Method for evaluating advanced control plan performance of aero-engine

Abstract

A method for evaluating the advanced control plan performance of an aero-engine comprises the following steps: acquiring an influence factor of the control plan performance of the engine based on an execution demand, and establishing a step hierarchical structure and a judgment matrix aij of the influence factor; defining an evaluation grade interval of advanced control plan performance and a corresponding quantitative index according to a performance evaluation model of a fuzzy theory, and calculating a performance value Wq after standard quantification of the influence factor; performing consistency check on the judgment matrix aij, performing hierarchical sorting according to the influence proportion of the influence factors on the overall performance of the control plan, and determining a weight matrix ki of the influence factors; according to the performance value Wq and the weight matrix ki of the influence factors, a comprehensive evaluation value AI is obtained, and whether the engine control plan meets the requirement or not is judged. According to the technical scheme, whether the control plan with the abstract concept meets the requirement or not can be quantitatively evaluated, and the advantages and disadvantages of engine control plan design can be evaluated in a multi-dimensional mode.

Description

Technical Field

[0001] The present invention relates to the technical field of aero-engine control, and in particular to a method for evaluating the performance of an aero-engine advanced control plan. Background Art

[0002] Engine control plan is an abstract concept, but it can be regarded as a product. Since there are many factors that actually affect the various performance indicators of the product, and each performance indicator is independent or restricted from each other, and it is impossible to control each indicator at the same performance level during production, and each indicator has different degrees of influence on the overall performance and different dimensions of influence, it is a relatively complex issue to comprehensively evaluate the performance level of the product.

[0003] Many researchers at home and abroad have proposed many different methods, such as: matter-element analysis method, gray system comprehensive evaluation method for product design quality, product quality evaluation method based on rough set theory and genetic algorithm, etc.

[0004] For example, patent publication number CN119356274A discloses a robustness evaluation method for an aircraft engine control plan. The method includes the following steps: defining mathematical models for aircraft engine components and obtaining robustness performance factors from these mathematical models; the mathematical models include: fan, compressor, high-pressure turbine, and low-pressure turbine mathematical models; defining an aircraft engine control plan robustness evaluation function, expressed as: where to Exn are the accumulated robustness metrics of key performance indicators under different performance factor perturbations, and α1, α2, …, αn are the proportions of the robustness metrics of the key performance indicators; and defining a robustness evaluation index, Q, to determine whether the robustness of the aircraft engine control plan meets the design standards based on the value of the evaluation function. This technical solution forms a comprehensive multi-index robustness calculation for control plans, enabling robustness evaluation of engine control plans. While this solution evaluates the robustness of control plans, it does not provide an evaluation method for indicators that are not in the same dimension and have different impacts on overall engine performance.

[0005] In 1965, American professor Zade proposed the "Fuzzy Set" theory. The fuzzy comprehensive evaluation method based on the fuzzy set theory can well realize multi-level comprehensive evaluation under the conditions of multiple indicators and index standard values being intervals, and has become a hot topic of current research in this field. This method can quantify the relative difference of indicators to the index standard value intervals at all levels, thereby determining the relative membership of indicators with index standard values being intervals.

[0006] Therefore, under the background of fuzzy comprehensive evaluation method, a special set of evaluation methods is designed to quantitatively evaluate whether the control plan with abstract concepts meets the requirements and whether the performance is optimal. Summary of the Invention

[0007] To solve the above technical problems, the present application provides a method for evaluating the performance of an advanced control plan for an aircraft engine, comprising the following steps:

[0008] Obtain the execution requirements of the engine, obtain the influencing factors of the engine control plan performance based on the execution requirements, establish a hierarchical structure of the influencing factors, and construct a judgment matrix a i j;

[0009] Define the evaluation level interval and corresponding quantitative index of the performance of the advanced control plan. The evaluation level interval is divided according to the performance evaluation model of fuzzy theory. The impact factor is standardized and mapped to the evaluation level interval. The performance value W after the impact factor standard quantification is calculated. q ;

[0010] For the judgment matrix a ij Perform consistency test. After completing the consistency test, perform hierarchical sorting based on the influence ratio of the influencing factors on the overall performance of the control plan, and determine the weight matrix k of the influencing factors. i ;

[0011] According to the performance value W q The weight matrix k of the impact factor i , using the linear weighted average method to obtain the comprehensive evaluation value AI; where i represents the row of the matrix, j represents the column of the matrix, and q represents the impact factor;

[0012] According to the comprehensive evaluation value AI, it is judged whether the engine control plan meets the requirements.

[0013] Furthermore, the calculation formula of the comprehensive assessment value AI is: Where i∈[1,n], n represents the matrix order.

[0014] Furthermore, the performance value W after the impact factor standard is quantified q The calculation formula is: q =W0+ΔW q , where W0 is the scoring base value, ΔW q It is the cumulative added points.

[0015] Furthermore, the cumulative added value ΔW q The value standard is:

[0016] when When ΔW q =2;

[0017] when When ΔW q =1;

[0018] when When ΔW q =0.5;

[0019] when When ΔW q =0,

[0020] Among them, A0 is the assessment index, A q For quantitative indicators.

[0021] Furthermore, the quantitative indicator A q The determination method is: normalize and map the impact factor to the evaluation grade interval. When the mapping result is close to the left value of the grade interval, the lower limit of the interval is taken; when the mapping result is close to the right value of the grade interval, the upper limit of the interval is taken.

[0022] Furthermore, the judgment matrix a ij For pairwise comparison matrices, specifically: using a 1-9 scaling method to sequentially scale and compare the impact factors of each matrix.

[0023] Furthermore, the calculation formula of the consistency ratio CR is:

[0024] Among them, CI is the consistency index and RI is the random consistency index.

[0025] Furthermore, the calculation formula of the consistency index CI is: The RI is the sample mean calculated by taking a sufficiently large sample of CI, where λ max is the judgment matrix a ij The maximum eigenvalue of .

[0026] Furthermore, the maximum eigenvalue λ max It is achieved through the sum-product method. The specific steps are:

[0027] First, the judgment matrix a ij Normalize each column element of

[0028] Secondly, the normalized matrix aij Sum by row;

[0029] Then, the matrix a ij The matrix obtained by row summation is normalized to obtain the matrix M i ,

[0030] Then, the approximate eigenvector a is established ij M i ;

[0031] Finally, calculate the maximum eigenvalue λ of the approximate eigenvector max ,

[0032] Furthermore, the weight matrix K of the impact factor i The calculation method is as follows:

[0033] First, the judgment matrix a ij Normalize each column element of

[0034] Secondly, the normalized matrix is averaged row by row;

[0035] Finally, the normalized matrix is combined into a square matrix by row-averaged matrix, and normalized with each non-zero element forming a set to obtain the normalized impact factor weight matrix k i .

[0036] The beneficial effects of the present invention are: through the constructed control plan performance evaluation method, performance influencing factors and quantified influencing factors can be evaluated in multiple dimensions, and whether the control plan with abstract concepts meets the requirements and whether the performance reaches the optimal evaluation method can be evaluated in multiple dimensions and in a relatively comprehensive manner. It has great reference significance for the performance evaluation of other similar abstract concepts and has great prospects for promotion and application. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 The present invention is a flowchart of a method for evaluating the performance of an advanced control plan for an aero-engine according to an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The technical solution of the present invention is further described below, but the scope of protection claimed is not limited to the description.

[0039] An embodiment of the present invention provides a method for evaluating the performance of an advanced control plan of an aircraft engine, comprising the following steps:

[0040] Step S100: Obtain the execution requirements of the engine, obtain the influencing factors of the engine control plan performance based on the execution requirements, establish a hierarchical structure of the influencing factors, and construct a judgment matrix a ij .

[0041] Step S200: define the evaluation level interval and corresponding quantitative index of the performance of the advanced control plan. The evaluation level interval is divided according to the performance evaluation model of fuzzy theory, the impact factor is standardized and mapped to the evaluation level interval, and the performance value W after the impact factor standard is quantified is calculated. q , where q represents the impact factor.

[0042] Step S300: the judgment matrix a ij Perform consistency test. After completing the consistency test, perform hierarchical sorting based on the influence ratio of the influencing factors on the overall performance of the control plan, and determine the weight matrix k of the influencing factors. i .

[0043] Step S400: According to the performance value W q The weight matrix k of the impact factor i , using the linear weighted average method to obtain the comprehensive evaluation value AI.

[0044] The calculation formula of the comprehensive assessment value AI is:

[0045]

[0046] Where i∈[1,n], n represents the matrix order.

[0047] Step S500: judging whether the engine control plan meets the requirements according to the comprehensive evaluation value AI. If the control plan performance represented by the comprehensive index AI meets the requirements and the standard quantified performance value W of the performance influencing factor is q When the requirements are also met, the control plan performance is considered to be fit for purpose.

[0048] In this embodiment, in step S200, the evaluation level interval is 4 levels: unqualified, good, qualified, and excellent. The corresponding relationship between the evaluation level interval and the quantitative index is shown in Table 1:

[0049] Table 1 Evaluation levels and corresponding quantitative indicators

[0050] Judging level Unqualified qualified good excellent Quantitative indicators [0,6) [6,7) [7,8) [8,10]

[0051] Performance impact factors need to be divided from top to bottom, and performance impact factor performance evaluation needs to start from the bottom up, that is, first carry out the first-level evaluation, and then start the second-level evaluation based on the evaluation results of the first level.

[0052] However, since different performance influencing factors represent different dimensions, when using the weighted average method, the performance of different dimensions needs to be mapped to the same interval. The mapping process is to establish a membership function, and the membership function is to establish a mapping to the standard domain to reflect the degree to which the control plan has a certain fuzzy property, or to use fuzzy mathematics theory for standardization. The mapping process can be implemented using the basic scoring and penalty function method.

[0053] In this embodiment, in step S200, the performance value W after the impact factor standard is quantified q The calculation formula is:

[0054] W q =W0+ΔW q (2)

[0055] Among them, W0 is the scoring base value, ΔW q It is the cumulative added points.

[0056] Based on the evaluation level interval mapped to the standardized impact factor, by comparing the quantitative index A q The difference between the assessment index A0 sets the cumulative bonus value ΔW q , the cumulative added value ΔW q The value standard is:

[0057] when When ΔW q =2;

[0058] when When ΔW q =1;

[0059] when When ΔW q =0.5;

[0060] when When ΔW q =0,

[0061] Among them, A0 is the assessment index, A q For quantitative indicators.

[0062] The quantitative indicator A q The determination method is as follows: the impact factor is standardized and mapped to the evaluation level interval. When the mapping result is close to the left value of the level interval, the lower limit value of the quantitative indicator corresponding to the interval is taken; when the mapping result is close to the right value of the level interval, the upper limit value of the quantitative indicator corresponding to the interval is taken.

[0063] In this embodiment, step S100, the control plan performance evaluation involves ranking and optimizing multiple criteria / factors, with each criterion / factor having a certain hierarchical relationship. Therefore, the AHP (Analytical Hierarchy Process) is employed, combining qualitative and quantitative methods, to refine the hierarchical relationships of the factors, thereby establishing a hierarchical structure of the influencing factors.

[0064] The judgment matrix a is established based on the hierarchical structure of the impact factors. ij , the judgment matrix a ijFor a pairwise comparison matrix, specifically: use the 1-9 scaling method to compare the influencing factors of each matrix in turn. Take the criterion / factor layer as the first element, and arrange the indicator layer elements belonging to this element in the first row and the first column in turn. To ensure that the judgment matrix a ij To be consistent, usually a ij =1 / a ji , and a ii =1, where i represents the row of the matrix and j represents the column of the matrix. The factors of each matrix are then compared in turn. The evaluation criteria of the 1-9 scaling method are shown in Table 2:

[0065] Table 2 1-9 scaling method and corresponding meaning

[0066] scale Definition and Explanation 1 Two factors are equally important for a criterion 3 When comparing two factors, one is slightly more important than the other 5 When comparing two factors, one is significantly more important than the other. 7 When comparing two factors, one is much more important than the other 9 When comparing two factors, one is extremely more important than the other. 2，4，6，8 It represents the scale when the adjacent scales mentioned above are compromised.

[0067] In this embodiment, step S300, considering the judgment matrix a ij The value of is inconsistent with the comparison benchmark and scale, that is, the weight coefficient of each performance influencing factor is unreasonable, resulting in unreasonable values mapped to the standard interval, and the final performance evaluation may be inaccurate. Therefore, it is necessary to adjust the judgment matrix a ij Conduct consistency test. For multi-order judgment models, test through consistency ratio CR.

[0068] The calculation formula of the consistency ratio CR is:

[0069]

[0070] Among them, CI is the consistency index and RI is the random consistency index.

[0071] The calculation formula of the consistency index CI is:

[0072]

[0073] The RI is the sample mean calculated by taking a sufficiently large sample of CI, where λ max is the judgment matrix a ij The maximum eigenvalue of , the random consistency index RI of the existing 1 to 15 order calculation results refer to Table 3:

[0074] Table 3 Random consistency index RI of order 1 to 15

[0075] n 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 RI 0 0.5 0.52 0.89 1.12 1.26 1.36 1.41 1.46 1.49 1.52 1.54 1.56 1.58 1.59

[0076] When the consistency ratio CR is less than 0.15, it is considered that the consistency is qualified and the evaluation of each performance factor is reasonable. If the consistency is greater than 0.15, it is necessary to re-value according to the method in Table 2.

[0077] The maximum eigenvalue λ max It is achieved through the sum-product method. The specific steps are:

[0078] First, the judgment matrix a ij Normalize each column element of

[0079] Secondly, the normalized matrix a ij Sum by row;

[0080] Then, the matrix a ij The matrix obtained by row summation is normalized to obtain the matrix M i ,

[0081] Then, the approximate eigenvector a is established ij M i ;

[0082] Finally, calculate the maximum eigenvalue λ of the approximate eigenvector max ,

[0083]

[0084] In this embodiment, step S300, after completing the consistency check, perform hierarchical sorting to determine the weight matrix k of each performance impact factor mapped to the same interval. i .

[0085] The weight matrix k of the impact factor i The calculation method is as follows:

[0086] First, the judgment matrix a ij Normalize each column element of

[0087] Secondly, the normalized matrix is averaged row by row;

[0088] Finally, the normalized matrix is combined into a square matrix by row-averaged matrix, and normalized with each non-zero element forming a set to obtain the normalized impact factor weight matrix k i .

[0089] In this embodiment, simulation, semi-physical testing, and retrograde verification may be used to support the validity of the control plan performance evaluation results.

[0090] The semi-physical test is carried out on a semi-physical tester of the engine under development. The controller that has passed the hardware-in-the-loop verification is connected to the semi-physical test platform for simulation. In addition to the controller, the simulation platform also needs to include the actuators of the fuel tank system, including the fuel pump and fuel regulator.

[0091] The verification process involves simulating the engine's steady-state performance at the design point on a physical-in-the-loop simulation platform using an onboard adaptive engine model. The AI output from the model is then compared with the simulated test parameters. If the accuracy requirements are met, the evaluation method is considered valid. Otherwise, subsequent calculations should be repeated using the values in Table 2.

[0092] The above disclosure is only a specific embodiment of the present invention, but the present invention is not limited thereto. Any changes that can be conceived by those skilled in the art should fall within the scope of protection of the present invention.

Claims

1. A method for evaluating the performance of an advanced control plan for an aircraft engine, characterized in that: The following steps are involved: Obtain the execution requirements of the engine, obtain the influencing factors of the engine control plan performance based on the execution requirements, establish a hierarchical structure of the influencing factors, and construct a judgment matrix a ij ; Define the evaluation level interval and corresponding quantitative index of the performance of the advanced control plan. The evaluation level interval is divided according to the performance evaluation model of fuzzy theory. The impact factor is standardized and mapped to the evaluation level interval. The performance value W after the impact factor standard quantification is calculated. q ; For the judgment matrix a ij Perform consistency test. After completing the consistency test, perform hierarchical sorting based on the influence ratio of the influencing factors on the overall performance of the control plan, and determine the weight matrix k of the influencing factors. i ; According to the performance value W q The weight matrix k of the impact factor i , using the linear weighted average method to obtain the comprehensive evaluation value AI; where i represents the row of the matrix, j represents the column of the matrix, and q represents the impact factor; According to the comprehensive evaluation value AI, it is judged whether the engine control plan meets the requirements.

2. The method for evaluating the performance of an aircraft engine advanced control plan according to claim 1, wherein: The calculation formula of the comprehensive assessment value AI is: Where i∈[1,n], n represents the matrix order.

3. The method for evaluating the performance of an aircraft engine advanced control plan according to claim 2, wherein: The performance value W after the impact factor standard is quantified q The calculation formula is: q =W0+ΔW q , where W0 is the scoring base value, ΔW q It is the cumulative added points.

4. The method for evaluating the performance of an aircraft engine advanced control plan according to claim 3, wherein: The cumulative added value ΔW q The value standard is: when When ΔW q =2; when When ΔW q =1; when When ΔW q =0.5; when When ΔW q =0, Among them, A0 is the assessment index, A q For quantitative indicators.

5. The method for evaluating the performance of an advanced control plan for an aircraft engine according to claim 4, wherein: The quantitative indicator A q The determination method is as follows: the impact factor is standardized and mapped to the evaluation level interval. When the mapping result is close to the left value of the level interval, the lower limit value of the quantitative indicator corresponding to the interval is taken; when the mapping result is close to the right value of the level interval, the upper limit value of the quantitative indicator corresponding to the interval is taken.

6. The method for evaluating the performance of an advanced control plan for an aircraft engine according to claim 5, wherein: The judgment matrix a ij For pairwise comparison matrices, specifically: using a 1-9 scaling method to sequentially scale and compare the impact factors of each matrix.

7. The method for evaluating the performance of an aircraft engine advanced control plan according to claim 6, wherein: The calculation formula of the consistency ratio CR is: Among them, CI is the consistency index and RI is the random consistency index.

8. The method for evaluating the performance of an aircraft engine advanced control plan according to claim 7, wherein: The calculation formula of the consistency index CI is: The RI is the sample mean calculated by taking a sufficiently large sample of CI, where λ max is the judgment matrix a ij The maximum eigenvalue of .

9. The method for evaluating the performance of an aircraft engine advanced control plan according to claim 8, wherein: The maximum eigenvalue λ max It is achieved through the sum-product method. The specific steps are: First, the judgment matrix a ij Normalize each column element of Secondly, the normalized matrix a ij Sum by row; Then, the matrix a ij The matrix obtained by row summation is normalized to obtain the matrix M i , Then, the approximate eigenvector a is established ij M i ; Finally, calculate the maximum eigenvalue λ of the approximate eigenvector max , 10. The method for evaluating the performance of an aircraft engine advanced control plan according to claim 9, wherein: The weight matrix k of the impact factor i The calculation method is as follows: First, the judgment matrix a ij Normalize each column element of Secondly, the normalized matrix is averaged row by row; Finally, the normalized matrix is combined into a square matrix by row-averaged matrix, and normalized with each non-zero element forming a set to obtain the normalized impact factor weight matrix k i .

Images