Credibility evaluation method of aircraft virtual maintenance training system
By constructing a multi-dimensional evaluation index system and a quantitative scoring method, the problem of inconsistent evaluation results in existing technologies has been solved, thereby improving the credibility and training effectiveness of the aircraft virtual maintenance training system.
Patent Information
- Application Number
- CN202511397437.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-28
- Publication Date
- 2026-02-03
AI Technical Summary
Existing reliability assessment methods for aircraft virtual maintenance training systems suffer from an overemphasis on visual effects while neglecting the accuracy of physical interaction simulation. They also lack objective quantitative indicators and fail to reflect the consistency with the real environment and personnel responses, resulting in inconsistent assessment results and poor training effectiveness.
A multi-dimensional evaluation index system was constructed, including visual simulation accuracy, physical interaction realism, environmental dynamic consistency, adaptability of personnel physiological and psychological reactions, and simulation accuracy of physical attributes of maintenance tools. The weights were determined by the analytic hierarchy process and quantitatively scored by the fuzzy comprehensive evaluation method to calculate the overall credibility of the system.
It significantly improved the objectivity and consistency of the evaluation results, enhanced the correlation between the evaluation results and the actual training effect, and strengthened the credibility of the system and the training effect.
Smart Images
Figure CN121458261A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application belongs to the technical field of aircraft virtual maintenance training, and particularly relates to a credibility evaluation method of an aircraft virtual maintenance training system. BACKGROUND
[0002] With the continuous development of the aviation industry, the aircraft virtual maintenance training system plays an increasingly important role in the training of aircraft maintenance personnel. The system can simulate the real aircraft maintenance environment and operation process, providing a safe and efficient training platform for maintenance personnel. The credibility of the system is a key indicator of whether it can be effectively used for actual maintenance training. A system with high credibility can better meet the training needs and improve the skill level of maintenance personnel.
[0003] Currently, the credibility evaluation method of the aircraft virtual maintenance training system has many defects. Some methods only focus on the visual effect of the system, but pay insufficient attention to the simulation accuracy of the physical interaction of the maintenance operation, making it difficult for the evaluation results to reflect the effectiveness of the system in actual maintenance training. At the same time, the evaluation relies heavily on expert subjective judgment, lacking objective quantitative indicators, resulting in large differences in evaluation results of the same system by different evaluators. In addition, existing methods often ignore the dynamic consistency of the virtual environment and the real maintenance environment, failing to reflect the influence of environmental factors changing over time on maintenance training, further reducing the reliability of the evaluation results. Furthermore, existing evaluation methods do not fully consider the simulation accuracy of the physical properties of maintenance tools, such as weight and hardness, which can affect the realism of maintenance personnel's operation and thus affect the training effect. Moreover, existing evaluation methods lack effective evaluation indicators for the physiological and psychological reactions of maintenance personnel during virtual maintenance, and cannot fully reflect the adaptability of the system to personnel training. Therefore, an improved credibility evaluation method of an aircraft virtual maintenance training system is designed. SUMMARY
[0004] In view of the above deficiencies in the prior art, the present application provides a credibility evaluation method of an aircraft virtual maintenance training system to solve the problems in the background art.
[0005] To solve the above technical problems, the present application adopts the following technical solutions: A credibility evaluation method of an aircraft virtual maintenance training system, comprising the following steps: a. Constructing a multi-dimensional evaluation index system, the evaluation index system comprising 5 evaluation dimensions, namely: visual simulation accuracy dimension, physical interaction reality dimension, environmental dynamic consistency dimension, personnel physiological and psychological reaction adaptability dimension, and maintenance tool physical property simulation accuracy dimension; b. Determining the weight of the comprehensive factor of each evaluation dimension using the analytic hierarchy process; c. The fuzzy comprehensive evaluation method is used to quantitatively score the comprehensive factors of each evaluation dimension, and the quantitative scoring results of the comprehensive factors are obtained; d. According to the quantitative scoring results of the comprehensive factors and the corresponding weights, the credibility comprehensive evaluation value C of the virtual aircraft maintenance training system is calculated, and the credibility level is determined according to the comprehensive evaluation value C.
[0006] Further, in step a, the comprehensive factor of the visual simulation accuracy dimension is a visual simulation comprehensive factor V, and the calculation formula is: , wherein is the maximum deviation of the three-dimensional coordinates of the geometric model, and the upper limit of the coordinate difference between the virtual model and the real part in the three-dimensional space, = 3 mm, which is the unqualified threshold of the index; is the texture resolution, the pixel density of the object surface texture in the virtual scene, = 4K, which is the excellent threshold of the texture resolution; is the light intensity distribution deviation rate, the deviation percentage of the light intensity distribution of the virtual scene from the real environment; is the color difference value, the perceived difference between the color of the virtual object and the color of the real object; = 8, which is the unqualified threshold of the color difference.
[0007] Further, in step a, the comprehensive factor of the physical interaction reality dimension is a physical interaction comprehensive factor P, and the calculation formula is: , wherein is the force feedback accuracy error, the relative deviation percentage of the force feedback value in the virtual operation from the real operation force value; is the collision detection accuracy rate, which is 1 when the collision detection rate is 0 and the false detection rate does not exceed 1%; is the maximum deviation of the motion trajectory, the maximum distance between the virtual operation trajectory and the real operation trajectory in space, = 15 mm, which is the unqualified threshold of the index; is the damping coefficient error, the relative deviation percentage of the damping coefficient in the virtual operation from the real operation damping coefficient.
[0008] Further, in step a, the comprehensive factor of the environmental dynamic consistency dimension is an environmental dynamic comprehensive factor E, and the calculation formula is: , wherein, is the maximum deviation of the temperature, the maximum difference between the virtual environment temperature and the real environment temperature; = 4℃; is the maximum deviation of the humidity, the maximum difference percentage between the virtual environment humidity and the real environment humidity, = 10%; is the noise decibel deviation, the difference between the virtual environment noise decibel value and the real environment noise decibel value, = 10 dB; is the vibration frequency deviation, the difference between the virtual environment vibration frequency and the real environment vibration frequency, = 10 Hz.
[0009] Further, in step a, the comprehensive factor of the personnel physiological and psychological reaction adaptability dimension is a man-machine adaptability comprehensive factor H, and a calculation formula thereof is: , wherein is a virtual and real operation heart rate difference, a difference between a heart rate measurement value of the maintenance personnel in virtual operation and a heart rate measurement value in real operation, = 20; is a key area fixation point coincidence rate, a coincidence proportion of a fixation point of an eye movement track of the maintenance personnel in virtual operation and a fixation point in a key area in real operation; is an operation pressure score difference, a subjective score difference of the maintenance personnel on operation pressure in virtual operation and real operation, = 4; is a fatigue index deviation rate, a relative deviation percentage of fatigue indexes of the maintenance personnel in virtual operation and real operation, = 30%.
[0010] Further, in step a, the comprehensive factor of the maintenance tool physical property simulation accuracy dimension is a tool property comprehensive factor T, and a calculation formula thereof is: , wherein is a weight simulation error, a relative deviation percentage of a virtual tool weight and a real tool weight; is a hardness value deviation, a difference between a virtual tool hardness simulation value and a real tool hardness measurement value, = 30 HB; is a size error, that is, a maximum difference between a virtual tool three-dimensional size and a real tool size, = 2 mm; is an operation hand feeling score difference, that is, a subjective score difference of the maintenance personnel on tool hand feeling in virtual operation and real operation, = 4.
[0011] Further, in step d, a calculation formula of the aircraft virtual maintenance training system credibility comprehensive evaluation value C is: , wherein is a weight of the visual simulation comprehensive factor, is a weight of the physical interaction comprehensive factor, is a weight of the environment dynamic comprehensive factor, is a weight of the man-machine adaptability comprehensive factor, is a weight of the tool property comprehensive factor.
[0012] Further, the maximum deviation of the three-dimensional coordinates of the geometric model is obtained The specific steps include obtaining the three-dimensional coordinate point cloud of the real aircraft part and the coordinate data of the virtual model by using a three-dimensional scanner, and calculating the maximum difference between the two sets of coordinates by using a point cloud matching algorithm. The specific steps include reading the texture file parameters of the virtual model. The light intensity distribution deviation rate ΔL is obtained by placing a light meter at the same position in the real scene and the virtual scene, measuring the light intensity of the corresponding points, and calculating according to the formula ΔL = | virtual light intensity - real light intensity | / real light intensity × 100%. The color difference value ΔE is obtained by measuring the Lab color values of the corresponding areas of the real part and the virtual model using a color difference meter, and calculating according to the formula ΔE = √[(ΔL)²+(Δa²+(Δb)²].
[0013] Further, the temperature deviation ΔT and the humidity deviation ΔH are obtained by recording the temperature and humidity data of the virtual environment and the real maintenance site at the same time period using a temperature and humidity sensor. The noise decibel deviation ΔN is obtained by collecting the noise values of the virtual scene and the real scene in the operation area using a decibel meter, and calculating the maximum value according to ΔN = | virtual noise value - real noise value |. The vibration frequency deviation ΔV is obtained by recording the vibration frequency of the equipment running in the real environment using a vibration sensor, extracting the frequency parameters of the corresponding vibration in the virtual scene, and calculating according to ΔV = | virtual frequency - real frequency |.
[0014] Further, the heart rate difference value ΔHR is obtained by allowing the maintenance personnel to complete the same maintenance task in the virtual system and the real scene, recording the heart rate during the whole process using a wrist heart rate monitor, calculating the heart rate difference value at the same time, and taking the maximum absolute value as ΔHR. The key area gaze point coincidence rate C o is obtained by recording the gaze point coordinates of the maintenance personnel during the operation using an eye tracking instrument, comparing the number of gaze points falling on the key maintenance area in the virtual and real operations, and calculating according to C o = number of coincident points / total number of gaze points.
[0015] Compared with the prior art, the present application has the following advantages: The application covers the core needs of "scene reality - operation reality - environment adaptability - man-machine coordination - tool simulation accuracy" in virtual maintenance training by constructing five dimensions of visual simulation accuracy, physical interaction reality, environment dynamic consistency, personnel physiological and psychological reaction adaptability and maintenance tool physical property simulation accuracy, and ensures that the evaluation result can truly reflect the support ability of the system to the actual training.
[0016] The application determines the weight by analytic hierarchy process (combined with multi-field expert consensus and consistency test), quantifies the score by fuzzy comprehensive evaluation method (based on measured data and membership calculation), converts qualitative description into quantitative indicators (such as visual factor V and physical factor P, which are both numerical values in the interval [0, 1]), so that the difference rate of evaluation results of different evaluators on the same system is reduced to within 10%, and the objectivity and consistency of evaluation are significantly improved.
[0017] By including the physiological and psychological reactions (heart rate, eye movement trajectory, fatigue degree, etc.) of maintenance personnel into the evaluation system, the adaptability of the system to personnel is quantified by the man-machine adaptation comprehensive factor H. This design breaks through the limitation of traditional "only evaluating the system itself", so that the correlation degree between the evaluation result and the actual training effect (such as skill transfer rate and operation proficiency) is improved by more than 30%, and a more accurate direction is provided for system optimization.
[0018] The application quantifies the simulation accuracy of weight, hardness, size and feel by the tool property comprehensive factor T, so as to ensure the operation feedback consistency of virtual tools and real tools. According to the actual measurement, the real feeling score of maintenance personnel on tool operation is improved by 25% by using the system evaluated and optimized by the application, and the skill discontinuity between virtual and real operation is significantly reduced. BRIEF DESCRIPTION OF DRAWINGS
[0019] Figure 1 The flowchart of the credibility evaluation method of the aircraft virtual maintenance training system of the application; DETAILED DESCRIPTION In order to enable those skilled in the art to better understand the application, the technical solutions of the application are further described below with reference to the drawings and examples.
[0020] Among them, the drawings are only used for illustrative explanation, and the representation is only a schematic diagram, not a physical diagram, and cannot be understood as a limitation of the patent; in order to better illustrate the embodiments of the application, some components of the drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product; for those skilled in the art, it is understandable that some known structures and their descriptions in the drawings may be omitted.
[0021] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "inner," and "outer" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present patent. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.
[0022] In the description of this invention, unless otherwise explicitly specified and limited, the term "connection" or similar designation indicating a connection between components should be interpreted broadly. For example, it can refer to a fixed connection, a detachable connection, or an integral part; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can refer to the internal communication between two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.
[0023] like Figure 1 As shown, this invention provides a reliability assessment method for an aircraft virtual maintenance training system, comprising the following steps: a. Construct a multi-dimensional evaluation index system, which includes 5 evaluation dimensions: visual simulation accuracy dimension, physical interaction realism dimension, environmental dynamic consistency dimension, personnel physiological and psychological reaction adaptability dimension, and maintenance tool physical attribute simulation accuracy dimension. b. Use the analytic hierarchy process (AHP) to determine the weights of the comprehensive factors for each evaluation dimension; Specifically, the analytic hierarchy process (AHP) is used to integrate fuzzy expert judgments, transforming qualitative evaluations into quantitative weights. The specific process is as follows: First, a three-level hierarchical structure was constructed (the target layer consists of determined weights, the criterion layer consists of 5 comprehensive factors, and the indicator layer consists of atomic indicators under each factor), and a 7-person expert group was formed (including 3 senior aviation maintenance engineers, 2 virtual simulation algorithm experts, and 2 human-machine ergonomics researchers) to ensure coverage of professional perspectives from multiple fields.
[0024] Next, the experts quantified the importance judgment between factors using triangular fuzzy numbers (composed of the minimum possible value, the most likely value, and the maximum possible value). For example, (1,2,3) represents "significantly important" and (1 / 3,1 / 2,1) represents "slightly unimportant". Each expert compared the five comprehensive factors pairwise to form a 5×5 fuzzy judgment matrix.
[0025] Then, the expert opinions are fused to generate a group matrix: weights are assigned according to the qualifications of the experts (senior engineers have a weight of 0.2, and the rest have a weight of 0.15), and a weighted average is taken to obtain the group fuzzy matrix; if the fuzzy consistency index of the matrix exceeds 0.1, the expert judgment that deviates the most from the group opinion is removed, and the fusion is repeated until the index meets the standard.
[0026] Then, the fuzzy matrix is converted into precise values, the expectation value method is used for calculation, the sum product method is used to solve the eigenvector and normalize it to obtain the initial weight vector; at the same time, the rationality of the weight is checked (the sum is 1 and the single weight is not less than 0.05), and if it does not meet the standard, the expert judgment is adjusted.
[0027] Finally, a dynamic correction mechanism is established: after accumulating 20 groups of evaluation data, if the actual contribution of a factor deviates from the expert's prediction by more than 15%, the weight of the factor is adjusted by the deviation ratio and then normalized to ensure the effectiveness of the weight with the system iteration.
[0028] c. The fuzzy comprehensive evaluation method is used to quantitatively score the comprehensive factors of each evaluation dimension, and the quantitative scoring results of the comprehensive factors are obtained; Specifically, this scheme uses fuzzy mathematical tools to convert the atomic index data of each evaluation dimension into quantitative scores of comprehensive factors. The core process is divided into five stages, as follows: I. Evaluation framework First, five comprehensive factors (visual, physical, environmental, human-machine, and tool) that need to be scored are determined, each factor includes 4-5 atomic indicators (such as visual factor including geometric precision, texture detail, etc.). At the same time, a five-level comment system is set up: "excellent, good, medium, qualified, unqualified", which corresponds to the quantitative intervals of 0.9-1, 0.8-0.9, 0.7-0.8, 0.6-0.7, and 0-0.6 respectively, as the scoring benchmark.
[0029] II. Raw data collection and standardization Data collection: geometric dimensions, force feedback, and other physical parameters are obtained through three-dimensional scanners, force sensors, and other equipment; subjective evaluations such as operation feel are collected using standardized scales (such as 1-10 points); physiological data such as heart rate and eye movement are recorded simultaneously; at least 30 samples from different scenarios are collected for each atomic indicator.
[0030] Data processing: for positive indicators (such as texture resolution, the larger the value, the better), use "measured value / maximum qualified value" for normalization; for negative indicators (such as geometric deviation, the smaller the value, the better), use "1 - measured value / maximum qualified value" for processing, to ensure that the data is within the 0-1 interval.
[0031] III. Membership calculation (core link) Membership function design: Continuous indicators (e.g. temperature deviation, force feedback error) use trapezoidal membership functions: with "excellent", "good" thresholds as nodes, membership degrees are calculated by linear interpolation. For example, force feedback error ≤ 5% is "excellent", 5%-10% is "good", if the actual error is 7%, then the "excellent" membership degree is 0.3, and the "good" membership degree is 0.7.
[0032] Discrete indicators (e.g. hand feel score) are proportionally mapped: if the full score is 5, 4 points correspond to "excellent" membership degree 0.8, "good" 0.2.
[0033] Generate membership degree vector: each atomic indicator calculates its membership degree in the five-level comment according to the measured value, forming a five-dimensional vector. For example, geometric deviation 1.2mm (excellent ≤1mm, good ≤2mm), the vector is (0.4, 0.6, 0, 0, 0), representing 40% "excellent", 60% "good".
[0034] Four, synthesis of comprehensive factor score Construct fuzzy matrix: arrange the membership degree vectors of all atomic indicators under the same comprehensive factor by rows to form a matrix (the number of rows is the number of atomic indicators, and the number of columns is 5-level comments).
[0035] Atomic indicator weighting: use AHP to calculate the weight of each atomic indicator in the comprehensive factor it belongs to (e.g. geometric accuracy accounts for 30% in visual factor, texture accounts for 20%, etc.), forming a weight vector.
[0036] Fuzzy synthesis and score conversion: multiply the weight vector and the fuzzy matrix (weighted average calculation) to get the membership degree vector of the comprehensive factor in the five-level comment; then use the benchmark score of the five-level comment (1.0, 0.8, 0.6, 0.4, 0.2) to calculate the final quantitative score (retain two decimal places). For example, if the membership degree vector of a factor is (0.12, 0.34, 0.28, 0.26, 0), the score = 0.12x1.0 + 0.34x0.8 + 0.28x0.6 + 0.26x0.4 = 0.65.
[0037] Five, result verification and adjustment Post-scoring check: if the result is out of the range of 0-1, the membership function or weight needs to be rechecked; if the same factor has large fluctuations in multiple scores (standard deviation > 0.15), increase the sample size (at least 50 groups) for re-evaluation, and finally output the quantitative scores of each comprehensive factor (e.g. visual 0.65, physical 0.82) as the basic data for comprehensive evaluation of system reliability.
[0038] d. According to the quantitative score results of each comprehensive factor and the corresponding weight, the credibility comprehensive evaluation value C of the aircraft virtual maintenance training system is calculated, and the credibility level is determined according to the comprehensive evaluation value C.
[0039] Specifically, in step c, the fuzzy comprehensive evaluation method comprises the following sub-steps: c1. Determine the evaluation levels of "excellent, good, medium, qualified, unqualified" and the corresponding reference points; c2. Establish a fuzzy relationship matrix combined with actual test data to determine the membership degree of each sub-index to the evaluation level; c3. Perform fuzzy synthesis operation by using a weighted average type synthesis operator; c4. Convert the fuzzy evaluation result into a specific quantitative score by using a weighted average method In step a, the comprehensive factor of the visual simulation accuracy dimension is the visual simulation comprehensive factor V, and the calculation formula is: , wherein is the maximum deviation of the three-dimensional coordinates of the geometric model, and the upper limit of the coordinate difference between the virtual model and the real part in the three-dimensional space, = 3mm, which is the unqualified threshold of this index; is the texture resolution, which is the pixel density of the object surface texture in the virtual scene, = 4K, which is the excellent threshold of the texture resolution; is the light intensity distribution deviation rate, which is the percentage of deviation of the virtual scene light intensity distribution from the real environment; is the color difference value, which is the perceived difference between the color of the virtual object and the color of the real object; = 8, which is the unqualified threshold of the color difference.
[0040] In step a, the comprehensive factor of the physical interaction reality dimension is the physical interaction comprehensive factor P, and the calculation formula is: , wherein is the force feedback accuracy error, which is the relative deviation percentage of the force feedback value in the virtual operation from the real operation force value; is the collision detection accuracy rate, which is 1 when the collision detection rate is 0 and the false detection rate does not exceed 1%; is the maximum deviation of the motion trajectory, which is the maximum distance between the virtual operation trajectory and the real operation trajectory in space, = 15mm, which is the unqualified threshold of this index; is the damping coefficient error, which is the relative deviation percentage of the damping coefficient in the virtual operation from the real operation damping coefficient.
[0041] In step a, the comprehensive factor of the environmental dynamic consistency dimension is the environmental dynamic comprehensive factor E, and the calculation formula is: , wherein, is a temperature maximum deviation, the maximum difference between the virtual environment temperature and the real environment temperature; = 4℃; is a humidity maximum deviation, the maximum difference between the virtual environment humidity and the real environment humidity percentage, = 10%; is a noise decibel deviation, the difference between the virtual environment noise decibel value and the real environment noise decibel value, = 10dB; is a vibration frequency deviation, the difference between the virtual environment vibration frequency and the real environment vibration frequency, = 10Hz.
[0042] In step a, the comprehensive factor of the personnel physiological and psychological reaction adaptability dimension is a man-machine adaptation comprehensive factor H, and the calculation formula is: , wherein is a virtual and real operation heart rate difference, the difference between the heart rate measured values of the maintenance personnel in virtual operation and real operation, = 20; is a key area fixation point coincidence rate, the coincidence proportion of the fixation points of the eye movement track of the maintenance personnel in virtual operation and the fixation points in the key area in real operation; is an operation pressure score difference, the subjective score difference of the maintenance personnel on the operation pressure in virtual operation and real operation, = 4; is a fatigue index deviation rate, the relative deviation percentage of the fatigue index of the maintenance personnel in virtual operation and real operation, = 30%.
[0043] In step a, the comprehensive factor of the maintenance tool physical property simulation accuracy dimension is a tool property comprehensive factor T, and the calculation formula is: , wherein is a weight simulation error, the relative deviation percentage of the virtual tool weight and the real tool weight; is a hardness value deviation, the difference between the virtual tool hardness simulation value and the real tool hardness measured value, = 30HB; is a size error, that is, the maximum difference between the virtual tool three-dimensional size and the real tool size, = 2mm; is an operation hand feeling score difference, that is, the subjective score difference of the maintenance personnel on the tool hand feeling in virtual operation and real operation, = 4.
[0044] In step d, the calculation formula of the credibility comprehensive evaluation value C of the aircraft virtual maintenance training system is: , wherein is the weight of the visual analog composite factor, is the weight of the physical interaction composite factor, is the weight of the environmental dynamics composite factor, is the weight of the human-machine fit composite factor, is the weight of the tool property composite factor.
[0045] wherein the maximum deviation of the three-dimensional coordinates of the geometric model is obtained , specifically comprising obtaining the three-dimensional coordinate point cloud of the real aircraft part and the coordinate data of the virtual model by using a three-dimensional scanner, and calculating the maximum difference of the two groups of coordinates by a point cloud matching algorithm; obtaining the texture resolution R specifically comprises reading the texture file parameters of the virtual model; The acquisition of the light intensity distribution deviation rate ΔL includes placing a light meter at the same position in the real scene and the virtual scene, measuring the light intensity of the corresponding points, and calculating according to the formula ΔL=| virtual light intensity - real light intensity | / real light intensity × 100%; The acquisition of the color difference value ΔE includes measuring the Lab color values of the corresponding areas of the real part and the virtual model by using a color difference meter, and calculating the color difference value according to the formula ΔE=√[(ΔL)²+(Δa²+(Δb)²].
[0046] wherein the acquisition of the temperature deviation ΔT and the humidity deviation ΔH includes recording the temperature and humidity data of the virtual environment and the real maintenance site at the same time period by using a temperature and humidity sensor; The acquisition of the noise decibel deviation ΔN includes collecting the noise values of the virtual scene and the real scene in the operation area by using a decibel meter, and calculating the maximum value according to ΔN=| virtual noise value - real noise value |; The acquisition of the vibration frequency deviation ΔV includes recording the vibration frequency of the equipment running in the real environment by using a vibration sensor, extracting the frequency parameters of the corresponding vibration in the virtual scene, and calculating according to ΔV=| virtual frequency - real frequency |.
[0047] wherein the acquisition of the heart rate difference value ΔHR includes making the maintenance personnel complete the same maintenance task in the virtual system and the real scene, recording the whole process heart rate by using a wrist heart rate monitor, calculating the heart rate difference value at the same time, and taking the maximum absolute value as ΔHR; The acquisition of the key area fixation point coincidence rate C o includes recording the fixation point coordinates of the maintenance personnel in the operation process by using an eye tracking instrument, comparing the number of fixation points falling in the key maintenance area in the virtual and real operations, and calculating according to C o = number of coincident points / total number of fixation points.
[0048] Example one: A certain type of Boeing 737 aircraft CFM56 engine virtual disassembly training system is used as the evaluation object, which is used to simulate the disassembly process of engine core components (such as high-pressure turbine, combustion chamber), which needs to be evaluated by the method.
[0049] II. Evaluation steps and data Step 1: Build evaluation index system Adopt the 5 evaluation dimensions described in the patent, and the comprehensive factors and atomic indicators of each dimension are as follows: Visual simulation accuracy (V): geometric model deviation (ΔG), texture resolution (R), light deviation (ΔL), color difference value (ΔE); Physical interaction authenticity (P): force feedback error (ΔF), collision detection accuracy (Ac), motion trajectory deviation (ΔM), damping error (ΔD); Environmental dynamic consistency (E): temperature deviation (ΔT), humidity deviation (ΔH), noise deviation (ΔN), vibration frequency deviation (ΔV); Physiological and psychological adaptability of personnel (H): heart rate difference (ΔHR), gaze point coincidence rate (Co), operation stress difference (ΔS), fatigue deviation rate (ΔF); Tool attribute simulation accuracy (T): weight error (ΔW), hardness deviation (ΔB), size error (ΔQ), hand feel score difference (ΔJ.
[0050] Step 2: Determine the weight using the analytic hierarchy process Assemble a 7-person expert group (3 aviation maintenance engineers, 2 virtual simulation experts, and 2 human-machine ergonomics researchers), calculate the weight through the judgment matrix as follows: ω V =0.20, ω P =0.30, ω E =0.10, ω H =0.25, ω T =0.15 (the sum is 1, CR=0.08<0.1, pass the consistency test).
[0051] Step 3: Quantitative scoring using fuzzy comprehensive evaluation method 3.1 Data collection (part of the key data) Visual dimension: Geometric model deviation ΔG=1.5mm (three-dimensional scanner measurement, maximum coordinate difference between real part and virtual model); Texture resolution R=4K (system texture file parameter); Light deviation ΔL=5% (light measurement, virtual and real scene light intensity deviation); Color difference value ΔE=2 (color difference instrument measurement, Lab color space calculation).
[0052] Physical interaction dimension: Force feedback error ΔF=8% (force sensor record, relative deviation of virtual tightening force and real force value); Collision detection accuracy A_c=1 (10 groups of collision scenes have no missed detection, and the false detection rate is 0.5%≤1%); Motion trajectory deviation ΔM=10mm (motion capture system record, maximum spatial difference of tool trajectory); Damping error ΔD=6% (damping tester compares the virtual and real operation resistance coefficients).
[0053] Environmental dimension: Temperature deviation ΔT=2℃ (temperature and humidity sensor record, maximum temperature difference between virtual environment and real maintenance workshop); Humidity deviation ΔH=5% (same as temperature collection, maximum humidity deviation); Noise deviation ΔN=3dB (decibel instrument measurement, difference between virtual engine noise and real value); Vibration frequency deviation ΔV=2Hz (vibration sensor record, frequency difference between virtual vibration and real idle vibration).
[0054] Human-machine dimension: Heart rate difference ΔHR=12 times / minute (after 10 maintenance personnel operate, the maximum heart rate difference between virtual and real operation); Gaze point coincidence rate C_o=0.75 (eye tracker record, gaze point coincidence rate of key operation area); Operation stress difference ΔS=1 (9-point scale score, stress score difference between virtual and real operation); Fatigue deviation rate ΔF=15% (fatigue monitor data, relative deviation of fatigue index between virtual and real operation).
[0055] Tool dimension: Weight error ΔW=5% (electronic scale measurement, relative deviation of virtual wrench and real wrench weight); Hardness deviation ΔB=10HB (Brinell hardness tester measurement, hardness difference between virtual and real tool); Size error ΔQ=0.8mm (vernier caliper measurement, maximum size difference between virtual tool and real tool); Hand feel score difference ΔJ=1 (5-point scale score, hand feel score difference between virtual and real operation).
[0056] 3.2 Comprehensive factor calculation Visual factor V=0.356 Physical factor P≈0.291 Environmental factor E = 0.14 Human-machine factor H = 0.1125 Tool factor T ≈ 0.285 Step 4: Calculate the comprehensive evaluation value and divide the level Comprehensive evaluation value C: C = 0.20x0.356 + 0.30x0.291 + 0.10x0.14 + 0.25x0.1125 + 0.15x0.285 ≈ 0.071 + 0.087 + 0.014 + 0.028 + 0.043 = 0.243 Level division: C = 0.243 < 0.6, judged as "unreliable".
[0057] III. Evaluation Conclusion The reliability of the engine virtual disassembly training system is "unreliable", and the main shortcomings are the physical interaction authenticity (such as large trajectory deviation) and human-machine adaptability (heart rate difference, insufficient gaze point coincidence rate), which need to optimize the system parameters (such as reducing the virtual tool trajectory deviation, adjusting the interaction interface to improve the personnel operation adaptability) for these dimensions.
[0058] The above is only an embodiment of the present application, and the circuit, electronic components and modules involved are all prior art. Those skilled in the art can implement them without further description. The content protected by this application does not involve the improvement of software and methods. The specific structure and characteristics of the scheme known to the public are not described in detail here. Those skilled in the art know all the ordinary technical knowledge in the field of the application before the application date or the priority date, can know all the prior art in this field, and have the ability to apply conventional experimental means before the date. Those skilled in the art can improve and implement the present scheme under the guidance of this application, and some typical known structures or known methods should not be an obstacle for those skilled in the art to implement this application. It should be noted that for those skilled in the art, without departing from the structure of the present application, a number of modifications and improvements can be made, which should be considered as the protection scope of the present application. These will not affect the effect and practicality of the patent.
Claims
1. A method for assessing the credibility of an aircraft virtual maintenance training system, the method comprising: The method comprises the following steps: a. Constructing a multi-dimensional evaluation index system, which comprises five evaluation dimensions, namely visual simulation accuracy dimension, physical interaction reality dimension, environment dynamic consistency dimension, personnel physiological and psychological reaction adaptability dimension, and maintenance tool physical property simulation accuracy dimension; b. Determining the weight of the comprehensive factor of each evaluation dimension by using the analytic hierarchy process; c. Quantitatively scoring the comprehensive factor of each evaluation dimension by using the fuzzy comprehensive evaluation method to obtain the quantitative scoring result of the comprehensive factor; d. According to the quantitative scoring result of each comprehensive factor and the corresponding weight, calculating the credibility comprehensive evaluation value C of the aircraft virtual maintenance training system, and dividing the credibility level according to the comprehensive evaluation value C.
2. The method of claim 1, wherein: In step a, the comprehensive factor of the visual simulation accuracy dimension is visual simulation comprehensive factor V, and the calculation formula is: , wherein is the maximum deviation of the geometric model three-dimensional coordinates, the upper limit of the coordinate difference between the virtual model and the real part in the three-dimensional space, = 3mm, the unqualified threshold of the index; is the texture resolution, the pixel density of the object surface texture in the virtual scene, = 4K, the excellent threshold of the texture resolution; is the light intensity distribution deviation rate, the deviation percentage of the virtual scene light intensity distribution and the real environment; is the color difference value, the perceived difference amount of the color of the virtual object and the color of the real object; = 8, the unqualified threshold of the color difference.
3. The method of claim 1, wherein: In step a, the comprehensive factor of the physical interaction authenticity dimension is a physical interaction comprehensive factor P, and a calculation formula thereof is: , wherein is a force feedback accuracy error, a relative deviation percentage of a force feedback value in virtual operation and a real operation force value; is a collision detection accuracy, and the value is 1 when a collision missed detection rate is 0 and a false detection rate is not more than 1%; is a maximum trajectory deviation, a maximum distance in space between a virtual operation trajectory and a real operation trajectory, = 15 mm, and the index unqualified threshold value is; is a damping coefficient error, a relative deviation percentage of a damping coefficient in virtual operation and a real operation damping coefficient.
4. The method of claim 1, wherein: In step a, the comprehensive factor of the environmental dynamic consistency dimension is an environmental dynamic comprehensive factor E, and a calculation formula thereof is: , wherein, is a temperature maximum deviation, a maximum difference between a virtual environment temperature and a real environment temperature; = 4℃; is a humidity maximum deviation, a maximum difference percentage between a virtual environment humidity and a real environment humidity, = 10%; is a noise decibel deviation, a difference between a virtual environment noise decibel value and a real environment noise decibel value, = 10dB; is a vibration frequency deviation, a difference between a virtual environment vibration frequency and a real environment vibration frequency, = 10Hz.
5. The method of claim 1, wherein: In step a, the comprehensive factor of the personnel physiological and psychological response adaptability dimension is human-machine adaptability comprehensive factor H, and the calculation formula is: , wherein is the difference between the virtual and real operation heart rates, the difference between the measured values of the heart rate of the maintenance personnel in the virtual operation and the real operation, = 20; is the key area fixation point coincidence rate, the coincidence proportion of the fixation points of the eye movement track of the maintenance personnel in the virtual operation and the fixation points in the key area in the real operation; is the operation pressure score difference value, the difference between the subjective scores of the operation pressure of the maintenance personnel in the virtual operation and the real operation, = 4; is the fatigue index deviation rate, the relative deviation percentage of the fatigue index of the maintenance personnel in the virtual operation and the real operation, = 30%.
6. The method of claim 1, wherein: In step a, the comprehensive factor of the maintenance tool physical property simulation accuracy dimension is a tool property comprehensive factor T, and a calculation formula thereof is: wherein is a weight simulation error, a relative deviation percentage of a virtual tool weight and an actual tool weight; is a hardness value deviation, a difference between a virtual tool hardness simulation value and an actual tool hardness measurement value, = 30 HB; is a size error, that is, a maximum difference between a virtual tool three-dimensional size and an actual tool size, = 2 mm; is an operation feeling score difference value, that is, a subjective score difference value of a tool feeling of a maintenance personnel in a virtual operation and an actual operation, = 4.
7. The method of claim 1, wherein: In step d, the calculation formula of the credibility comprehensive evaluation value C of the aircraft virtual maintenance training system is: , wherein is the weight of the visual simulation comprehensive factor, is the weight of the physical interaction comprehensive factor, is the weight of the environment dynamic comprehensive factor, is the weight of the man-machine adaptation comprehensive factor, is the weight of the tool attribute comprehensive factor.
8. The method of claim 2, wherein: the virtual maintenance training system is a virtual aircraft maintenance training system; and the virtual maintenance training system comprises a plurality of virtual maintenance training system components, wherein each of the plurality of virtual maintenance training system components is associated with a respective one of a plurality of virtual maintenance training system component types. Wherein the maximum deviation of the three-dimensional coordinates of the geometric model is obtained The specific includes that the three-dimensional coordinate point cloud of the real aircraft part and the coordinate data of the virtual model are respectively obtained by using a three-dimensional scanner, and the maximum difference value of the two groups of coordinates is calculated by a point cloud matching algorithm. The texture resolution R is obtained by reading the texture file parameters of the virtual model; The light intensity distribution deviation rate AL is obtained by placing a light meter at the same position in the real scene and the virtual scene, measuring the light intensity of the corresponding points, and calculating according to the formula AL=| virtual light intensity - real light intensity | / real light intensity × 100%; The color difference value ΔE is obtained by measuring the Lab color values of the corresponding areas of the real part and the virtual model with a color difference meter, and calculating according to the formula ΔE=√[(ΔL)²+(Δa²+(Δb)²].
9. The method of claim 4, wherein: The temperature deviation AT and the humidity deviation AH are obtained by recording the temperature and humidity data of the virtual environment and the real maintenance site at the same time period with a temperature and humidity sensor; The noise decibel deviation ΔN is obtained by collecting the noise values of the virtual scene and the real scene in the operation area with a decibel meter, and calculating the maximum value according to ΔN=| virtual noise value - real noise value |; The vibration frequency deviation ΔV is obtained by recording the vibration frequency of the equipment running in the real environment with a vibration sensor, extracting the frequency parameters of the corresponding vibration in the virtual scene, and calculating according to ΔV=| virtual frequency - real frequency |.
10. The method of claim 5, wherein: the virtual maintenance training system is for an aircraft; and the virtual maintenance training system is for a plurality of aircraft. 5 The heart rate difference value ΔHR is obtained by making the maintenance personnel complete the same maintenance task in the virtual system and the real scene, recording the whole heart rate with a wrist heart rate monitor, calculating the heart rate difference value at the same time, and taking the maximum absolute value as ΔHR. Key area gaze point coincidence rate C o The acquisition includes recording the gaze point coordinates of the maintenance personnel during the operation process with an eye tracking instrument, comparing the number of gaze points falling on the key maintenance area in virtual and real operations, and calculating according to C o = the number of coincidences / the total number of gaze points.