Aircraft maintenance scheme execution effect evaluation dynamic monitoring method

Abstract

The invention discloses an aircraft maintenance scheme execution effect evaluation dynamic monitoring method. The method comprises the following steps: step 1, calculating monthly fault indexes before and after optimization of a maintenance scheme; 2, whether the two overall mean values are equal or not is checked, and distribution of differences of sample mean values of the two data sets before and after optimization of the maintenance scheme is given; step 3, obtaining a combined standard deviation of two data set samples before and after optimization of the maintenance scheme; and step 4, establishing an alarm value of the failure rate of the fleet system, i.e., determining the alarm value by calculating the standard deviation of the failure rate (x) of the fleet system per thousand flight hours in adjacent three months by using the flight hours in the past 14 months, the unit report failure and the ground maintenance discovery times. The invention belongs to the technical field of aircraft maintenance reliability evaluation and monitoring, and performs evaluation and dynamic monitoring on the maintenance scheme optimization effect of long-period maintenance data through reasonable classification and integration of maintenance data, setting of original hypothesis and inconsistent hypothesis and the like.

Description

Technical Field

[0001] The present invention belongs to the technical field of aircraft maintenance reliability evaluation and monitoring, and specifically refers to a dynamic monitoring method for evaluating the execution effect of an aircraft maintenance program. Background Art

[0002] According to the provisions of the Civil Aviation Administration of China (CCAR-121), the certificate holder shall prepare a maintenance plan for each aircraft it operates and continuously monitor the effectiveness of the maintenance plan based on the reliability plan. China Civil Aviation Advisory Circular AC-121 / 135-53 (Maintenance Plan for Civil Aircraft) 6.4(5) requires that the modification of maintenance intervals shall be gradually implemented through controllable means (such as sampling tests, statistical analysis of maintenance data or other supporting data, etc.), but the approval of the corresponding civil aviation regional administration must be obtained in advance. The modification of maintenance intervals should be monitored by controllable methods to ensure the reliability of related systems after the interval modification.

[0003] According to the requirements of the above-mentioned Chinese Civil Aviation Regulations and Advisory Circulars, the maintenance plan needs to be continuously monitored through the reliability plan, and the maintenance interval modification of the maintenance plan items should be monitored through a controllable method to ensure the reliability of the relevant systems after the interval modification. At present, aviation operators generally follow the "7.4 Statistical Performance Standard (Warning Value)" of the Chinese Civil Aviation Advisory Circular AC-121-54R1 "Reliability Plan" to establish corresponding warning values ​​based on the monthly system or component failure fluctuations of the fleet using the "standard deviation" or "Poisson distribution" statistical method, and monitor on a monthly basis.

[0004] For example, China Civil Aviation Advisory Circular AC-121-54R1 "Reliability Program" requires: "For control objects (such as systems and components) that use alert value performance standards, statistical performance standard analysis methods should be used to compare the collected data with the currently adopted performance standards." The specific implementation methods are as follows:

[0005] Statistical performance standards (warning values)

[0006] (1) Normally, the performance of an aircraft is measured numerically based on crew reports, flight delays or cancellations, unplanned component replacement, confirmed component failure rate, service time, or other events, and is used as the basis for establishing standards. The following factors should be considered in determining performance standards (alert values):

[0007] (a) Past and current operating experience of independent operators and industry. If such experience is used, the operator's reliability control program must include provisions and measures to review performance standards after one year of operating experience has been accumulated;

[0008] (b) performance analysis of similar equipment currently in use;

[0009] (c) Reliability engineering analysis by aircraft or equipment manufacturers;

[0010] (d) In-service experience based on reliability standards acceptable to the aviation community.

[0011] (2) The establishment and modification of control limits or warning values ​​are usually based on acceptable statistical methods, such as standard deviation or Poisson distribution. Common performance standard calculations (but not limited to) are as follows:

[0012] (a) Unit reported faults: (at least 12 months of cumulative data)

[0013] Calculation method 1: 1.3*average reported failure rate per thousand hours

[0014] Calculation method 2: average (rounded) of reported failure rate per thousand hours + 3 times standard deviation (rounded)

[0015] Calculation method 3: average reported failure rate per thousand hours + standard deviation of the average value of the adjacent month + 3 times the standard deviation (rounded)

[0016] (b) Unplanned replacement rate of components: (at least 21 months of cumulative data)

[0017] Calculation method 4: average value of unplanned component replacement rate per quarter + 2 times standard deviation

[0018] Calculation method 5: Use the Poisson distribution table to find the maximum acceptable number of unplanned replacements per quarter with a probability of 95%

[0019] (c) Component confirmation failure:

[0020] Calculation method 6: (at least 21 months of cumulative data)

[0021] Adjusted mean + adjusted standard deviation of quarterly component confirmation thousand-hour failure rate

[0022] (3) The program should include procedures for adjusting and periodically reviewing the adopted performance standards. Aircraft performance standards should be adjusted based on the operator's operating experience and should reflect the effects of seasonal and environmental factors.

[0023] (4) The program should also include a performance monitoring program for the new aircraft until sufficient operating data and experience have been accumulated on which performance standards can be calculated, which should normally take one year.

[0024] The existing "statistical performance standard (warning value)" method is a short-cycle abnormal fluctuation alarm logic. The advantage of this method is that it can detect the monthly system or component failure deterioration trend so that timely measures can be taken to correct it. However, the disadvantage of this method is that it is difficult to reflect the impact of changes in maintenance project intervals on the fleet, because there are many reasons for the fluctuation of the fleet's monthly data, such as changes in the external environment or operating conditions, repeated multiple failures in the fleet, and policy factors. In addition, the fault alarm cycle of this method is too short, and the execution cycle of the maintenance plan project is very long. The use of short-cycle abnormal fluctuation alarm logic to judge the long-term impact of fleet maintenance plan changes is not effective.

[0025] The present invention is mainly applied to the field of aircraft maintenance reliability evaluation and monitoring. The present invention provides a method for dynamic evaluation and monitoring of aircraft maintenance program execution effect based on hypothesis testing. The method is based on "7.2 Requirements for Data Analysis System" of China Civil Aviation Advisory Circular AC-121-54R1 "Reliability Program", and invents a traditional reliability monitoring method that is different from "using statistical performance standard analysis for control objects (such as systems and components) using warning value performance standards". Through reasonable classification and integration of maintenance data, setting of null hypothesis and paradox hypothesis, etc., the optimization effect of maintenance program on long-term maintenance data is evaluated and dynamically monitored. Summary of the invention

[0026] The technical solution adopted by the present invention is as follows: A method for dynamically monitoring the execution effect evaluation of an aircraft maintenance plan proposed by the present invention comprises the following steps:

[0027] Step 1: Record the monthly failure indicators (fleet failure rate, crew report rate, ground discovery rate, and delay rate) before and after the maintenance plan optimization as

[0028] Step 2: Check whether the two population means are equal. First, give the distribution of the difference between the sample means of the two data sets before and after the maintenance plan optimization.

[0029] Step 3: The combined standard deviation of the samples of the two data sets before and after the maintenance plan optimization;

[0030] Step 4: Establish a warning value for the fleet system failure rate, which is determined by calculating the standard deviation of the fleet system failure rate (x) per thousand flight hours in the three consecutive months using the flight hours, crew reported failures and ground maintenance findings in the past 14 months;

[0031] Step 5: The warning value is to use the Poisson cumulative probability distribution table to find the maximum acceptable number of unplanned replacements in a quarter with a probability of 95%, and compare it with the number of unplanned replacements in the current quarter to determine whether it exceeds the limit and alarm;

[0032] Step six: Use Poisson cumulative distribution to monitor the reliability of accessories.

[0033] Furthermore, in step 1, the monthly failure indicators (fleet failure rate, crew report rate, ground discovery rate, and delay rate) before and after the maintenance plan optimization are recorded as The sample mean and sample standard deviation of the monthly failure index before and after optimization are:

[0034]

[0035] Furthermore, in step 2, because the monthly failure indicators before and after the maintenance plan optimization come from the same data set, the variances of the samples of the two data sets are equal and unknown, and the sample variance satisfies The overall variance of the data is To test whether the two population means are equal, first give the distribution of the difference between the sample means of the two data sets before and after the maintenance plan optimization. According to the assumption, we get:

[0036]

[0037]

[0038] By the Chi-square theorem:

[0039]

[0040] Since the sample distributions of the two data sets before and after the maintenance plan optimization are independent, Independence, by additivity of chi-square:

[0041]

[0042] Defined by the construction of the distribution, After simplification, we can get:

[0043]

[0044] in:

[0045]

[0046] Furthermore, in step 3, similarly, under the condition that the sample means of the two data sets before and after the maintenance plan optimization are equal to the overall mean, μ1-μ2=0;

[0047] Because the test is two-sided, the degrees of freedom are n-1, and the two-sided α gives the two-tailed critical value Establishing the hypothesis:

[0048] Null hypothesis: There is no significant change in the samples of the two data sets before and after the maintenance plan optimization

[0049] H0: Paradox hypothesis: There is a significant change in the samples of the two data sets before and after the maintenance plan optimization

[0050] H a : if:

[0051]

[0052] Then the null hypothesis is rejected, and it is believed that the sample means of the two data sets before and after the optimization of the two maintenance schemes are not equal, otherwise the null hypothesis is not rejected;

[0053] If the null hypothesis is rejected,

[0054]

[0055] It represents The mean is significantly less than For example, after the maintenance plan is optimized, the data set is The data set before maintenance plan optimization is The mean failure rate of the data set after the maintenance scheme is optimized is significantly lower than that of the data set before the maintenance scheme is optimized, and the failure rate decreases significantly after the maintenance scheme is optimized;

[0056] If the null hypothesis is rejected,

[0057]

[0058] It represents The mean is significantly greater than If the maintenance plan is optimized, the data set is The data set before maintenance plan optimization is The mean failure rate of the data set after the maintenance plan is optimized is significantly higher than that of the data set before the maintenance plan is optimized, and the failure rate has increased significantly after the maintenance plan is optimized.

[0059] Furthermore, in step 4, the warning value of the fleet system failure rate is established according to the performance standard "Calculation Method 2" provided in AC-121-54, that is, the warning value is determined by calculating the standard deviation of the fleet system failure rate (x) per thousand flight hours in the three consecutive months using the flight hours, crew reported failures and ground maintenance findings in the past 14 months;

[0060]

[0061] In the formula:

[0062] X = fleet system failure rate per thousand flight hours in the three consecutive months;

[0063] N = the number of samples in three consecutive months, normally 12;

[0064] = the average value of fleet system failure rate (x) per thousand flight hours in the three consecutive months;

[0065] Σ = sum of 12 ratios;

[0066] Warning value calculation:

[0067]

[0068]

[0069] Furthermore, in step 5, according to the performance standard "Calculation Method 5" Poisson distribution monitoring annex provided in AC-121-54, the warning value is to find out the maximum acceptable number of unplanned replacements in a quarter with a probability of 95% using the Poisson cumulative probability distribution table, and compare the number of unplanned replacements in the current quarter to determine whether it exceeds the standard and issue an alarm;

[0070] Suppose all possible values ​​of the random variable X are 0, 1, 2, ..., and the probability of taking each value is:

[0071] where k = 0, 1, 2, ...,

[0072] Where λ is a constant, then X is said to obey the Poisson distribution with parameter λ;

[0073] where k = 0, 1, 2, ...,

[0074] The above formula is the probability that the random variable X is less than or equal to the value k, which is called Poisson cumulative distribution.

[0075] Furthermore, in step six, according to calculation method five of advisory circular AC-121-54, the reliability of the attachment is monitored by using Poisson cumulative distribution;

[0076] Warning value = the maximum acceptable number of unplanned replacements in a quarter with a probability of 95% found using the Poisson cumulative probability distribution table;

[0077] Components: Autoflight pitch amplifier;

[0078] The number of parts per aircraft is n = 1;

[0079] The number of flight plan replacements in the first 21 months is N = 62;

[0080] The fleet usage hours in the first 21 months were H = 36,840;

[0081] Component usage hours in the first 21 months T = (n × H) = 36840;

[0082] Fleet usage hours in the past three months h = 5895;

[0083] Component usage hours in the past three months t = (n × h) = 5895;

[0084] The number of unplanned replacements in the past three months is x = 12;

[0085] Average unplanned replacement rate λ = N / T = 0.00168;

[0086] The expected number of unplanned replacements in the past three months = λt = 0.00168 × 5895 = 9.9 rounded to the nearest integer = 10;

[0087] Referring to the Poisson cumulative probability diagram: the intersection line of λt=10 and 95% probability, it can be concluded that the maximum acceptable number of unplanned replacements (A value) is 15.

[0088] Comparing the current number of unplanned replacements x=12, it can be concluded that the warning value has not been exceeded.

[0089] The beneficial effects achieved by the present invention using the above structure are as follows:

[0090] 1. The maintenance data monitoring system established by the present invention is a long-term periodic data monitoring system, which solves the pain point that the maintenance project execution time span is long and difficult to monitor.

[0091] 2. The present invention can well integrate the traditional reliability monitoring system based on "statistical performance standards (warning values)", and the two monitoring systems can complement each other's advantages.

[0092] 3. The present invention can reflect the positive or negative effects obtained before and after the execution of the maintenance project.

[0093] 4. The present invention can quantify the normal fluctuation range and margin of the fleet system failure rate, and effectively reduce the impact of other factors on the fleet failure rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0094] Figure 1 Schematic diagram of "S1-Aircraft Maintenance Program Execution Effect Evaluation Dynamic Monitoring System" and "S2-Statistical Performance Standard (Warning Value) Monitoring System" of an embodiment of the present invention;

[0095] Figure 2 This is a schematic diagram of the display interface of the S1.1 monitoring system;

[0096] Figure 3 A schematic diagram of the elements of a monitoring system according to an embodiment of the present invention;

[0097] Figure 4 A schematic diagram of a maintenance solution before and after optimization using hypothesis testing according to an embodiment of the present invention;

[0098] Figure 5 Schematic diagram of a presentation method of an embodiment of the present invention.

[0099] Among them, Figure 2 In Chinese: 1. ATA chapter number; 2. ATA chapter description; 3. Maintenance plan optimization "fleet failure rate, crew reporting rate, ground detection rate, delay rate" coordinate axis; 4. Maintenance plan optimization effect evaluation performance interval; 5. Orange columns represent the average failure indicators (fleet failure rate, crew reporting rate, ground detection rate, delay rate) within a period of time after the maintenance plan is modified; 6. Blue columns represent the average failure indicators (fleet failure rate, crew reporting rate, ground detection rate, delay rate) within a period of time before the maintenance plan is modified; 7. Areas with good maintenance plan optimization effects; 8. Areas with stable maintenance plan optimization effects Domain; 9. Area where the optimization effect of the maintenance plan is not good; 10. The optimization effect of the maintenance plan has improved significantly. The more the block is to the left, the better the optimization effect is. The color of the block is dark green; 11. The changes in various system indicators before and after the optimization of the maintenance plan are within a reasonable range. The color of the block is light blue; 12. The optimization effect of the maintenance plan has deteriorated significantly. The more the block is to the right, the worse the optimization effect is. The color of the block is dark red; 13. The progressive color bar from red to green represents the Performace scale axis. The shorter the bar, the redder it is, the worse the data performance is. The longer the bar, the greener it is, the better the data performance is.

[0100] exist Figure 3 Chinese: 1. Double-click to pop up the fault rate details; 2. Average fault indicators before maintenance plan optimization (fleet fault rate, crew reporting rate, ground detection rate, delay rate); 3. Average fault indicators after maintenance plan optimization (fleet fault rate, crew reporting rate, ground detection rate, delay rate); 4. Monthly fault indicators before maintenance plan optimization (fleet fault rate, crew reporting rate, ground detection rate, delay rate); 5. Monthly fault indicators after maintenance plan optimization (fleet fault rate, crew reporting rate, ground detection rate, delay rate). DETAILED DESCRIPTION

[0101] The present technical solution is mainly composed of the following figure "S1-dynamic monitoring system for evaluating the implementation effect of aircraft maintenance program" and "S2-statistical performance standard (warning value) monitoring system", wherein S2 is mainly composed of the methods mentioned in AC-121-54R1 "Reliability Program" "S2.1 Alarm monitoring system based on abnormal unplanned replacement of components" and "S2.2 Monitoring system based on reliability warning value performance standard". The present invention establishes a mutual correlation relationship between S1 and S2 through the Air Transport Association of America (Air Transport Association of America) ATA100 specification and key component part numbers or group numbers.

[0102] The monthly failure indicators (fleet failure rate, crew report rate, ground discovery rate, and delay rate) before and after the maintenance plan optimization are recorded as The sample mean and sample standard deviation of the monthly failure index before and after optimization are:

[0103]

[0104] Because the monthly failure indicators before and after the maintenance plan optimization come from the same data set, the variances of the samples of the two data sets are equal and unknown, and the sample variance satisfies The overall variance of the data is To test whether the two population means are equal, first give the distribution of the difference between the sample means of the two data sets before and after the maintenance plan optimization. According to the assumption, we get:

[0105]

[0106] By the Chi-square theorem:

[0107]

[0108] Since the sample distributions of the two data sets before and after the maintenance plan optimization are independent, Independence, by additivity of chi-square:

[0109]

[0110] Defined by the construction of the distribution, After simplification, we can get:

[0111]

[0112] in:

[0113]

[0114] is the pooled standard deviation of the two data set samples before and after the maintenance plan is optimized.

[0115] Similarly, under the condition that the sample means of the two data sets before and after the maintenance plan optimization are equal to the overall mean, μ1-μ2=0.

[0116] Because the test is two-sided, the degrees of freedom are n-1, and the two-sided α gives the two-tailed critical value Establishing the hypothesis:

[0117] Null hypothesis: There is no significant change in the samples of the two data sets before and after the maintenance plan optimization

[0118] H0:

[0119] Paradox hypothesis: There is a significant change in the samples of the two data sets before and after the maintenance plan optimization

[0120] H a :

[0121] if:

[0122]

[0123] The null hypothesis is rejected, and it is believed that the sample means of the two data sets before and after the optimization of the two maintenance schemes are not equal. Otherwise, the null hypothesis is not rejected.

[0124] If the null hypothesis is rejected,

[0125]

[0126] It represents The mean is significantly less than For example, after the maintenance plan is optimized, the data set is The data set before maintenance plan optimization is The mean failure rate of the data set after the maintenance scheme is optimized is significantly lower than that of the data set before the maintenance scheme is optimized, and the failure rate decreases significantly after the maintenance scheme is optimized.

[0127] If the null hypothesis is rejected,

[0128]

[0129] It represents The mean is significantly greater than If the maintenance plan is optimized, the data set is The data set before maintenance plan optimization is The mean failure rate of the data set after the maintenance plan is optimized is significantly higher than that of the data set before the maintenance plan is optimized, and the failure rate has increased significantly after the maintenance plan is optimized.

[0130] S2-Statistical performance standard (warning value) monitoring system

[0131] S2.1 Alarm monitoring system based on abnormal unplanned replacement of components

[0132] Determination of Failure Rate Warning Value of Fleet System

[0133] The warning value of the fleet system failure rate is established according to the performance standard "Calculation Method 2" provided in AC-121-54, that is, using the flight hours, crew-reported failures and ground maintenance findings in the past 14 months, and determined by calculating the standard deviation of the fleet system failure rate (x) per thousand flight hours in the adjacent three months.

[0134]

[0135] In the formula:

[0136] X = fleet system failure rate per thousand flight hours in the three consecutive months

[0137] N = the number of samples in three consecutive months, normally 12

[0138] = Average value of fleet system failure rate (x) per thousand flight hours in the three consecutive months

[0139] Σ = sum of 12 ratios

[0140] Warning value calculation:

[0141]

[0142]

[0143] S2.2 Monitoring system based on reliability warning value performance standard

[0144] According to the performance standard "Calculation Method 5" Poisson distribution monitoring annex provided in AC-121-54, the warning value is the maximum acceptable number of unplanned replacements in a quarter with a probability of 95% found using the Poisson cumulative probability distribution table, and compared with the number of unplanned replacements in the current quarter to determine whether an alarm is issued if the standard is exceeded.

[0145] Suppose all possible values ​​of the random variable X are 0, 1, 2, ..., and the probability of taking each value is:

[0146] where k = 0, 1, 2, ...,

[0147] Where λ is a constant, then X is said to obey the Poisson distribution with parameter λ.

[0148] where k = 0, 1, 2, ...,

[0149] The above formula is the probability that the random variable X is less than or equal to the value k, which is called Poisson cumulative distribution. S2.2.1 Monitoring method

[0150] According to calculation method 5 of advisory circular AC-121-54, Poisson cumulative distribution is used to monitor the reliability of accessories. The specific method is shown in the following example.

[0151] Warning value = the maximum acceptable number of unplanned replacements per quarter with a probability of 95% using the Poisson cumulative probability distribution table.

[0152] Component: Autoflight pitch amplifier

[0153] Number of parts per aircraft n = 1

[0154] The number of flight plan replacements in the first 21 months is N = 62

[0155] Fleet usage hours in the first 21 months H = 36840

[0156] Component usage hours in the first 21 months T = (n × H) = 36840

[0157] Fleet usage hours in the past three months h=5895

[0158] Component usage hours in the past three months t=(n×h)=5895

[0159] Number of unplanned replacements in the past three months x = 12

[0160] Average unplanned replacement rate λ = N / T = 0.00168

[0161] Expected number of unplanned replacements in the past three months = λt = 0.00168 × 5895 = 9.9 rounded to the nearest integer = 10

[0162] Referring to the Poisson cumulative probability diagram: the intersection line of λt=10 and 95% probability, it can be concluded that the maximum acceptable number of unplanned replacements (A value) is 15.

[0163] Comparing the current number of unplanned replacements x=12, it can be concluded that the warning value has not been exceeded.

[0164] It can be seen from the calculation formula in Section 3.2.1 that the Poisson cumulative distribution is positively correlated with the number of faults. The greater the number of faults, the larger the Poisson cumulative distribution. The greater the number of faults, the greater the failure rate and the operation guarantee pressure. It can be seen that the Poisson cumulative distribution reflects the following two operational significances: 1. The Poisson cumulative distribution reflects the level of accessory failure rate; 2. The Poisson cumulative distribution reflects the level of accessory operation guarantee pressure.

[0165] The way S1 and S2 are related to each other

[0166] 1. Perform system chapter correlation analysis on maintenance plan items through the ATA100 specification of the Air Transport Association of America, obtain the ATA100 chapter number of the relevant inspection content of the maintenance plan item, and monitor the fault indicators (fleet failure rate, crew reporting rate, ground discovery rate, delay rate) corresponding to the ATA100 chapter number using the S1 and S2 algorithms respectively.

[0167] 2. Associate by key component number or group number

[0168] Analyze the maintenance plan items, obtain the component numbers or component number groups of the inspection contents related to the maintenance plan items, and use the "S2.2 Monitoring system based on reliability warning value performance standards" method for monitoring.

[0169] The above is the entire usage process of a dynamic monitoring method for evaluating the implementation effect of an aircraft maintenance program.

[0170] The present invention establishes a set of dynamic monitoring and analysis processes and methods for evaluating the execution effect of aircraft maintenance programs based on the double-sample equal variance assumption. The key points include:

[0171] (1) An analytical framework that integrates the “S1-Dynamic Monitoring System for Aircraft Maintenance Program Implementation Effectiveness Evaluation” with the “S2-Statistical Performance Standard (Warning Value) Monitoring System”.

[0172] (2)S1.1 Monitoring system display interface, element composition and meaning

[0173] (3) S1.2 Algorithm principle and actual presentation, setting logic of null hypothesis and paradox hypothesis before and after maintenance plan optimization.

[0174] (4) The method of establishing a mutual relationship between S1 and S2 through the Air Transport Association of America (ATA100) specification and key component part numbers or group numbers

[0175] The description of the dynamic monitoring and analysis process and method for evaluating the execution effect of the aircraft maintenance program optimization from S1 to S2 is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that the process sequence can also be modified and improved in several ways without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

[0176] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. A method for dynamically monitoring the execution effect of an aircraft maintenance program, characterized in that: The following steps are involved: Step 1: Record the monthly failure indicators before and after the maintenance plan is optimized; Step 2: Check whether the two population means are equal. First, give the distribution of the difference between the sample means of the two data sets before and after the maintenance plan optimization. Step 3: The combined standard deviation of the samples of the two data sets before and after the maintenance plan optimization; Step 4: Establish a warning value for the fleet system failure rate, which is determined by calculating the standard deviation of the fleet system failure rate (x) per thousand flight hours in the three consecutive months using the flight hours, crew reported failures and ground maintenance findings in the past 14 months; Step 5: The warning value is to use the Poisson cumulative probability distribution table to find the maximum acceptable number of unplanned replacements in a quarter with a probability of 95%, and compare it with the number of unplanned replacements in the current quarter to determine whether it exceeds the limit and alarm; Step six: Use Poisson cumulative distribution to monitor the reliability of accessories.

2. The method for dynamically monitoring the execution effect of an aircraft maintenance plan according to claim 1, characterized in that: In step 1, the monthly failure indicators (fleet failure rate, crew reporting rate, ground discovery rate, and delay rate) before and after the maintenance plan optimization are recorded as The sample mean and sample standard deviation of the monthly failure index before and after optimization are:

3. The method for dynamically monitoring the execution effect of an aircraft maintenance plan according to claim 2, characterized in that: In step 2, because the monthly failure indicators before and after the maintenance plan optimization come from the same data set, the variances of the samples of the two data sets are equal and unknown, and the sample variance satisfies The overall variance of the data is To test whether the two population means are equal, first give the distribution of the difference between the sample means of the two data sets before and after the maintenance plan optimization. According to the assumption, we get: By the Chi-square theorem: Since the sample distributions of the two data sets before and after the maintenance plan optimization are independent, Independence, by additivity of chi-square: Defined by the construction of the distribution, After simplification, we can get: in:

4. The method for dynamically monitoring the execution effect of an aircraft maintenance plan according to claim 3, characterized in that: In the step 3, similarly, under the condition that the sample means of the two data sets before and after the maintenance plan optimization are equal to the overall mean, μ1-μ2=0; Because the test is two-sided, the degrees of freedom are n-1, and the two-sided α gives the two-tailed critical value Establishing the hypothesis: Null hypothesis: There is no significant change in the samples of the two data sets before and after the maintenance plan optimization Paradox hypothesis: There is a significant change in the samples of the two data sets before and after the maintenance plan optimization if: Then the null hypothesis is rejected, and it is believed that the sample means of the two data sets before and after the optimization of the two maintenance schemes are not equal, otherwise the null hypothesis is not rejected; If the null hypothesis is rejected, It represents The mean is significantly less than For example, after the maintenance plan is optimized, the data set is The data set before maintenance plan optimization is The mean failure rate of the data set after the maintenance scheme is optimized is significantly lower than that of the data set before the maintenance scheme is optimized, and the failure rate decreases significantly after the maintenance scheme is optimized; If the null hypothesis is rejected, It represents The mean is significantly greater than If the maintenance plan is optimized, the data set is The data set before maintenance plan optimization is The mean failure rate of the data set after the maintenance plan is optimized is significantly higher than that of the data set before the maintenance plan is optimized, and the failure rate increases significantly after the maintenance plan is optimized.

5. The method for dynamically monitoring the execution effect of an aircraft maintenance plan according to claim 4, characterized in that: In step 5, according to the performance standard "Calculation Method 5" Poisson distribution monitoring attachment provided by AC-121-54, the warning value is to use the Poisson cumulative probability distribution table to find the maximum acceptable number of unplanned replacements in a quarter with a probability of 95%, and compare the number of unplanned replacements in the current quarter to determine whether it exceeds the standard and alarm; Suppose all possible values ​​of the random variable X are 0, 1, 2, ..., and the probability of taking each value is: where k = 0, 1, 2, ..., Where λ is a constant, then X is said to obey the Poisson distribution with parameter λ; where k = 0, 1, 2, ..., The above formula is the probability that the random variable X is less than or equal to the value k, which is called Poisson cumulative distribution.

6. The method for dynamically monitoring the execution effect of an aircraft maintenance plan according to claim 5, characterized in that: Step 6, according to calculation method 5 of advisory circular AC-121-54, adopt Poisson cumulative distribution to monitor the reliability of the attachment; Warning value = the maximum acceptable number of unplanned replacements in a quarter with a probability of 95% found using the Poisson cumulative probability distribution table; Components: Autoflight pitch amplifier; The number of parts per aircraft is n = 1; The number of flight plan replacements in the first 21 months is N = 62; The fleet usage hours in the first 21 months were H = 36,840; Component usage hours in the first 21 months T = (n × H) = 36840; Fleet usage hours in the past three months h = 5895; Component usage hours in the past three months t = (n × h) = 5895; The number of unplanned replacements in the past three months is x = 12; Average unplanned replacement rate λ = N / T = 0.00168; The expected number of unplanned replacements in the past three months = λt = 0.00168 × 5895 = 9.9 rounded to the nearest integer = 10; Referring to the Poisson cumulative probability diagram: the intersection line of λt=10 and 95% probability, it can be concluded that the maximum acceptable number of unplanned replacements (A value) is 15.

Images