Methods, systems, storage media, and terminal equipment for assessing the performance degradation of aircraft engines
By employing multi-operating-point analysis and an improved differential evolution algorithm, the underdetermined problem and the local optimum trap problem in solving the degradation factor of aero-engines are solved, achieving high-precision engine performance degradation assessment, which is applicable to the health management and performance monitoring of aero-engines.
Patent Information
- Application Number
- CN202211028117.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-08-25
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2042-08-25
AI Technical Summary
In complex nonlinear systems, traditional genetic algorithms and differential evolution algorithms suffer from low computational accuracy, premature convergence, and susceptibility to local optima when solving engine degradation factors. Furthermore, the insufficient number of sensors makes it difficult to solve underdetermined problems.
The multi-operating-point analysis method is used to extend the equation system for solving the degradation factor. The fitness function is constructed by combining the engine performance calculation model. The algorithm's early-stage exploration capability and late-stage development capability are enhanced by improving the nonlinear population size reduction and the parameter adaptive differential evolution algorithm based on success history.
It improves the computational accuracy and robustness of engine performance degradation assessment, enabling accurate assessment of engine performance degradation under various operating conditions, and provides a breakthrough solution for engineering applications.
Smart Images

Figure CN115808875B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of aero-engine performance evaluation technology, specifically to aero-engine performance degradation evaluation methods, systems, storage media, and terminal equipment. Background Technology
[0002] As the primary power source for aircraft, the reliability of aero-engines is crucial for flight safety. Operating under harsh environments such as high temperature, high pressure, and high speed, engines inevitably experience tip wear and blade corrosion in their airflow components, leading to performance degradation and frequent malfunctions such as overheating, stall, and surge. Therefore, to improve engine reliability, it is necessary to conduct performance degradation assessment studies on aero-engines.
[0003] The performance degradation assessment of aero-engines is mainly achieved by estimating engine degradation factors (the efficiency and flow capacity of various components). Currently, engine degradation factor estimation methods are mainly divided into physical model-based methods and data-driven methods. Data-driven methods train a large number of "measurement parameter-degradation factor" mapping samples, enabling the model to calculate the engine performance degradation trend from the measurement parameters. This method has shown excellent predictive ability on simulated degradation datasets. However, the performance degradation factors of actual engines cannot be measured, making data-driven methods difficult to apply to actual engine performance degradation assessment. Physical model-based methods, on the other hand, do not require prior knowledge of engine degradation and can estimate engine degradation factors based on the deviation between the engine performance calculation model and engine sensor data. However, this estimation requires that the number of measurement parameters p be greater than or equal to the number of component correction coefficients n. Due to factors such as installation space and weight, the number of sensors in actual engines is limited, making it difficult to meet the requirement that the number of measurement parameters be greater than or equal to the number of component correction coefficients, thus making the solution an underdetermined problem, theoretically unsolvable. To address the aforementioned underdetermined problem, a multi-operating-point analysis method can be applied. By introducing multiple operating points, the equations for solving the degradation factor can be extended, thereby resolving the underdetermined problem in solving the nonlinear equations.
[0004] After solving the underdetermined problem of the degradation factor, another challenge in solving it is that traditional optimization algorithms are prone to getting trapped in local optima, especially when dealing with complex multi-peak problems such as nonlinear systems of aero-engines.
[0005] The difficulty in solving the aforementioned technical problems lies in the fact that, since aero-engines are complex nonlinear systems, existing methods in this field suffer from low computational accuracy and premature convergence when solving for engine degradation factors. While differential evolution algorithms have advantages such as simple structure, few control parameters, and high computational accuracy, their linear population reduction strategy can affect population diversity during the initial rapid population reduction, easily leading to local optima. This results in poor performance when dealing with the multi-peak problem of solving aero-engine degradation factors, and the algorithm also lacks robustness. Summary of the Invention
[0006] To address the technical problems existing in the prior art, this invention provides a method, system, storage medium, and terminal device for evaluating the performance degradation of aero-engines. First, it utilizes a multi-operating-point analysis method, introducing multiple operating points to expand the degradation factor solution equations, thus solving the problem of insufficient sensor count in flight parameter data. Second, it constructs a fitness function based on the engine performance calculation model and sensor data, transforming the degradation estimation problem into an optimization problem. Third, based on standard linear population size reduction and differential evolution algorithms with parameter adaptation based on success history, it proposes a nonlinear population reduction strategy and a weighted mutation strategy, thereby enhancing the algorithm's exploration capability in the early stages of iteration and its development capability in the later stages. Finally, through experiments comparing with other differential evolution algorithms, the convergence accuracy and robustness of the invented method are verified.
[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for evaluating the performance degradation of an aero-engine, comprising the following steps:
[0008] Step 1: Establish a nonlinear engine model that considers component performance degradation, and construct an estimation expression for the engine component degradation factor;
[0009] Step 2: Use the multi-operating-point analysis method to solve the equation system by introducing multiple operating points to expand the degradation factor.
[0010] Step 3: Construct a fitness function using the differences between engine measurement parameters and engine model;
[0011] Step 4: Apply the improved nonlinear population size reduction and the adaptive differential evolution algorithm based on successful history parameters to solve the degradation factor in reverse, thereby realizing the assessment of engine performance degradation.
[0012] Preferably, the nonlinear engine model established in Step 1 is an expression for the engine measurement parameters:
[0013] z(t)=g(x(t),u(t),θ)
[0014] In the formula, x is the engine state vector, u is the engine input vector, z is the engine measurement parameter vector, θ is the vector composed of the performance degradation factors of each rotating part of the engine, and t is time;
[0015] Based on the expression for engine measurement parameters, the estimation formula for engine component degradation factor θ=G is constructed by reverse solving. -1 (x(t), u(t), z(t)).
[0016] Preferably, Step 2 specifically includes the following steps:
[0017] Step 21: Find the engine steady-state operating point within the speed range of 86%N2-92%N2, and select the point with the largest operating difference within this speed range in a single flight as the multiple operating point. N2 is the engine high-pressure rotor speed.
[0018] Step 22: After selecting a suitable multi-working point, within a single region, apply the equation θ = G -1 Extending (x(t), u(t), z(t)) yields θ = G -1 (x k (t), u k (t), z k (t)), k = 1, 2, ..., q;
[0019] In the formula, q is the number of selected working points, and p×q≥n is satisfied, p is the number of sensor measurement parameters, and n is the number of engine health factors.
[0020] Preferably, the fitness function constructed in Step 3 is:
[0021]
[0022] In the formula, This represents the j-th actual measured parameter value at the i-th working point; Calculate parameter values for the j-th model at the i-th operating point; p is the number of sensor measurement parameters; q is the number of selected operating points; ω j This refers to the weights of the measurement parameters in the fitness function.
[0023] Preferably, Step 4 specifically includes the following steps:
[0024] Step 41, Initialization: Randomly generate an initial population of candidate solutions within the constraints of the population vector, where the j-th dimension component x of the i-th population vector is... i,j for:
[0025] x i,j =x min,j +rand[0,1]·(x max,j -xmin,j )
[0026] In the formula, rand[0,1] is a random number uniformly distributed between 0 and 1; the subscripts max and min are the upper and lower limits of the population range; i = 1, 2, ..., N, where N is the size of the population; j = 1, 2, ..., d, where d is the dimension of the population vector;
[0027] Step 42, Mutation: For the i-th parent vector x of the G-th generation population... i,G Perform the mutation operation to generate the mutation vector v. i,G The specific mutation strategy is as follows:
[0028] v i,G =x i,G +F w ·(x pbest,G -x i,G )+F i ·(x r1,G -x r2,G )
[0029] F w = (0.7 + 0.05 · round[10 · n) fesr ])·F
[0030] In the formula, x pbeSt Randomly select from the elite individuals of the first p·N generations (p∈[0,1], product rounded to the nearest integer) of the Gth generation; while x r1,G ∈P G Randomly select from the N population vector set PG of the Gth generation; x r2,G ∈P G ∪A, where A is an external storage unit. Before storage, individuals that have failed are selected in the first G-1 generations. When the population size exceeds N, an old individual from A is randomly selected to replace the new individual; F w The weighted scaling factor is ; F is the scaling factor; round[] is the rounding function; n fesr =N fes / N fesmax , which is the ratio of the current number of evaluations to the maximum number of evaluations; N fes N represents the current number of evaluations. ifesmax This represents the maximum number of evaluations.
[0031] Step 43, Crossover: Experimental vector individual u i,j,G Individual v with mutation vector i,j,G and parent vector individual x i,j,G Cross generation:
[0032]
[0033] In the formula, CR is the crossover rate; j rand A random integer in [1:D];
[0034] Step 44: Selection: Calculate the fitness function values of the experimental vector and the parent vector respectively. Following the principle of survival of the fittest, determine the next generation of individuals, retaining the better-performing vectors for the next generation. Mathematically, this is expressed as:
[0035]
[0036] In the formula, u i,G Let x be the test vector. i,G The parent vector;
[0037] Step 45: The accuracy of the differential evolution algorithm is controlled by adaptive parameter adjustment. The population reduction algorithm used in the adaptive adjustment is as follows:
[0038]
[0039] In the formula, N G+1 The population size for the next generation; round[] is the floor function; N min =4 represents the final population size after the evolutionary process; N init nfes is the initial population size. r =N fes / N fesmax N is the ratio of the current number of assessments to the maximum number of assessments. fes N represents the current number of evaluations. fesmax This represents the maximum number of evaluations.
[0040] Preferably, in the mutation operation of Step 42, if the individual numerical vector value exceeds the limit range, the value is corrected by the following formula:
[0041]
[0042] In the formula, v i,j,G For the individual with the mutation vector, x i,j,G For the parent vector individual, x max,j To optimize the upper bound of the problem, x min,j To optimize the lower bound of the problem.
[0043] Another object of the present invention is to provide an evaluation system for implementing the aforementioned method for reducing the performance degradation of aero-engines, the evaluation system comprising:
[0044] The engine nonlinear model module is used to construct estimation expressions for the degradation factors of engine components;
[0045] The degeneration factor solution equation system multi-operating-point extension module is used to extend the degeneration factor solution equation system;
[0046] The fitness function construction module is connected to the multi-operating point extension module for solving the degradation factor equation system, and is used to transform the equation system for solving the degradation factor into an optimization problem;
[0047] The model solving module, connected to the engine nonlinear model module and the fitness function construction module, uses an improved nonlinear population size reduction and an adaptive differential evolution algorithm based on successful history parameters to solve for the degradation factor in reverse.
[0048] Another object of the present invention is to provide a computer-readable storage medium storing a computer program that is loaded by a processor and executed by the aforementioned method for assessing the performance degradation of aero-engines.
[0049] Another objective of this invention is to provide a terminal device comprising a memory and a processor. The memory stores a computer program, which is loaded and executed by the processor to evaluate the aforementioned aircraft engine performance degradation method. The processor is used to implement the various computer programs within the computer program.
[0050] Compared with the prior art, the present invention has the following beneficial effects:
[0051] This invention utilizes a multi-operating-point analysis method, introducing multiple operating points to expand the equations for solving the degradation factor, thus addressing the problem of insufficient sensor count in flight parameter data. Based on the engine performance calculation model and sensor data, a fitness function is constructed, transforming the degradation estimation problem into an optimization problem. Building upon standard linear population reduction and differential evolution algorithms based on success history and parameter adaptation, a nonlinear population reduction strategy and a weighted mutation strategy are proposed. The former maintains a relatively constant population size in the early stages of iteration, providing sufficient evolutionary time for the initial generation and enhancing the algorithm's diversity. The latter improves both the exploration capability in the early stages of iteration and the development capability in the later stages. This invention provides a groundbreaking solution for evaluating the performance degradation of aero-engines under various operating environments and timelines, possessing considerable engineering application potential. Attached Figure Description
[0052] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention, but do not constitute a limitation thereof. In the drawings:
[0053] Figure 1 This is a flowchart illustrating the implementation of the aero-engine performance degradation assessment method provided by the present invention.
[0054] Figure 2This is a comparison chart of population size changes during the optimization and iteration process between the population reduction strategy provided by this invention and existing strategies.
[0055] Figure 3 This is a flowchart of the improved nonlinear population size reduction and adaptive differential evolution algorithm based on success history parameters proposed in this invention.
[0056] Figure 4 This is a comparison graph of the fitness function of the optimization algorithm provided by this invention with other optimization algorithms during the iteration process, wherein, Figure 4 (a) is a comparative diagram of Example 1; Figure 4 (b) is a comparative diagram of example 2. Detailed Implementation
[0057] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.
[0058] This embodiment proposes a method for evaluating the performance degradation of aero-engines, and the method steps are as follows:
[0059] S1: Establish a nonlinear engine model that considers component performance degradation, and construct an estimation expression for the engine component degradation factor.
[0060] Specifically, considering the performance degradation of engine components, a nonlinear model of the engine is established, and the expression for the engine measurement parameters is obtained as z(t)=g(x(t),u(t),θ), where x is the engine state vector, u is the engine input vector, z is the engine measurement parameter vector, θ is the vector composed of the performance degradation factors (flow capacity SW and efficiency characteristics SE) of each rotating part of the engine, and t is time.
[0061] Considering that the degradation factor θ of the component cannot be directly measured or calculated directly from the sensor measurement parameters, the component degradation factor is estimated as θ = G by solving the equation z(t) = g(x(t), u(t), θ) in reverse from the measurement parameters. -1 (x(t), u(t), z(t)).
[0062] S2: To address the problem of insufficient measurement parameters in traditional degradation factor estimation methods, a multi-operating-point analysis method is used to expand the degradation factor solution equations by introducing multiple operating points, thereby solving the underdetermined problem in solving nonlinear equations.
[0063] When implementing the above scheme, the estimation of the degradation factor requires that the number of measured parameters p be greater than or equal to the number of component correction coefficients n. Due to factors such as installation space and weight, the actual number of engine sensors is limited, making it difficult to meet the requirement that the number of measured parameters be greater than or equal to the number of component correction coefficients. This makes the above equation an underdetermined problem, theoretically unsolvable. Through analysis of actual flight parameter data from multiple sorties of the engine under study, it was found that the steady-state operating point of this type of engine is mainly concentrated in a small speed range (86% N2-92% N2). Since this speed range is small, it is approximately assumed that the component performance degradation factor is equal within this range. At the same time, in the selection of multiple operating points, points with large operating differences within this speed range in a single flight are selected as much as possible to alleviate the multicollinearity problem in solving the equation system, thereby obtaining better equation well-posedness. Therefore, within a single region, the equation θ = G -1 Extending (x(t), u(t), z(t)) yields:
[0064] θ=G -1 (x k (t), u k (t), k (t)), k = 1, 2, ..., q
[0065] In the formula, q represents the number of selected work points, and p×q≥n is satisfied.
[0066] In this example, the engine under study has five air path sensors: fuel flow rate Wf, high-pressure rotor speed N2, low-pressure rotor speed N1, total pressure after high-pressure compressor Pt3, and total temperature after low-pressure turbine Tt5. Wf is the input to the engine model, and the model outputs four variables p. The degradation factors of the study object are the efficiency and flow capacity of the engine's rotating components (fan, low-pressure compressor, high-pressure compressor, high-pressure turbine, and low-pressure turbine), denoted as SE12, SW12, SE2, SW2, SE26, SW26, SE41, SW41, SE46, and SW46, respectively, with a quantity n of 10. q is chosen as 3.
[0067] S3: Construct a fitness function using the differences between engine measurement parameters and the engine model. To solve for the degradation factor, the equation system θ=G -1 (x k (t), u k (t), z kSolving the problem for (t) with k = 1, 2, ..., q is transformed into an optimization problem, and the key to transforming it into an optimization problem is to determine the fitness function.
[0068] The weights ω of the measurement parameters in the fitness function are determined by comprehensively considering the error range of the measurement parameters, the relationship between the measurement parameters, and the ratio of modeling error to measurement parameter value. j The constructed fitness function is:
[0069]
[0070] In the formula For the j-th actual measured parameter value at the i-th working point, Calculate the parameter values for the j-th model at the i-th operating point. The degradation factor can be estimated by minimizing this fitness function.
[0071] S4: An improved nonlinear population size reduction and an adaptive differential evolution algorithm based on success history parameters are used to inversely solve for the degradation factor, thereby enabling the assessment of engine performance degradation.
[0072] When solving for the degradation factor using the algorithm proposed above, initialization is performed first. That is, an initial population of candidate solutions is randomly generated within the constraints of the population vector, where the j-th dimension component x of the i-th population vector is... i,j for:
[0073] x i,j =x min,j +rand[0,1]·(x max,j -x min,j )
[0074] In the formula: rand[0,1] is a random number uniformly distributed between 0 and 1; the subscripts max and min are the upper and lower limits of the population range; i = 1, 2, ..., N, N is the size of the population; j = 1, 2, ..., d, d is the dimension of the population vector.
[0075] In the differential evolution algorithm, the parent vector x of the i-th generation of the population in the G-th generation is... i,G It will perform a mutation operation, producing a mutation vector v. i,G To address the performance degradation optimization problem of aero-engines, this application proposes a new mutation strategy:
[0076] v i,G =x i,G +F w ·(x pbest,G -x i,G )+F i ·(x r1,G -x r2,G )
[0077] F w = (0.7 + 0.05 · round[10 · n) fesr ])·F
[0078] In the formula: x pbest Randomly select from the elite individuals of the first p·N generations (p∈[0,1], product rounded to the nearest integer) of the Gth generation; x r1,G ∈P G From the Nth population vector set P of the Gth generation G Random selection from; x r2,G ∈P G ∪A, where A is an external storage unit. During the first G-1 generations, failed individuals are selected. When the population size exceeds N, an older individual from A is randomly selected to replace the new individual. The introduction of storage unit A increases population diversity; F w The weighted scaling factor is F; the scaling factor is round[]; the round[] function is the rounding function; nfes r =N fes / N fesmax N is the ratio of the current number of assessments to the maximum number of assessments. fes N represents the current number of evaluations. fesmax This represents the maximum number of evaluations.
[0079] Mutation operations may cause a value in an individual's numerical vector to exceed a certain limit. The value is corrected using the following formula:
[0080]
[0081] In the formula: rand[0,1] is a random number uniformly distributed between 0 and 1; the subscripts max and min are the upper and lower limits of the population range; i = 1, 2, ..., N, N is the size of the population; j = 1, 2, ..., d, d is the dimension of the population vector.
[0082] Test vector individual u i,j,G Individual v with mutation vector i,j,G and parent vector individual x i,j,G Cross generation:
[0083]
[0084] In the formula: CR is the crossover rate, j rand For a random integer in [1:D]. By introducing j rand This ensures that at least one dimension of each parent vector undergoes a crossover operation, thereby improving the convergence speed of the algorithm.
[0085] Calculate the fitness function values of the experimental vector and the parent vector respectively, and determine the next generation of individuals according to the principle of survival of the fittest. Vectors with better performance are retained in the next generation. Mathematically, this can be expressed as:
[0086]
[0087] In the formula: u i,G Let x be the test vector. i,G The parent vector;
[0088] The control parameters determine the accuracy of the differential evolution algorithm. However, manually setting these parameters cannot satisfy all optimization problems; therefore, adaptive adjustment of the control parameters is necessary. In adaptive adjustment, a larger population size provides better population diversity and exploration capabilities, while a smaller population size increases the utilization of promising areas, resulting in better development capabilities. Therefore, a technique called population reduction is used for adaptive adjustment during parameter iteration. However, existing population reduction strategies suffer from low algorithmic efficiency and poor convergence.
[0089] To address the shortcomings of existing population reduction technologies, this application proposes a new population reduction algorithm:
[0090]
[0091] Where: N G+1 The population size for the next generation; round[] is the floor function; N min =4 represents the final population size in the evolutionary process, N init n is the initial population size; fesr =N fes / N fesmax , where N is the ratio of the current number of evaluations to the maximum number of evaluations. fes N represents the current number of evaluations. fesmax This represents the maximum number of evaluations.
[0092] This algorithm, while maintaining a relatively stable population size in the early stages, allows the first generation of individuals ample time to evolve, greatly enhancing the algorithm's diversity.
[0093] Table 1 shows the estimation and evaluation results of degradation factors using JADE, L-SHADE, NL-SHADE-RSP and the method of this application on the two groups of pre-implanted performance degradation data, where the degradation factor search range of each algorithm is [0.95, 1.05].
[0094] Table 1. Engine performance degradation implantation and estimation results of various algorithms.
[0095]
[0096]
[0097] As can be seen from the comparison results in Table 1, the calculation results of the present invention are the best, with the smallest average error, which can meet the engineering accuracy requirements for estimating engine degradation factors.
[0098] Figure 4 A comparison diagram of the fitness function during the iteration process of the optimization algorithm provided by this invention and other optimization algorithms.
[0099] Depend on Figure 4 As can be seen from the comparison of the iteration processes, JADE, due to the lack of a historical memory, has control parameters that are affected by previous success values during the iteration process, resulting in poor convergence accuracy. L-SHADE and NL-SHADE-RSP, influenced by their population reduction strategies, have the fastest iteration convergence speed, but are prone to getting stuck in local optima in the later stages of iteration. The method of this invention enhances the population diversity in the early stages of algorithm iteration and the development capability in the later stages of algorithm iteration, and has higher computational accuracy than the other three algorithms.
[0100] This invention is applicable to research on engine health management and performance monitoring. Given known flight data and the establishment of an engine nonlinear energy model, it can provide sufficiently realistic and reliable degradation factor assessment results; and it has significant reference value for the performance evaluation of other equipment such as gas turbines.
[0101] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the scope and spirit of the invention, the scope of which is defined by the claims and their equivalents.
Claims
1. A method for assessing the performance degradation of aircraft engines, characterized in that, Includes the following steps: Step 1: Establish a nonlinear engine model that considers component performance degradation, and construct an estimation expression for the engine component degradation factor; Step 2: Use the multi-operating-point analysis method to solve the equation system by introducing multiple operating points to expand the degradation factor. Step 3: Construct a fitness function using the differences between engine measurement parameters and engine model; Step 4: Apply the improved nonlinear population size reduction and the adaptive differential evolution algorithm based on success history parameters to solve the degradation factor in reverse, so as to realize the assessment of engine performance degradation. Step 4 details Includes the following steps: Step 41, Initialization: Randomly generate an initial population of candidate solutions within the constraints of the population vector, where the j-th dimension component x of the i-th population vector is... i,j for: ; In the formula, rand[0,1] is a random number uniformly distributed between 0 and 1; the subscripts max and min are the upper and lower limits of the population range; i=1,2,…,N, where N is the size of the population; j=1,2,…,d, where d is the dimension of the population vector; Step 42, Mutation: Targeting the... G The generation i Parent vectors of each population x i,G Perform the mutation operation to generate a mutation vector. v i,G The specific mutation strategy is as follows: ; ; In the formula, x pbest,G From the G Generations ago Randomly selected from the elite individuals of the generation, among whom ;and From the G Given a set of N population vectors P G Random selection from; Where A is an external storage unit, before storage G - Select the unsuccessful individuals in the first generation, when the population size exceeds N When the time is right, randomly select an old individual from A and replace it with a new individual; F w This is the weighted scaling factor; F is the scaling factor; round[] is the rounding function; = N fes / N fesmax , which is the ratio of the current number of assessments to the maximum number of assessments; N fes This represents the current number of assessments. N fesmax This represents the maximum number of evaluations. Step 43, Crossover: Experimental Vector Individuals u i,j,G Individuals with mutation vectors v i,j,G and parent vector individual x i,j,G Cross generation: ; In the formula, CR Crossover rate; j rand A random integer in [1:D]; Step 44: Selection: Calculate the fitness function values of the experimental vector and the parent vector respectively. Following the principle of survival of the fittest, determine the next generation of individuals, retaining the better-performing vectors for the next generation. Mathematically, this is expressed as: ; In the formula, For the test vector, The parent vector; Step 45: The accuracy of the differential evolution algorithm is controlled by adaptive parameter adjustment. The population reduction algorithm used in the adaptive adjustment is as follows: ; In the formula, N G+1 The population size for the next generation; round[] is the floor function; N min =4 indicates the final population size after the evolutionary process; N init This is the initial population size; nfes r = N fes / N fesmax This is the ratio of the current number of assessments to the maximum number of assessments. N fes This represents the current number of assessments. N fesmax This represents the maximum number of evaluations.
2. The method for assessing the performance degradation of an aero-engine according to claim 1, characterized in that, The nonlinear engine model established in Step 1 is the expression for the engine measurement parameters: ; In the formula, x This is the engine state vector. u It is the engine input vector. It is an engine measurement parameter vector. is a vector composed of performance degradation factors of each rotating component of the engine, where t is time; Based on the expression of engine measurement parameters, an estimation formula for engine component degradation factors is constructed by inversely solving the expression. .
3. The method for assessing the performance degradation of an aero-engine according to claim 2, characterized in that, Step2 specific Includes the following steps: Step 21: Find the engine steady-state operating point within the speed range of 86%N2-92%N2, and select the point with the largest operating difference within this speed range in a single flight as the multiple operating point. N2 is the engine high-pressure rotor speed. Step 22: After selecting a suitable multi-working point, within a single area, adjust the formula... To expand upon this, we can obtain... ; In the formula, q The number of selected work points, and satisfying the following conditions: , p The number of parameters measured by the sensor. n This refers to the number of engine health factors.
4. The method for assessing the performance degradation of an aero-engine according to claim 3, characterized in that, The fitness function constructed in Step 3 is: ; In the formula, For the first i The first work site j One actual measured parameter value; For the first i The first work site j Each model calculates parameter values; p The number of parameters measured by the sensor; q The number of work points selected; This refers to the weights of the measurement parameters in the fitness function.
5. The method for assessing the performance degradation of an aero-engine according to claim 1, characterized in that, In Step 42, during the mutation operation, if the individual numerical vector value exceeds the limit range, the value is corrected using the following formula: ; In the formula, For individuals with mutated vectors, For the parent vector individual, To optimize the upper limit of the problem, To optimize the lower bound of the problem.
6. An evaluation system for implementing the aero-engine performance degradation evaluation method as described in any one of claims 1 to 5, characterized in that, The evaluation system includes: The engine nonlinear model module is used to construct estimation expressions for the degradation factors of engine components; The degeneration factor solution equation system multi-operating-point extension module is used to extend the degeneration factor solution equation system; The fitness function construction module is connected to the multi-operating point extension module for solving the degradation factor equation system, and is used to transform the equation system for solving the degradation factor into an optimization problem; The model solving module, connected to the engine nonlinear model module and the fitness function construction module, uses an improved nonlinear population size reduction and an adaptive differential evolution algorithm based on successful history parameters to solve for the degradation factor in reverse.
7. A computer-readable storage medium, characterized in that, The readable storage medium stores a computer program, which is loaded by a processor and executed as the assessment method for the performance degradation of an aero-engine according to any one of claims 1-5.
8. A terminal device, characterized in that, It includes a memory and a processor. The memory stores a computer program, which is loaded by the processor and executed as described in any one of claims 1 to 5, to assess the performance degradation of an aero-engine. The processor is used to implement the various computer programs within the computer program.
Citation Information
Patent Citations
Large-bypass-ratio civil aviation engine performance diagnosis method and system
CN107977526A
Aero-engine gas path fault diagnosis method based on DBN-DTL
CN114136622A
