Rotor blade sorting method based on single parent improved genetic algorithm
By optimizing blade sorting using a single-parent improved genetic algorithm, an objective function for the initial imbalance of the impeller is constructed. By combining gene recombination and adaptive mutation strategies, the problem of large initial imbalance after rotor impeller assembly is solved, and a smaller imbalance is achieved.
Patent Information
- Application Number
- CN202511168526.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-20
- Publication Date
- 2025-11-28
AI Technical Summary
The existing technology has failed to effectively address the following technical challenges in the rotor blade assembly process: it has failed to effectively reduce the initial imbalance of the rotor impeller, and the classic arrangement model has failed to consider the difference between the impeller's own imbalance and the blade's static torque, resulting in a large initial imbalance of the rotor after assembly.
A technical solution based on a single-parent improved genetic algorithm was adopted. By constructing the blade structure and the initial impeller structure, a separate technical solution was used to optimize the blade arrangement. The objective function for the initial impeller imbalance was constructed. Combined with gene recombination, gene translocation, and gene reversal operations, greedy selection and adaptive mutation strategies were added to optimize the blade arrangement.
The technical solution effectively reduced the rotor blades, decreased the initial imbalance of the impeller, and optimized the blade assembly effect.
Smart Images
Figure CN121031338A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application relates to the technical field of aero-engines and discloses a rotor blade sequencing method based on a single-parent improved genetic algorithm. BACKGROUND
[0002] In actual machining and production of rotor blades, geometric and mass differences exist among the blades due to machining tolerances and material properties, etc., so that an unbalance exists in the rotor impeller after assembly. In order to reduce the unbalance, the blade sequence is arranged according to a classical arrangement model of blade static moment in engineering, so that the static moment of the blade group after arrangement tends to zero. However, the classical arrangement model only arranges the blades according to the relative size of the blade static moment, without considering the differences of different batches of products and the unbalance of the wheel disc itself, so that the initial unbalance of the rotor is still large after the blade group after arrangement is assembled on the wheel disc. SUMMARY
[0003] The application aims to provide a rotor blade sequencing method based on a single-parent improved genetic algorithm, which optimizes blade arrangement and can effectively reduce the initial unbalance of the rotor impeller.
[0004] In order to achieve the above technical effects, the application adopts the following technical scheme: A rotor blade sequencing method based on a single-parent improved genetic algorithm, comprising the following steps: Step 1: constructing a target function about the initial unbalance of the impeller according to the unbalance of the wheel disc and the static moment value of each blade; Step 2: regarding blade groups with different arrangement sequences as species populations, regarding each blade group as a chromosome, and encoding the blade arrangement sequence; initializing the population by using the improved single-parent genetic algorithm, and generating a parent chromosome at random as an initial iteration point; Step 3: performing gene recombination operation on the parent chromosome to obtain a corresponding child chromosome, taking the unbalance of each chromosome as the fitness of the corresponding chromosome, and selecting the chromosome with the minimum fitness from the parent chromosome and the child chromosome by using a greedy selection algorithm; Step 4: judging whether the chromosome with the minimum fitness meets the end iteration condition, if yes, ending the iteration, taking the chromosome with the minimum fitness as the optimal chromosome, and outputting the corresponding blade arrangement sequence; otherwise, analyzing the mutation probability of the chromosome with the minimum fitness by using a mutation operator, and performing an adaptive mutation strategy to obtain a corresponding chromosome; Step 5: Between the chromosome with the lowest fitness before the mutation strategy is executed and the chromosome obtained after the mutation strategy is executed, a greedy selection algorithm is used to select the chromosome with the lowest fitness. If the chromosome with the lowest fitness satisfies the termination condition, the iteration ends and the chromosome with the lowest fitness is taken as the optimal chromosome, and its corresponding leaf arrangement order is output. Otherwise, the chromosome with the lowest fitness is taken as the next generation parent chromosome and steps 3-5 are repeated for iteration until the termination condition is met, and the leaf arrangement order of the optimal chromosome is output.
[0005] Furthermore, the objective function for the initial imbalance of the impeller is constructed as follows: ; in, This represents the initial imbalance of the impeller. The initial imbalance of the roulette wheel itself Axial components, The sum of the static moments of the blade assembly Axial components, The sum of the static moments of the blade assembly Axial components, The initial imbalance of the roulette wheel itself Axial components.
[0006] Furthermore, the gene recombination includes gene translocation, gene shifting, and gene reversal. The gene translocation includes single-point translocation and multi-point translocation, the gene shifting includes single-point shifting and multi-point shifting, and the gene reversal includes single-point inversion and multi-point inversion.
[0007] Furthermore, in step 2, parameters are set before initializing the population using the improved single-parent genetic algorithm. These parameters include population size, total number of generations, maximum mutation probability, minimum mutation probability, number of consecutive generations with optimal fitness, and allowable fitness.
[0008] Furthermore, the mutation operator is: ; in: This represents the probability of chromosome variation. The maximum mutation probability is preset; The preset minimum mutation probability; For chromosome fitness; The average fitness of a population is equal to the ratio of the sum of the fitness of all chromosomes in the population to the total number of chromosomes.
[0009] Furthermore, in step 4, if continuous The fitness of the chromosome with the lowest current fitness in a generation remains unchanged, and this applies to consecutive generations. If the number of generations is less than the total number of generations, then the fitness of the chromosome with the lowest current fitness is forced to be equal to the maximum mutation probability, the mutation probability of the chromosome with the lowest current fitness is increased, and then an adaptive mutation strategy is executed.
[0010] Furthermore, in step 4, the adaptive mutation strategy is as follows: a random number between 0 and 1 is randomly generated. If the random number is less than or equal to the mutation probability of the chromosome with the lowest current fitness, then the chromosome with the lowest current fitness is mutated to generate a new chromosome; otherwise, the chromosome with the lowest current fitness is not mutated.
[0011] Furthermore, the termination condition is: determining whether the number of iterations of the chromosome has reached the maximum number of iterations, or determining whether the fitness of the chromosome has reached the allowable fitness.
[0012] Furthermore, the static moment of the blade is measured by a static moment measuring device, which includes: The weighing beam is provided with a knife edge for supporting the weighing beam, and is also provided with a level. A balancing mechanism is provided on the weighing beam and is used to balance the weighing beam before measurement is performed; A weighing pan, which is installed at one end of the weighing beam, and contains standard weights; A connecting mechanism is installed at the other end of the weighing bar to fix the blade.
[0013] Compared with the prior art, the beneficial effects of this invention are: This invention employs a rotor blade sorting method based on an improved single-parent genetic algorithm. When constructing the objective function for the initial imbalance of the impeller, it considers the imbalance inherent in the impeller itself and the static torque of the blades. Furthermore, it utilizes single-chromosome gene translocation, gene shifting, and gene reversal operations, and adds greedy selection and adaptive mutation strategies to iteratively optimize the blade arrangement. Compared with the classic arrangement method, this invention effectively reduces the initial imbalance of the impeller. Attached Figure Description
[0014] Figure 1 This is a flowchart illustrating a rotor blade sorting method based on a single-parent improved genetic algorithm in an embodiment. Figure 2 This is a schematic diagram of the impeller and coordinate system in the embodiment; Figure 3 This is a schematic diagram of the static moment measuring device in the embodiment; Among them, 1-base, 2-weighing beam, 3-level, 4-sliding weight, 5-weighing pan, 6-standard weight, 7-connecting mechanism, 8-weighing blade, 9-equipped with smoothing rod. Detailed Implementation
[0015] The present invention will now be described in further detail with reference to the embodiments and accompanying drawings. However, this should not be construed as limiting the scope of the above-described subject matter of the present invention to the following embodiments; all technologies implemented based on the content of the present invention fall within the scope of the present invention.
[0016] Example See Figure 1 and Figure 2 A rotor blade sorting method based on a single-parent improved genetic algorithm, comprising: Step 1: Based on the imbalance of the impeller and the static torque value of each blade, construct an objective function for the initial imbalance of the impeller; Specifically, the impeller includes a disk and blades installed in the grooves of the disk. All blades on the impeller are defined as a blade group. Since the blade arrangement needs to consider both the magnitude and position of the blade static moment, i.e., the calculation process is a vector operation, an orthogonal decomposition calculation is performed by establishing a coordinate system, as shown in the figure. Figure 2 As shown, then: In the mortise At that place, the leaf static moment The axial components are: ; In the mortise At that place, the leaf static moment The axial components are: ; The sum of the static moments of the blade assembly The axial components are: ; The sum of the static moments of the blade assembly The axial components are: ; The sum of the static moments of the blade assembly is: ; Considering the impeller's own imbalance, the objective function for the initial impeller imbalance is constructed as follows: ; in, This represents the initial imbalance of the impeller. The initial imbalance of the roulette wheel itself Axial components, The sum of the static moments of the blade assembly Axial components, The sum of the static moments of the blade assembly Axial components, The initial imbalance of the roulette wheel itself Axial components.
[0017] Step 2: Treat leaf groups with different arrangement orders as species populations, where each leaf group is treated as a chromosome, and the leaf arrangement order is encoded; set the parameters of the improved single-parent genetic algorithm, including population size and total number of generations. Maximum mutation probability Minimum mutation probability Optimal fitness continuous algebra And allow fitness; initialize the population, and randomly generate a parent chromosome. , the parent chromosome Represented as: ,in, Indicates the first i One gene, Indicates the first n One gene.
[0018] Step 3: First, analyze the parent chromosomes... Gene recombination is performed to obtain corresponding offspring chromosomes. Specifically, the gene recombination includes gene translocation, gene shifting, and gene reordering.
[0019] Gene translocation includes single-point translocation and multi-point translocation. A single-point translocation involves two genes on the same chromosome exchanging positions, while a multi-point translocation involves multiple genes on the same chromosome exchanging positions. Both translocation methods can generate new chromosomes based on probability, and the parent chromosome... The offspring chromosomes were obtained through single-point transposition and multi-point transposition, respectively. and They are respectively: ; .
[0020] Gene translocation refers to the random selection of a daughter gene strand on a chromosome based on translocation probability, shifting that daughter gene strand one position backward, and placing the last gene in the first position of that daughter gene strand. Gene translocation includes single-point translocation and multi-point translocation, and the parent chromosome... After single-point translocation and multi-point translocation, offspring chromosomes were obtained respectively. and They are respectively: ; .
[0021] Gene inversion refers to reversing the sequence of a gene strand on a chromosome from beginning to end based on probability, with the inverted strand and its length being randomly selected. Gene inversion includes single-point inversion and multi-point inversion. (The parent chromosome...) The offspring chromosomes were obtained through single-point inversion and multi-point inversion, respectively. and They are respectively: ; .
[0022] Then, the imbalance value of each chromosome is calculated through the objective function, and the imbalance value of the chromosome is used as its corresponding fitness. The greedy selection algorithm is used to select the chromosome with the smallest fitness among the parent chromosome and all offspring chromosomes. Specifically, in the paternal chromosome Offspring chromosomes obtained through gene translocation and In this process, the first greedy selection operator is used to perform greedy selection, selecting the chromosome with the smallest fitness. ; in, This refers to the chromosome with the lowest fitness corresponding to the gene transposition operation. The paternal chromosome, These are the offspring chromosomes obtained through a single-point transposition operation. These are the offspring chromosomes obtained through multi-point transposition. For fitness functions. In the parent chromosome... Offspring chromosomes obtained through gene translocation and If the parent chromosome If the fitness is minimized, then the greedy selection operator selects the parent chromosome. As And output, if the offspring chromosome If the fitness is minimized, then the greedy selection operator selects the offspring chromosome. As And output, if the offspring chromosome If the fitness is minimized, then the greedy selection operator selects the offspring chromosome. As And output it. It should be noted that chromosomes... subscript It serves only as a symbol to distinguish it from other chromosomes and has no practical meaning; the subscript will appear later in the text. Similarly.
[0023] Similarly, in the paternal chromosome Offspring chromosomes obtained from gene translocation and In this process, the following second greedy selection operator is used to perform greedy selection, selecting the chromosome with the minimum fitness; ; in, This refers to the chromosome with the lowest fitness corresponding to the gene transfer operation. In the parent chromosome... Offspring chromosomes obtained from gene translocation and If the parent chromosome If the fitness is minimized, then the greedy selection operator selects the parent chromosome. As And output, if the offspring chromosome If the fitness is minimized, then the greedy selection operator selects the offspring chromosome. As And output, if the offspring chromosome If the fitness is minimized, then the greedy selection operator selects the offspring chromosome. As And output it.
[0024] In paternal chromosomes Offspring chromosomes obtained by gene reversal and The following third greedy selection operator is used to perform greedy selection, selecting the chromosome with the smallest fitness; ; in, This is the chromosome with the lowest fitness corresponding to the gene reordering operation. In the parent chromosome... Offspring chromosomes obtained by gene reversal and If the parent chromosome If the fitness is minimized, then the greedy selection operator selects the parent chromosome. As And output, if the offspring chromosome If the fitness is minimized, then the greedy selection operator selects the offspring chromosome. As And output, if the offspring chromosome If the fitness is minimized, then the greedy selection operator selects the offspring chromosome. As And output it.
[0025] This embodiment increases population diversity by adding a first mutation operator, a second mutation operator, and a third mutation operator.
[0026] Based on the three chromosomes selected above , , Then, a greedy selection algorithm is used to select the chromosome with the lowest fitness. Step 4: The iteration termination condition is determined by whether the current iteration count of the chromosome has reached the preset maximum iteration count, or whether the chromosome's fitness has reached the preset allowable fitness. First, it is determined whether the chromosome with the lowest fitness (i.e., the chromosome with the lowest fitness output in Step 3) satisfies the iteration termination condition. If it does, the iteration ends, and the chromosome with the lowest fitness is taken as the optimal chromosome, and its corresponding leaf arrangement order is output. Otherwise, the mutation operator is used to analyze the mutation probability of the chromosome with the lowest fitness, and an adaptive mutation strategy is executed to obtain the corresponding chromosome.
[0027] Specifically, assuming that the chromosome with the lowest fitness output in step 3 does not satisfy the termination iteration condition, the mutation operator for analyzing the mutation probability of the chromosome with the lowest fitness is: ; in: This represents the probability of chromosome variation. The maximum mutation probability is preset; The preset minimum mutation probability; For chromosome fitness; The average fitness of the population is equal to the ratio of the sum of the fitness of all chromosomes to the total number of chromosomes. When calculating the mutation probability of the chromosome with the lowest fitness, which is the final output in step 3, this is based on the fitness of that chromosome. The average fitness of the current population corresponding to this chromosome. The mutation probability of the chromosome is obtained through mutation operator analysis.
[0028] After obtaining the mutation probability of the chromosome with the lowest fitness as the final output of step 3, the adaptive mutation strategy is executed. The steps are as follows: randomly generate a random number. , If random number The mutation probability of the chromosome with the lowest fitness, which is less than or equal to the final output of step 3. Then the chromosome mutation generates a mutated chromosome. The chromosome is mutated and output as the result of the adaptive mutation strategy; otherwise, the chromosome does not mutate and is output as the result of the adaptive mutation strategy. It should be noted that the chromosome with the lowest fitness output in step 3 is mutated using the mutation operation of a genetic algorithm in the prior art to generate a mutated chromosome. .
[0029] It should be noted that if consecutive In step 3, the fitness of the chromosome with the lowest fitness remains unchanged, and this applies to consecutive generations. Less than the total number of generations If so, then the fitness of the chromosome with the lowest fitness in the final output of step 3 must be equal to the maximum mutation probability. This increases the mutation probability of the chromosome with the lowest fitness in the final output of step 3, and then an adaptive mutation strategy is executed.
[0030] Step 5: After executing the adaptive mutation strategy, between the chromosome with the lowest fitness before executing the mutation strategy and the chromosome obtained by executing the mutation strategy, a greedy selection algorithm is used to select the chromosome with the lowest fitness as the current chromosome with the lowest fitness. If the chromosome with the lowest fitness satisfies the termination iteration condition, the iteration ends and the chromosome with the lowest fitness is selected as the optimal chromosome, and its corresponding leaf arrangement order is output. Otherwise, the chromosome with the lowest fitness is selected as the next generation parent chromosome and steps 3-5 are repeated for iteration until the termination iteration condition is met, and the leaf arrangement order of the optimal chromosome is output.
[0031] This invention employs a rotor blade sorting method based on an improved single-parent genetic algorithm. When constructing the objective function for the initial imbalance of the impeller, it considers the imbalance inherent in the impeller itself and the static torque of the blades. Furthermore, it utilizes single-chromosome gene translocation, gene shifting, and gene reversal operations, and adds greedy selection and adaptive mutation strategies to iteratively optimize the blade arrangement. Compared with the classic arrangement method, this invention effectively reduces the initial imbalance of the impeller.
[0032] In this embodiment, the static moment of the blade is measured by a static moment measuring device, see [link to documentation]. Figure 3 The static torque measuring device includes: The weighing beam 2 is provided with a weighing knife edge 8 for supporting the weighing beam 2. The weighing knife edge 8 is installed on the base 1 and serves as the lever fulcrum of the weighing beam 2. The base 1 is a fixed point. The weighing beam 2 is provided with a scale. The weighing beam 2 is also provided with a level 3 for determining whether the weighing beam 2 has reached a balanced state. The weighing pan 5 is mounted on one end of the weighing rod 2 via a sliding sleeve. The sliding sleeve can slide along the weighing rod 2. When the sliding sleeve slides to the appropriate position, it is locked to the weighing rod 2 by a set screw or other locking structure. The weighing pan 5 is used to place standard weights 6. The connecting mechanism 7 and the weighing pan 5 are located at both ends of the weighing rod 2. The connecting mechanism 7 can be a connecting rod. The upper end of the connecting rod is installed on the weighing rod 2. The connecting rod can slide along the weighing rod 2. When the connecting rod slides to the appropriate position, it is locked to the weighing rod 2 by a set screw or other locking structure. The lower end of the connecting rod is provided with a clamp for fixing the blade. The clamp can be an existing blade clamping device with a tenon groove. The balancing mechanism includes a balancing rod 9 and a sliding weight 4 mounted on the balancing rod 9. The balancing rod 9 is fixed on the weighing beam 2. The balancing mechanism is used to balance the weighing beam 2 before measurement.
[0033] The method for measuring the static moment of a blade using a static moment measuring device is as follows: Step 1: Based on the installation dimensions of the blade being tested on the rotor disc, adjust and fix the position of the connecting mechanism 7, ensuring that the distance between the connecting mechanism 7 and the weighing blade 8 is [missing information]. It is equal to the installation dimension of the blade being tested on the rotor disc. It should be noted that the deviation between the distance from the connecting mechanism 7 to the weighing blade 8 and the installation dimension of the blade being tested on the rotor disc should be within the design range.
[0034] Step 2: Based on the installation dimensions of the blade being tested on the rotor disc, adjust and fix the position of the weighing pan 5, ensuring the distance between the weighing pan 5 and the weighing blade 8 is [specific dimensions]. It is equal to the installation dimensions of the blade being tested on the rotor disc.
[0035] Step 3: Adjust the sliding weight 4 to level the scale rod 2 before measurement until the level instrument 3 shows horizontal, and then fix the sliding weight 4.
[0036] Step 4: Fix the blade to be measured, and place the standard weight 6 on the weighing pan 5 until the level indicator 3 shows horizontal. The mass of the standard weight 6 is... Compared to the mass of the tested blade Slightly larger; after the level indicator 3 shows horizontal, the measuring device reaches a balanced state again.
[0037] Step 5: Through Calculate the static moment of the blade under test. And record it.
[0038] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A rotor blade sorting method based on a single-parent improved genetic algorithm, characterized in that, include: Step 1: Based on the imbalance of the impeller and the static torque value of each blade, construct an objective function for the initial imbalance of the impeller; Step 2: Treat leaf groups with different arrangement orders as species populations, with each leaf group as a chromosome, and encode the leaf arrangement order; The population is initialized using an improved single-parent genetic algorithm, with a randomly generated parent chromosome serving as the initial point for iteration. Step 3: Perform gene recombination on the parent chromosome to obtain the corresponding offspring chromosome. Use the imbalance of each chromosome as the fitness of the corresponding chromosome, and use a greedy selection algorithm to select the chromosome with the lowest fitness from the parent chromosome and offspring chromosome. Step 4: Determine whether the chromosome with the lowest fitness satisfies the termination condition. If it does, end the iteration and take the chromosome with the lowest fitness as the optimal chromosome, and output its corresponding leaf arrangement order; otherwise, use the mutation operator to analyze the mutation probability of the chromosome with the lowest fitness and execute the adaptive mutation strategy to obtain the corresponding chromosome. Step 5: Between the chromosome with the lowest fitness before the mutation strategy is executed and the chromosome obtained after the mutation strategy is executed, a greedy selection algorithm is used to select the chromosome with the lowest fitness. If the chromosome with the lowest fitness satisfies the termination condition, the iteration ends and the chromosome with the lowest fitness is taken as the optimal chromosome, and its corresponding leaf arrangement order is output. Otherwise, the chromosome with the lowest fitness is taken as the next generation parent chromosome and steps 3-5 are repeated for iteration until the termination condition is met, and the leaf arrangement order corresponding to the optimal chromosome is output.
2. The rotor blade sorting method according to claim 1, characterized in that, The objective function for the initial imbalance of the impeller is constructed as follows: ； in, This represents the initial imbalance of the impeller. The initial imbalance of the roulette wheel itself Axial components, The sum of the static moments of the blade assembly Axial components, The sum of the static moments of the blade assembly Axial components, The initial imbalance of the roulette wheel itself Axial components.
3. The rotor blade sorting method according to claim 1, characterized in that, The gene recombination includes gene translocation, gene shifting, and gene reversal. The gene translocation includes single-point translocation and multi-point translocation. The gene shifting includes single-point shifting and multi-point shifting. The gene reversal includes single-point inversion and multi-point inversion.
4. The rotor blade sorting method according to claim 1, characterized in that, In step 2, parameters are set before initializing the population using the improved single-parent genetic algorithm. These parameters include population size, total number of generations, maximum mutation probability, minimum mutation probability, number of consecutive generations with optimal fitness, and allowable fitness.
5. The rotor blade sorting method according to claim 4, characterized in that, The mutation operator is: ； in: This represents the probability of chromosome variation. The maximum mutation probability is preset; The preset minimum mutation probability; For chromosome fitness; The average fitness of a population is equal to the ratio of the sum of the fitness of all chromosomes in the population to the total number of chromosomes.
6. The rotor blade sorting method according to claim 5, characterized in that, In step 4, if continuous The fitness of the chromosome with the lowest current fitness in a generation remains unchanged, and this applies to consecutive generations. If the number of generations is less than the total number of generations, then the fitness of the chromosome with the lowest current fitness is forced to be equal to the maximum mutation probability, the mutation probability of the chromosome with the lowest current fitness is increased, and then an adaptive mutation strategy is executed.
7. The rotor blade sorting method according to claim 6, characterized in that, In step 4, the adaptive mutation strategy is as follows: a random number between 0 and 1 is randomly generated. If the random number is less than or equal to the mutation probability of the chromosome with the lowest current fitness, then the chromosome with the lowest current fitness is mutated to generate a new chromosome; otherwise, the chromosome with the lowest current fitness is not mutated.
8. The rotor blade sorting method according to claim 4, characterized in that, The termination condition is: determining whether the number of iterations of the chromosome has reached the maximum number of iterations, or determining whether the fitness of the chromosome has reached the allowable fitness.
9. The rotor blade sorting method according to claim 1, characterized in that, The static moment of the blade is measured by a static moment measuring device, which includes: The weighing beam (2) is provided with a knife edge (8) for supporting the weighing beam (2), and is also provided with a level (3); A balancing mechanism is provided on the weighing rod (2) and is used to balance the weighing rod (2) before measurement is carried out; A weighing pan (5) is installed at one end of the weighing rod (2), and a standard weight (6) is provided in the weighing pan (5); A connecting mechanism (7) is installed at the other end of the weighing bar (2) to fix the blade.