Improved genetic algorithm-based fan lattice type tower optimization design system and method

Abstract

The invention discloses a fan lattice tower optimization design system and method based on an improved genetic algorithm, and the system comprises a parameter input module which is used for receiving and preprocessing tower design parameters; the population initialization module is used for randomly generating a population and adapting to the change of the overall structure layer number of the tower by adopting a variable-length coding strategy; the finite element modeling and analysis module is used for constructing a parameterized finite element model for individuals of each population based on a finite element platform and performing finite element analysis; the algorithm optimization module is used for calculating the total cost and penalty term of each individual, and then taking the total cost of the tower as an optimization target, and using an improved genetic algorithm to complete optimization iteration; and the result output module is used for outputting an optimal individual parameter and a total cost optimal value when optimization iteration reaches the maximum evolutionary algebra or meets a convergence condition. The manufacturing cost of the tower can be remarkably reduced on the premise that constraint conditions such as strength, stability and frequency are met.

Description

Technical Field

[0001] This invention relates to the technical field of wind power generation, and in particular to a wind turbine lattice tower optimization design system and method based on an improved genetic algorithm. Background Technology

[0002] As wind turbine units become larger, the load on the turbine head is increasing, posing significant challenges to the design, manufacturing, and transportation of traditional conical steel towers. While steel-concrete towers offer advantages in load-bearing capacity, they still suffer from difficulties in quality control, long construction periods, and inconvenient dismantling. Against this backdrop, lattice towers, with their high stiffness, ease of on-site assembly, and flexible component transportation, are gradually emerging as a promising alternative. However, these towers are structurally complex, comprising main members, diagonal braces, and steel strands, with design variables encompassing both geometric and shape parameters, exhibiting high dimensionality and strong coupling. Furthermore, the design must meet multiple regulatory constraints, including strength, stability, natural frequency, and structural requirements, further increasing the optimization difficulty. Currently, the design of lattice towers largely relies on empirical trial-and-error algorithms or traditional optimization tools. The former suffers from low automation, low efficiency, and dependence on human experience, making global optimization difficult; the latter is prone to slow convergence or premature convergence under multi-variable and multi-constraint conditions, failing to meet the dual requirements of efficiency and accuracy in engineering practice. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of existing technologies and provide an optimized design system for wind turbine lattice towers based on an improved genetic algorithm. By constructing a parameterized finite element model and using the total tower cost as the optimization objective, the system automatically optimizes structural parameters while meeting various specification constraints. Furthermore, considering the unique characteristics of lattice tower optimization problems, several improvements are introduced on top of the standard genetic algorithm: a variable-length encoding mechanism adapts to changes in the number of tower layers; embedded specification checks ensure solution feasibility; a dynamic adaptive mutation strategy improves convergence performance; and elite retention, restart mechanisms, and population diversity control balance global and local searches, improving the algorithm's search efficiency and optimization accuracy. This significantly reduces tower manufacturing costs while meeting constraints such as strength, stability, and frequency.

[0004] Another objective of this invention is to provide an optimization design method for wind turbine lattice towers based on an improved genetic algorithm.

[0005] The objective of this invention is achieved through the following technical solution: A wind turbine lattice tower optimization design system based on an improved genetic algorithm includes, The parameter input module is used to receive tower design parameters and perform preprocessing. The population initialization module is used to randomly generate a population based on the preprocessed tower design parameters as constraints, and to adopt a variable-length coding strategy to adapt to changes in the number of layers in the overall tower structure. The finite element modeling and analysis module is used to construct parametric finite element models for each individual in the population and perform finite element analysis based on the finite element platform; wherein, the parametric finite element model includes main members, diagonal bars, steel strands, transition sections and towers; The algorithm optimization module is used to calculate the total cost and penalty term of each individual based on the finite element analysis results of each individual. Then, with the total cost of the tower as the optimization objective, the improved genetic algorithm, which introduces dynamic adaptive mutation strategy, elite retention strategy, restart mechanism strategy and population diversity strategy, completes the optimization iteration. The results output module is used to output the optimal individual parameters and the optimal total cost when the optimization iteration reaches the maximum number of generations or meets the convergence condition.

[0006] Furthermore, the tower design parameters include multi-condition ultimate load data, tower foundation geometric parameters, material property parameters, and specification constraint parameters.

[0007] Furthermore, the population initialization module specifically performs the following operations: Based on the preprocessed tower design parameters as constraints, a random population is generated, with each individual corresponding to a set of tower design structural parameters. A variable-length encoding strategy is adopted to adapt to the non-fixed number of design variables. A standard constraint check algorithm is embedded in the population initialization process to ensure that all initial individuals meet the construction measures requirements and to avoid invalid individuals from entering the subsequent optimization process.

[0008] Furthermore, the tower design structural parameters include: the angle α between the main member and the horizontal ground, the angle β between the diagonal member and the main member, the number of steel strands N, and the diameter D and thickness T of each main member, and the diameter d and thickness t of each diagonal member.

[0009] Furthermore, the main limbs are simulated using elastic beam-column elements. Based on the equivalent combined stiffness formula, the material parameters of the steel pipe and concrete of the tower are input to simulate the mechanical behavior of the steel pipe and concrete working together. The diagonal members are simulated using elastic beam-column elements, with only the steel material parameters input. The steel strands are simulated using truss elements, with the initial strain set by the initial strain material to simulate the prestress effect, bearing only axial force. The transition section and tower are simulated using ASD four-node shell elements, and the connection between the transition section and the main limbs is achieved through equal degree of freedom constraint commands. To transfer the load from the upper part of the wind turbine system, a central node is set at the top of the tower, and a rigid connection command is used to establish a rigid connection between the central node and the top node of the tower. The nacelle, hub, and upper mass of the wind turbine are simulated as concentrated mass points through mass commands and rigidly connected to the central node through rigid connections. The bottom node of the tower is set as a consolidation constraint using consolidation commands. By simultaneously performing finite element analysis on multiple individuals in the population, the first natural frequency of the tower, the strength stress ratio of the main limbs and diagonal members, and the stability stress ratio are output.

[0010] Furthermore, the algorithm optimization module specifically performs the following operations: Fitness assessment: The optimization objective is the total cost of the tower, which includes the cost of concrete, steel and steel strands, and a penalty term is introduced. The penalty term is used to constrain the aspect ratio of the main members, the hoop coefficient, the solidity ratio and the steel content of the section, the slenderness ratio of the diagonal members, the first natural frequency of the tower, the strength stress ratio of the main members and the diagonal members, and the stability stress ratio. Selection operation: The tournament selection mechanism is used to simulate the natural selection process, and superior individuals are selected from the current population to enter the mating pool; each time, k individuals are randomly selected from the population to form a tournament group, and the individual with the best fitness value is selected to enter the mating pool; this process is repeated until the offspring population size reaches the preset population size; Two-point crossover: For randomly paired parent individuals, two crossover points are randomly generated, and the gene segments between these two crossover points are exchanged; Dynamic mutation: Mutation operations increase the diversity of the population by introducing random perturbations. By introducing a dynamic adaptive mutation strategy, the algorithm can help escape local optima. The dynamic mutation probability and dynamic mutation length are both adaptively adjusted with the number of generations. Elite retention: In each generation, the parent population and the offspring population generated by crossover and mutation are merged. The merged population is sorted according to fitness value, and only the top n individuals with the best fitness are retained as the population for the next iteration. Restart Mechanism: A convergence stagnation monitoring mechanism is set up. When the optimal fitness is not improved for L consecutive generations, the population restart is automatically triggered. After the population restart is triggered, a first preset proportion of new individuals are generated to replace the individual with the worst fitness in the current population, thereby forcing the population to jump out of the local optimum. Diversity control: The Euclidean distance between individuals is calculated every m generations. If the average Euclidean distance is less than a preset value, the individuals in the second preset proportion are mutated to enhance population diversity and promote local search.

[0011] Furthermore, to maintain the stability of the overall tower structure's number of layers, the angle α between the main member representing the main topology and the horizontal ground, and the angle β between the diagonal member and the main member, do not participate in the cross-operation; only the number of subsequent steel strands and the parameters representing the cross-sectional dimensions are exchanged.

[0012] Furthermore, the dynamic mutation probability decays linearly according to the formula Pm = initial_mutation × (1 - gen / max_gen) + 0.1, where gen is the current generation, max_gen is the maximum number of generations, and initial_mutation is the initial mutation probability value. The dynamic mutation time decays with the generation, which is step = max(limit_value, 0.1 × (1 -gen / max_gen)). For the size parameter, a perturbation of ±step is randomly added to its current value, and the step size is set to a lower limit limit_value. All parameters are immediately checked and repaired for constraints after mutation to ensure the feasibility of the new individual.

[0013] Furthermore, the objective function for optimizing the total cost of the tower is: Cost=V_c_Main×0.1+Mass_s_Main×1.1e-3+Mass_brace×1.0e-3+Mass_prestress×1.15e-3 +∑penalty Where Cost is the total cost of the tower, V_c_Main, Mass_s_Main, Mass_brace and Mass_prestress are the engineering quantities of main column concrete, main column steel, diagonal brace steel and steel strand respectively, and penalty is the penalty introduced when the structure does not meet the design requirements of the specification.

[0014] Another objective of this invention is achieved through the following technical solution: A wind turbine lattice tower optimization design method based on an improved genetic algorithm, based on the aforementioned wind turbine lattice tower optimization design system based on an improved genetic algorithm, includes the following steps: S1. Receive tower design parameters and perform preprocessing; S2. Based on the preprocessed tower design parameters as constraints, a population that satisfies the constraints is randomly generated, and a variable-length coding strategy is adopted to adapt to the change in the number of layers of the overall tower structure. S3. Based on the finite element platform, construct a parametric finite element model for each individual generated in step S2 and perform finite element analysis; S4. Calculate the total cost and penalty for each individual based on the finite element analysis results. S5. Taking the total cost of the tower as the optimization objective, an improved genetic algorithm with the introduction of dynamic adaptive mutation strategy, elite retention strategy, restart mechanism strategy and population diversity strategy is used to complete the optimization iteration. S6. When the optimization iteration reaches the maximum number of generations or meets the convergence condition, output the optimal individual parameters and the optimal total cost.

[0015] Compared with the prior art, the present invention has the following advantages and beneficial effects: 1. Cost Advantage: By using a precise parametric finite element model and a reasonable fitness function, combined with an improved genetic algorithm, this invention can significantly reduce the manufacturing cost of lattice towers for wind turbine generators while meeting constraints such as strength, stability, frequency, and structural measures. Actual case studies have verified that the cost reduction can reach 10%-15%, providing strong support for cost control in wind power projects.

[0016] 2. Accuracy and Reliability: Based on the finite element platform, the model is constructed, taking into account the combined effect of steel-concrete composite, prestressing effect, and connection behavior of each component. It can accurately simulate the mechanical properties of the tower and ensure the reliability of the optimization results. At the same time, the penalty term in the fitness function can cover various performance constraints, further ensuring that the optimized tower meets the requirements of engineering specifications.

[0017] 3. Improved algorithm performance: The improved genetic algorithm effectively solves the problems of low search efficiency and easy getting trapped in local optima in traditional genetic algorithms by dynamically adjusting the mutation probability and step size, controlling population diversity, and introducing a restart mechanism. The algorithm's convergence speed is improved by 20%-30%, and it can find a better global solution.

[0018] 4. Efficiency and Applicability: The multi-process parallel computing framework significantly reduces optimization time and improves design efficiency several times over, meeting the efficiency requirements of actual engineering projects. This invention is not only applicable to onshore wind turbine towers, but can also be used for the optimization design of offshore wind turbine jackets after appropriate adjustments, demonstrating its wide applicability.

[0019] 5. Automation and Intelligence: Achieve automated processes from load processing, model building, parameter optimization to result output, reduce manual intervention, decrease reliance on designers' experience, improve the intelligence level of tower design, and provide technical support for large-scale design in the wind power industry. Attached Figure Description

[0020] Figure 1This is a framework diagram of the optimized design system of the present invention.

[0021] Figure 2 This is a schematic diagram of the optimized parameters for the lattice tower of the present invention.

[0022] Figure 3 This is a schematic diagram of the model composition and element types of the parametric finite element model of the present invention.

[0023] Figure 4 This is a framework diagram of the improved genetic algorithm of the present invention.

[0024] Figure 5 This is a schematic diagram of the improved genetic algorithm mutation probability and variable asynchronous length of the present invention.

[0025] Figure 6 This is a fitness evolution curve diagram of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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 some embodiments of the present invention, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0027] Example 1: like Figure 1 As shown in the figure, this embodiment provides an optimized design system for wind turbine lattice towers based on an improved genetic algorithm.

[0028] The lattice tower consists of three parts: a lower prestressed steel-concrete composite lattice structure, a middle transition section, and an upper conical steel tower. The lower structure uses steel-concrete composite as its main members, with pre-reserved ducts in the center of the concrete for arranging steel strands. The diagonal braces are hollow circular steel tubes, and the prestressed tendons are tensioned from the bottom to the top of the tower. Figure 2 As shown.

[0029] The wind turbine lattice tower optimization design system includes, Parameter input module 1 is used to receive tower design parameters and perform preprocessing. The tower design parameters include multi-condition ultimate load data (such as axial force, bending moment, shear force, and torsion at different heights), tower foundation geometric parameters (such as blade tip clearance distance, total tower height, tower top diameter, and nose mass), material property parameters (such as elastic modulus, density, and strength design value of steel, concrete, and steel strand), and specification constraint parameters (such as strength safety factor, frequency limit, and diameter-to-thickness ratio limit).

[0030] Population initialization module 2 is used to randomly generate a population based on the preprocessed tower design parameters as constraints, and employs a variable-length encoding strategy to adapt to changes in the overall number of tower structure layers; specifically, it performs the following operations: Based on the preprocessed tower design parameters as constraints (such as the diameter-to-thickness ratio of main members and diagonal members), a random population is generated. Each individual corresponds to a set of tower design structural parameters, including: the angle α between the main member and the horizontal ground, the angle β between the diagonal member and the main member, the number of steel strands N, and the diameter D and thickness T of each main member and the diameter d and thickness t of each diagonal member. In particular, since the total number of tower layers is dynamically affected by the angle α between the main member and the horizontal ground and the angle β between the diagonal member and the main member, a variable-length coding strategy is adopted to adapt to the non-fixed number of design variables. A standard constraint checking algorithm is embedded in the population initialization process to ensure that all initial individuals meet the construction requirements and to avoid invalid individuals from entering the subsequent optimization process.

[0031] Module 3, Finite Element Modeling and Analysis, is used to construct parametric finite element models for each individual in the population and perform finite element analysis based on the OpenSeesPy open-source finite element platform. The parametric finite element model consists of the main member 301, diagonal member 302, steel strand 303, transition section 304, and tower 305, with element types shown in Figure 3. The main member is simulated using elasticBeamColumn elements, based on the equivalent combined stiffness formula in the "Technical Specification for Steel-Concrete Composite Structures" (GB50936-2014), with the material parameters of the steel tube and concrete of the tower input to simulate the mechanical behavior of the steel-concrete composite structure working together. The diagonal member is simulated using elasticBeamColumn elements, with only the steel material parameters input. The steel strand is simulated using Truss elements, with the initial strain set using InitStrainMaterial material to simulate the prestressing effect, bearing only axial force. The transition section and tower are simulated using ASDShellQ4. The simulation uses shell elements (ASD four-node shell elements), and the transition section is connected to the main body using the equalDOF command (equal degree of freedom constraint command). To transfer the load from the upper part of the wind turbine system, a central node is set at the top of the tower, and the rigidLink command (rigid connection command) is used to establish a rigid connection between the central node and the top node of the tower. The nacelle, hub, and upper 306 mass of the wind turbine are simulated as concentrated mass points using the mass command (mass command), and rigidly connected to the central node using rigidLink. The bottom node of the tower is set as a consolidation constraint using the fix command (consolidation command). The module supports multi-process parallel computing and can perform finite element analysis on multiple individuals in the population at the same time, outputting the first natural frequency of the tower, the strength stress ratio of the main body and the diagonal member, and the stability stress ratio.

[0032] Algorithm optimization module 4 calculates the total cost and penalty term for each individual based on the finite element analysis results. Then, using the total tower cost as the optimization objective, it employs an improved genetic algorithm incorporating dynamic adaptive mutation, elite retention, restart mechanisms, and population diversity strategies to complete the optimization iteration. The overall operation flow of this module is as follows: Figure 4 As shown, the details are as follows: Fitness Assessment: The optimization objective is the total cost of the lattice tower, including concrete cost, steel cost, and steel strand cost, while introducing a penalty term. The penalty term constrains the aspect ratio of the main members, the hoop coefficient, the solidity ratio, the steel content of the cross-section, the slenderness ratio of the diagonal members, the first natural frequency of the tower, the ratio of strength stress in the main members to that in the diagonal members, and the ratio of stability stress. Furthermore, the objective function for optimizing the total cost of the tower is: Cost = V_c_Main × 0.1 + Mass_s_Main × 1.1e-3 + Mass_brace × 1.0e-3 + Mass_prestress × 1.15e-3 + ∑penalty Where Cost is the total cost of the tower, V_c_Main, Mass_s_Main, Mass_brace, and Mass_prestress are the engineering quantities (volume or mass) of the main body concrete, main body steel, diagonal brace steel, and steel strand, respectively, and penalty is the penalty introduced when the structure does not meet the design requirements of the specification.

[0033] Selection operation: The tournament selection mechanism is used to simulate the natural selection process, and superior individuals are selected from the current population to enter the mating pool; each time, k individuals are randomly selected from the population to form a tournament group, and the individual with the best fitness value is selected to enter the mating pool; this process is repeated until the offspring population size reaches the preset population size; Two-point crossover: For randomly paired parent individuals, two crossover points are randomly generated, and the gene segments between these two crossover points are exchanged; to maintain the stability of the overall tower structure layer number, the angle α between the main limb representing the main topology and the horizontal ground, and the angle β between the diagonal bar and the main limb do not participate in the crossover operation, only the subsequent number of steel strands and the parameters representing the cross-sectional dimensions are exchanged; this design effectively maintains the rationality of the structural parameters while exploring new solutions; Dynamic mutation: Mutation operations increase population diversity by introducing random perturbations, and a dynamic adaptive mutation strategy helps the algorithm escape local optima; both the dynamic mutation probability and the dynamic mutation length are adaptively adjusted with the number of generations; for example... Figure 5As shown, the dynamic mutation probability decays linearly according to the formula Pm = initial_mutation × (1 - gen / max_gen) + 0.1, where gen is the current generation, max_gen is the maximum number of generations, and initial_mutation is the initial mutation probability value. This design allows the algorithm to maintain strong global exploration capabilities in the early stages of evolution, while focusing on fine-tuning local searches in the later stages. The dynamic mutation probability decays with the number of generations, and is given by step = max(limit_value, 0.1 × (1 - gen / max_gen)). For the size parameter, a perturbation of ±step is randomly added to its current value, and a lower limit limit_value is set for the step size to ensure that the algorithm still has fine-tuning capabilities in the later stages. All parameters undergo constraint checks and repairs immediately after mutation to ensure the feasibility of the new individual.

[0034] Elite Preservation: To prevent the loss of the currently obtained optimal solution during evolution, the algorithm introduces an elite preservation strategy. In each generation, the parent population and the offspring population generated by crossover and mutation are merged. The merged population is sorted according to its fitness value, and only the top n individuals with the best fitness are retained as the population for the next iteration. This strategy guarantees the convergence of the algorithm and ensures that the optimal solution is not destroyed by crossover and mutation operations.

[0035] Restart Mechanism: A convergence stagnation monitoring mechanism is set up. When the optimal fitness is not improved for L consecutive generations, the population restart is automatically triggered. After the population restart is triggered, a% of new individuals are generated to replace the individual with the worst fitness in the current population, thereby forcing the population to jump out of the local optimum.

[0036] Diversity control: The Euclidean distance between individuals is calculated every m generations. If the average Euclidean distance is less than the preset value b, c% of the individuals are mutated to enhance population diversity and promote local search.

[0037] Result output module 5 is used to output the optimal individual parameters and the optimal total cost when the optimization iteration reaches the maximum number of generations or meets the convergence condition.

[0038] The modules mentioned above are connected sequentially through data interfaces to form a closed-loop automated optimization process. The system architecture is shown in Figure 2.

[0039] Example 2: This embodiment provides an optimization design method for wind turbine lattice towers based on an improved genetic algorithm, including the following steps: S1. Receive tower design parameters and preprocess them; the tower design parameters specifically include: Basic parameters: The rated speed of the unit is 8 rpm, the blade tip height is 12.0 m, the blade tip clearance is 20.16 m, the tower top diameter is 4.484 m, the lattice tower height is 117.0 m, and the total tower height is 135.0 m. These parameters directly determine the basic geometric shape and spatial layout of the tower.

[0040] Material properties: Q355 steel: elastic modulus 2.06e11Pa, shear modulus 7.923e10Pa, density 7850kg / m³, design strength 305e6Pa, standard strength 355e6Pa; C80 concrete: elastic modulus 3.8e10Pa, shear modulus 1.583e10Pa, density 2550kg / m³, design strength 35.9e6Pa; steel strand: elastic modulus 1.95e11Pa, reserved duct diameter 0.340m.

[0041] Load data: Ultimate load information for each working condition at different heights, providing a basis for subsequent load interpolation calculations at each floor height of the lattice tower.

[0042] S2. Based on the preprocessed tower design parameters as constraints, a population satisfying the constraints is randomly generated, and a variable-length coding strategy is adopted to adapt to changes in the overall number of tower structure layers. The constraints are: the diameter-to-thickness ratio of the main limb is less than 135×235 / fy (fy is the standard value of steel), and the diameter-to-thickness ratio of the diagonal members is less than 100×235 / fy. The design structural parameter set for each individual tower is: [α, β, N, [D1, T1, d1, t1], [D2, T2, d2, t2], ...].

[0043] S3. Based on the OpenSeesPy open-source finite element platform, construct a parametric finite element model for each individual generated in step S2 and perform finite element analysis. The transition section has a height of 3.6m, is divided into 32 nodes in the circumferential direction, and is divided into 6 layers in the height direction, with a shell thickness of 0.095m. The total height of the tower is the total height minus the height of the lattice section and the transition section, with a shell thickness of 0.028m. It is divided into 16 layers along the height, sharing nodes with the top of the transition section. Other components are established according to material properties and element types.

[0044] S4. Calculate the total cost and penalty for each individual based on the finite element analysis results; according to market research, the unit cost of the main concrete is 0.1 million yuan / m³. 3The unit cost of main steel is 11,000 yuan / ton, the unit cost of diagonal steel is 10,000 yuan / ton, and the unit cost of steel strand is 11,500 yuan / ton. The penalty settings are as follows: When the main member diameter-to-thickness ratio (ratio) > 20, penalty = (ratio - 20)² × 1e3; when the main member clamping ratio (θ) < 0.5, penalty = (θ - 0.5)² × 1e6; when θ > 2.0, penalty = (θ - 2.0)² × 1e6; when the main member steel content (αs) < 0.06, penalty = (αs - 0.06)² × 1e7; when αs > 0.20, penalty = (αs - 0.20)² × 1e7; when the main member solidity (ψ) > 0.25, penalty = (ψ - 0.25)² × 1e7; when the diagonal member slenderness ratio (Lr) > 180, penalty = (Lr - 180)² × 1e3; when the tower's first natural frequency < 0.307 Hz (wind turbine frequency), penalty = (0.307-frequencies[0])²×1e7; When the strength or stability verification result of the main member and diagonal member is σ≥0.95, penalty= (σ-0.95)² ×1e7. Finally, the formula for calculating the total cost of the tower is: Cost =V_c_Main×0.1+Mass_s_Main×1.1e-3+Mass_brace×1.0e-3+Mass_prestress×1.15e-3 +∑penalty.

[0045] S5. Using the total cost of the tower as the optimization objective, an improved genetic algorithm incorporating dynamic adaptive mutation, elite retention, restart mechanisms, and population diversity strategies is used to complete the optimization iteration. Specifically: Initial parameters are set as follows: population size 350, maximum generation count 120, crossover probability 0.7, and initial mutation probability 0.3. A tournament selection strategy is adopted, randomly selecting 5 individuals from the parent population each time, and choosing the individual with the lowest fitness for the next generation. This process is repeated 350 times to form a new population. Crossover is performed at a probability of 0.7. The dynamic mutation probability decreases linearly according to the formula 0.3 * (1 - gen / max_gen) + 0.1, with an initial mutation probability of 0.4, decreasing to 0.1 at the end. The dynamic mutation probability is calculated using the formula max(0.005, 0.1 * (1 - gen / max_gen)). In each generation of optimization, the parent generation and newly generated offspring are merged and reordered, and the top 350 individuals are retained based on fitness to achieve elite preservation. A restart mechanism check is performed; if the optimal fitness value has not improved for five consecutive generations, new individuals equivalent to 20% of the population are generated to replace the worst-performing individuals in the current population. Furthermore, a population diversity check is performed every 10 generations; if the average Euclidean distance between individuals is less than 0.8, mutation operations are performed on 30% of the individuals to enhance diversity.

[0046] S6. When the optimization iteration reaches the maximum number of generations or meets the convergence condition, output the optimal individual parameters and the optimal total cost value; for example, in this embodiment, when the iteration reaches the 120th generation, the termination condition is met, and the optimal individual and the minimum fitness value are output. Figure 6 The figure shows the fitness change curve during the algorithm optimization process in this embodiment.

[0047] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope disclosed in the present invention, based on the technical solution and inventive concept of the present invention, shall fall within the scope of protection of the present invention.

Claims

1. A wind turbine lattice tower optimization design system based on an improved genetic algorithm, characterized in that: include, The parameter input module is used to receive tower design parameters and perform preprocessing. The population initialization module is used to randomly generate a population based on the preprocessed tower design parameters as constraints, and to adopt a variable-length coding strategy to adapt to changes in the number of layers in the overall tower structure. The finite element modeling and analysis module is used to construct parametric finite element models for each individual in the population and perform finite element analysis based on the finite element platform; wherein, the parametric finite element model includes main members, diagonal bars, steel strands, transition sections and towers; The algorithm optimization module is used to calculate the total cost and penalty term of each individual based on the finite element analysis results of each individual. Then, with the total cost of the tower as the optimization objective, the improved genetic algorithm, which introduces dynamic adaptive mutation strategy, elite retention strategy, restart mechanism strategy and population diversity strategy, completes the optimization iteration. The results output module is used to output the optimal individual parameters and the optimal total cost when the optimization iteration reaches the maximum number of generations or meets the convergence condition.

2. The wind turbine lattice tower optimization design system based on improved genetic algorithm according to claim 1, characterized in that: The tower design parameters include multi-condition ultimate load data, tower foundation geometric parameters, material property parameters, and specification constraint parameters.

3. The wind turbine lattice tower optimization design system based on improved genetic algorithm according to claim 1, characterized in that: The population initialization module performs the following operations: Based on the preprocessed tower design parameters as constraints, a random population is generated, with each individual corresponding to a set of tower design structural parameters. A variable-length encoding strategy is adopted to adapt to the non-fixed number of design variables. A standard constraint check algorithm is embedded in the population initialization process to ensure that all initial individuals meet the construction measures requirements and to avoid invalid individuals from entering the subsequent optimization process.

4. The wind turbine lattice tower optimization design system based on improved genetic algorithm according to claim 3, characterized in that: The tower design structural parameters include: the angle α between the main member and the horizontal ground, the angle β between the diagonal member and the main member, the number of steel strands N, and the diameter D and thickness T of each main member, and the diameter d and thickness t of each diagonal member.

5. The wind turbine lattice tower optimization design system based on improved genetic algorithm according to claim 1, characterized in that: The main limbs are simulated using elastic beam-column elements. Based on the equivalent combined stiffness formula, the material parameters of the steel pipe and concrete of the tower are input to simulate the mechanical behavior of the steel pipe and concrete working together. The diagonal members are simulated using elastic beam-column elements, with only the steel material parameters input. The steel strands are simulated using truss elements, with the initial strain set by the initial strain material to simulate the prestress effect, bearing only axial force. The transition section and tower are simulated using ASD four-node shell elements, and the connection between the transition section and the main limbs is achieved through equal degree of freedom constraint commands. To transfer the load from the upper part of the wind turbine system, a central node is set at the top of the tower, and a rigid connection command is used to establish a rigid connection between the central node and the top node of the tower. The nacelle, hub, and upper mass of the wind turbine are simulated as concentrated mass points through mass commands and rigidly connected to the central node through rigid connections. The bottom node of the tower is set as a consolidation constraint using consolidation commands. By simultaneously performing finite element analysis on multiple individuals in the population, the first natural frequency of the tower, the strength stress ratio of the main limbs and diagonal members, and the stability stress ratio are output.

6. The wind turbine lattice tower optimization design system based on improved genetic algorithm according to claim 1, characterized in that: The algorithm optimization module performs the following operations: Fitness assessment: The optimization objective is the total cost of the tower, which includes the cost of concrete, steel and steel strands, and a penalty term is introduced. The penalty term is used to constrain the aspect ratio of the main members, the hoop coefficient, the solidity ratio and the steel content of the section, the slenderness ratio of the diagonal members, the first natural frequency of the tower, the strength stress ratio of the main members and the diagonal members, and the stability stress ratio. Selection operation: A tournament selection mechanism is used to simulate the natural selection process, selecting superior individuals from the current population to enter the mating pool; Each time, k individuals are randomly selected from the population to form a tournament group, and the individual with the best fitness value is selected to enter the mating pool; this process is repeated until the offspring population size reaches the preset population size. Two-point crossover: For randomly paired parent individuals, two crossover points are randomly generated, and the gene segments between these two crossover points are exchanged; Dynamic mutation: Mutation operations increase the diversity of the population by introducing random perturbations. By introducing a dynamic adaptive mutation strategy, the algorithm can help escape local optima. The dynamic mutation probability and dynamic mutation length are both adaptively adjusted with the number of generations. Elite retention: In each generation, the parent population and the offspring population generated by crossover and mutation are merged. The merged population is sorted according to fitness value, and only the top n individuals with the best fitness are retained as the population for the next iteration. Restart Mechanism: A convergence stagnation monitoring mechanism is set up. When the optimal fitness is not improved for L consecutive generations, the population restart is automatically triggered. After the population restart is triggered, a first preset proportion of new individuals are generated to replace the individual with the worst fitness in the current population, thereby forcing the population to jump out of the local optimum. Diversity control: The Euclidean distance between individuals is calculated every m generations. If the average Euclidean distance is less than a preset value, the individuals in the second preset proportion are mutated to enhance population diversity and promote local search.

7. The wind turbine lattice tower optimization design system based on improved genetic algorithm according to claim 6, characterized in that: To maintain the stability of the overall tower structure's number of layers, the angle α between the main member representing the main topology and the horizontal ground, and the angle β between the diagonal member and the main member, do not participate in the cross-operation. Only the number of subsequent steel strands and the parameters representing the cross-sectional dimensions are exchanged.

8. The wind turbine lattice tower optimization design system based on improved genetic algorithm according to claim 6, characterized in that: The dynamic mutation probability decays linearly according to the formula Pm = initial_mutation × (1 - gen / max_gen) + 0.1, where gen is the current generation, max_gen is the maximum number of generations, and initial_mutation is the initial mutation probability value. The dynamic mutation step decays with the generation, which is step = max(limit_value, 0.1 × (1 - gen / max_gen)). For the size parameter, a perturbation of ±step is randomly added to its current value, and the step size is set to a lower limit limit_value. All parameters are immediately checked and repaired for constraints after mutation to ensure the feasibility of the new individual.

9. The wind turbine lattice tower optimization design system based on improved genetic algorithm according to claim 1, characterized in that: The objective function for optimizing the total cost of the tower is: Cost=V_c_Main×0.1+Mass_s_Main×1.1e-3+Mass_brace×1.0e-3+Mass_prestress×1.15e-3 +∑penalty Where Cost is the total cost of the tower, V_c_Main, Mass_s_Main, Mass_brace and Mass_prestress are the engineering quantities of main column concrete, main column steel, diagonal brace steel and steel strand respectively, and penalty is the penalty introduced when the structure does not meet the design requirements of the specification.

10. A method for optimizing the design of wind turbine lattice towers based on an improved genetic algorithm, characterized in that: The wind turbine lattice tower optimization design system based on the improved genetic algorithm according to any one of claims 1 to 9 includes the following steps: S1. Receive tower design parameters and perform preprocessing; S2. Based on the preprocessed tower design parameters as constraints, a population that satisfies the constraints is randomly generated, and a variable-length coding strategy is adopted to adapt to the change in the number of layers of the overall tower structure. S3. Based on the finite element platform, construct a parametric finite element model for each individual generated in step S2 and perform finite element analysis; S4. Calculate the total cost and penalty for each individual based on the finite element analysis results. S5. Taking the total cost of the tower as the optimization objective, an improved genetic algorithm with the introduction of dynamic adaptive mutation strategy, elite retention strategy, restart mechanism strategy and population diversity strategy is used to complete the optimization iteration. S6. When the optimization iteration reaches the maximum number of generations or meets the convergence condition, output the optimal individual parameters and the optimal total cost.

Images