is used to perform fault mode and effect analysis on the key components of the aircraft according to the historical maintenance data, predict the possible fault types and occurrence probabilities, and calculate the fault risk priority number according to the fault types and occurrence probabilities to quantify the risk of the fault;

[0078] The maintenance analysis module is used to analyze the maintenance historical data to obtain the fault maintenance cycle and maintenance cost data, and obtain the maintenance time of the current fault according to the fault maintenance cycle;

[0079] The plan evaluation module is used to construct a maintenance plan evaluation model based on the machine learning algorithm by combining the fault risk priority number, the maintenance interval and the maintenance cost data of the fault, obtain the maintenance plan evaluation data, and analyze the optimization effect of the maintenance plan according to the maintenance plan evaluation data;

[0080] A model optimization module is used to compare and analyze the maintenance plan evaluation data with the predicted values of maintenance requirements, and adjust the decision tree parameters in the random forest regression model according to the results of the comparison and analysis.

[0081] The above formulas are all dimensionless and take their numerical calculations. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0082] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product.

[0083] Those of ordinary skill in the art can realize that the modules and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0084] In addition, the functional modules in each embodiment of this application can be integrated into a processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0085] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claimed rights.

[0086] Finally: The above is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An aircraft maintenance program optimization analysis method and system, characterized in that: The following steps are involved: Collect historical maintenance data of aircraft, preprocess and extract features of the historical maintenance data, and establish a maintenance demand prediction model for aircraft parts based on the random forest regression model according to the historical maintenance data; Conduct failure mode and effect analysis on key aircraft components based on historical maintenance data, predict possible failure types and their probability of occurrence, and calculate failure risk priority numbers based on failure types and probability of occurrence to quantify failure risks; The fault repair cycle and repair cost data are obtained by analyzing the repair history data, and the repair time of the current fault is obtained according to the fault repair cycle; The fault risk priority number is combined with the fault repair time and repair cost data to build a maintenance plan evaluation model based on a machine learning algorithm to obtain maintenance plan evaluation data, and the optimization effect of the maintenance plan is analyzed based on the maintenance plan evaluation data; The maintenance plan evaluation data and the maintenance demand prediction value are compared and analyzed, and the decision tree parameters in the random forest regression model are adjusted according to the comparative analysis results.

2. The aircraft maintenance program optimization analysis method according to claim 1, characterized in that: The historical maintenance data includes maintenance records, component usage, failure modes, environmental factors, and component replacement status.

3. The aircraft maintenance program optimization analysis method according to claim 2, characterized in that: The method of establishing a maintenance demand prediction model for aircraft parts based on a random forest regression model comprises the following steps: Input the historical maintenance data that can be quantitatively analyzed after feature extraction, and divide the historical maintenance data into a training set and a test set; The random forest regression model is trained using the training set data. The model fits the data by constructing several decision trees to obtain a model that can predict maintenance needs. The formula for training the random forest regression model using the training set data is as follows: In the formula, is the maintenance demand prediction value of the i-th decision tree, x i is the input feature of the decision tree, and θ is the parameter of the tree; Use the test set data to test the obtained random forest regression model for predicting maintenance needs and adjust the parameters of the decision tree; The formula for obtaining the final maintenance demand prediction value according to the maintenance demand prediction model is as follows: Where y is the maintenance demand forecast value, is the maintenance demand prediction value of the i-th decision tree, and N is the number of trees.

4. The aircraft maintenance program optimization analysis method according to claim 3, characterized in that: The steps of performing failure mode and effect analysis on key aircraft components and predicting possible failure types and probability of occurrence are as follows: Identify the possible failure modes of each critical component and analyze the potential consequences of each failure mode; And obtain past fault history data based on historical maintenance data; Acquire the fault type according to the fault history data, wherein the fault type includes the severity of the fault and the difficulty of fault detection; And based on the Bayesian network Sunaf calculates the frequency of fault occurrence.

5. The aircraft maintenance program optimization analysis method according to claim 4, characterized in that: The formula for calculating the fault risk priority number based on the fault type and occurrence probability is as follows: RPN=Severity×Occurrence×Detection; RPN is the fault risk priority number, Severity indicates the severity of the fault, Occurrence indicates the frequency of the fault, and Detection indicates the difficulty of fault detection.

6. The aircraft maintenance program optimization analysis method according to claim 5, characterized in that: The maintenance cost data includes material cost, labor cost, and operating loss due to aircraft grounding; The fault repair cycle includes the time interval between two faults, and the specific calculation formula is as follows: j =T j -T j-1 ; In the formula, Z j is the time interval between the jth fault repair and the last fault repair, T j The jth fault repair time, T j-1 is the j-1th fault repair time; The steps of obtaining the repair time of the current fault according to the fault repair cycle are as follows: Calculate the average of all historical maintenance cycles based on the time interval between each failure; The repair time of the current fault is obtained by adding the last repair time of the current fault to the average value of the historical repair cycles; The specific calculation formula for the repair time of the current fault is as follows: Where, T k is the repair time of the current fault, T K-1 is the last repair time of the current fault; is the average of the historical maintenance cycles, k-1 is the number of failure repairs, is the sum of all fault cycles, Z j is the time interval between the jth fault repair and the previous fault repair.

7. The aircraft maintenance program optimization analysis method according to claim 6, characterized in that: The specific calculation formula of the maintenance plan evaluation data is as follows: P = μ1T k -μ2C-μ3RPN; where P is the maintenance plan evaluation data, T k is the repair time of the current fault, C is the repair cost data, RPN is the fault risk priority number, μ1 is the weight factor of the repair time of the current fault, μ2 is the weight factor of the repair cost data, and μ3 is the fault risk priority number, where μ1, μ2, and μ3 are all greater than 0.

8. The aircraft maintenance program optimization analysis method according to claim 7, characterized in that: The specific process of analyzing the optimization effect of the maintenance plan based on the maintenance plan evaluation data is as follows: Compare and analyze the maintenance plan evaluation data with the preset maintenance plan evaluation threshold; If the maintenance plan evaluation data is greater than the preset maintenance plan evaluation threshold, the optimization effect of the maintenance plan is good and there is no need to adjust the maintenance plan temporarily; If the maintenance plan evaluation data is less than the preset maintenance plan evaluation threshold, the optimization effect of the maintenance plan is not good. It is necessary to improve the fault repair efficiency to shorten the maintenance downtime and reduce the frequency of maintenance downtime, and extend the maintenance interval to achieve the optimization of the maintenance plan.

9. The aircraft maintenance program optimization analysis method according to claim 8, characterized in that: The specific steps of comparing and analyzing the maintenance plan evaluation data with the maintenance demand forecast value and adjusting the decision tree parameters in the random forest regression model according to the comparative analysis results are as follows: Compare and analyze the maintenance plan evaluation data with the maintenance demand forecast value; and calculate the absolute error between the maintenance plan evaluation data and the maintenance demand forecast value; The absolute error is compared with the preset error threshold. If the absolute error is greater than the preset error threshold, the model prediction is inaccurate. The decision tree parameters in the random forest regression model need to be adjusted so that the absolute error is less than the preset error threshold.

10. A system for optimizing and analyzing an aircraft maintenance plan, characterized in that: It includes demand prediction module, fault assessment module, maintenance analysis module, solution evaluation module, and model optimization module. There are connections between the modules: The demand prediction module is used to collect historical maintenance data of aircraft, preprocess and extract features of the historical maintenance data, and establish a maintenance demand prediction model for aircraft parts based on the random forest regression model according to the historical maintenance data; The fault assessment module is used to analyze the failure modes and effects of key aircraft components based on historical maintenance data, predict the possible types of failures and their probability of occurrence, and calculate the failure risk priority number based on the failure type and probability of occurrence to quantify the risk of failure; The maintenance analysis module is used to obtain the fault maintenance cycle and maintenance cost data based on the analysis of the maintenance history data, and obtain the maintenance time of the current fault based on the fault maintenance cycle; A solution evaluation module is used to construct a maintenance solution evaluation model based on a machine learning algorithm by combining the fault risk priority number with the maintenance time and maintenance cost data of the fault, obtain maintenance solution evaluation data, and analyze the optimization effect of the maintenance solution based on the maintenance solution evaluation data; The model optimization module is used to compare and analyze the maintenance plan evaluation data with the maintenance demand prediction value, and adjust the decision tree parameters in the random forest regression model according to the comparative analysis results.

Images