A method for cutter layout of a circular cutterhead for a tunnel boring machine.

Abstract

This invention relates to the field of tunnel boring machine (TBM) design technology and discloses a tool layout method for a circular cutterhead of a TBM. The method includes: constructing a mathematical model of tool layout incorporating cutting force data from a CSM model; determining the optimization objective and constraints for minimizing eccentric force and eccentric moment; initializing a gray wolf population by mapping discrete permutations to continuous vectors using a continuous encoding strategy and calculating fitness values; executing a global iteration of the gray wolf optimization algorithm, using social hierarchy to guide population updates; embedding a tabu search strategy in the global iteration to perform local neighborhood search on the optimal individual to generate and selectively replace the optimal solution; and finally outputting the tool layout scheme and verifying its feasibility. This invention effectively solves the problem of tool layout easily getting trapped in local optima through the synergistic cooperation of the gray wolf algorithm and tabu search, improving the mechanical balance and solution efficiency of the cutterhead design, and reducing design and production costs.

Description

Technical Field

[0001] This invention relates to the field of tunnel boring machine design technology, specifically to a method for the cutter layout of a circular cutterhead in a tunnel boring machine. Background Technology

[0002] Currently, in the fields of tunnel engineering and underground space development, tunnel boring machines (TBMs) are the core excavation equipment, and their operational performance directly determines the project progress and cost. The circular cutterhead, as a key component for rock breaking at the front end of the TBM, has a crucial element in the arrangement of its cutters. A reasonable cutter layout not only affects the efficiency of rock breaking but also directly impacts the overall force balance, vibration amplitude, and service life of the main bearings on the cutterhead. Cutter layout design is essentially a complex combined optimization problem involving multiple constraints such as geometric interference, mechanical balance, and cutting sequence, requiring the search for the optimal solution for resource allocation within the limited space of the cutterhead.

[0003] To address the aforementioned tool layout design requirements, existing technologies often employ semi-automated layout based on empirical formulas or a single heuristic algorithm for design assistance. Designers typically begin by determining the total number of tools based on geological parameters, then set preliminary tool installation positions according to an Archimedean spiral or concentric circle trajectory. Subsequently, a simplified mechanical model is used to calculate the thrust and torque of the tool head, and the radial spacing and installation angle of the tools are fine-tuned manually or through simple iterative algorithms until the solution passes geometric interference checks and static balance verification. This type of method focuses on meeting basic geometric layout requirements, approximating a feasible solution through continuous trial and error.

[0004] However, existing technologies still have shortcomings in handling high-dimensional nonlinear constraints. First, traditional mechanical calculation models often neglect the cooperative cutting effect between tools, leading to a large deviation between theoretical forces and actual working conditions, which can easily cause uneven wear of the tool head or uneven loading of the main bearing. Second, single optimization algorithms have limited global search capabilities when facing complex discrete layout spaces, are prone to getting trapped in local optima, and struggle to simultaneously minimize eccentric forces and balance torque. Furthermore, the iterative calculation process relying on manual experience is inefficient, making it difficult to obtain a high-precision layout scheme that meets multiple engineering constraints in a short time, increasing the trial-and-error costs of subsequent manufacturing and construction.

[0005] Therefore, the present invention provides a cutter layout method for a circular cutterhead of a tunnel boring machine to overcome the shortcomings of the prior art. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a method for the cutter layout of a circular cutterhead for tunnel boring machines (TBMs). This method solves the technical problems of low solution efficiency, susceptibility to local optima, and difficulty in balancing mechanical equilibrium and multiple engineering constraints that arise from the reliance on manual experience and a single algorithm in existing TBM cutterhead cutter layout methods.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for cutter layout of a circular cutterhead in a tunnel boring machine, comprising the following steps:

[0008] A mathematical model of tool layout is constructed. Combining the geometric parameters of the tool head and the cutting force data of the CSM model, the optimization objectives of minimizing eccentric force and eccentric moment and the constraints that the mathematical model of tool layout must satisfy are determined.

[0009] A continuous coding strategy is used to initialize the gray wolf population. The continuous position vectors of individual gray wolves are decoded into discrete tool arrangement schemes, and the fitness value is calculated based on the mathematical model of the tool layout.

[0010] The gray wolf optimization algorithm is executed globally iterated. Based on the fitness value, α wolf, β wolf and δ wolf are selected. The position vectors of the α wolf, β wolf and δ wolf are used to guide the continuous position vector update of other gray wolf individuals in the gray wolf population.

[0011] In the global iteration, a tabu search strategy is embedded. When the triggering condition is met, the α wolf with the best fitness is determined as the initial solution, and a local neighborhood search is performed to generate and replace the optimal solution.

[0012] The process terminates when the preset maximum number of iterations is reached, outputs the tool layout scheme corresponding to the α wolf at the final moment, and uses the tool layout mathematical model to verify the feasibility of the tool layout scheme.

[0013] By adopting the above technical solution, since a mechanical objective function and multiple constraints based on the CSM model are established, and the discrete layout is mapped to continuous encoding, combined with the global guidance of the Grey Wolf algorithm and the local breakthrough mechanism of tabu search, the tool layout scheme that satisfies static balance and optimal mechanical performance can be solved automatically. This solves the problems of low efficiency and difficulty in taking multiple constraints into account in traditional manual layout, while overcoming the defect of single algorithms being prone to getting trapped in local optima, thus improving the mechanical balance and design efficiency of tool head design.

[0014] Preferably, the process of constructing the mathematical model of the tool layout includes: establishing a spatial rectangular coordinate system, defining a tool set and a candidate slot set, and determining the spatial coordinates and installation angle of each slot in the candidate slot set; setting the optimization objective as minimizing the sum of the absolute values ​​of the resultant eccentric forces and the sum of the absolute values ​​of the resultant eccentric moments of the tool head in the three coordinate axis directions; setting the constraints, which include: an allocation constraint to ensure that each slot is assigned only one tool and each tool is assigned only to one slot; a structural constraint to ensure that the tool model matches the slot type and is not installed in a prohibited area; and a static balance constraint to limit the offset of the tool head's centroid within the allowable tolerance range.

[0015] By adopting the above technical solution, the geometric benchmark, physical objective and engineering boundary conditions for optimization solution are clarified, ensuring the applicability and compliance of the optimization results in actual engineering.

[0016] Preferably, the process of determining the optimization objective of minimizing eccentric force and eccentric moment includes: calculating the resultant force components of the eccentric force of the cutter head in the three coordinate axes based on the normal force and lateral force in the cutting force data of the CSM model, combined with the installation angle of the slot and the decision variables; calculating the resultant force components of the eccentric moment of the cutter head in the three coordinate axes based on the normal force, the lateral force and the spatial coordinates of the slot, combined with the decision variables; and summing the absolute values ​​of the resultant force components of the eccentric force in the three directions with the absolute values ​​of the resultant force components of the eccentric moment in the three directions, and determining the sum as the optimization objective value.

[0017] By adopting the above technical solution, a quantitative assessment of the complex stress state during cutterhead tunneling was achieved, transforming multi-dimensional mechanical balance indicators into a single calculable target, and providing a clear direction for algorithm optimization.

[0018] Preferably, the process of initializing the gray wolf population using a continuous coding strategy and decoding the continuous position vectors of individual gray wolves into a discrete tool arrangement scheme includes: generating the continuous position vector with the same dimension as the number of tools, and limiting the value range of each dimension in the continuous position vector to a preset closed interval; sorting the values ​​in the continuous position vector in ascending order and recording the original index of each value in the original vector; determining the original index sequence generated after sorting as the installation order of the tools, and sequentially assigning the corresponding numbered tools to the corresponding numbered slots, thereby decoding the continuous position vector into the discrete tool arrangement scheme.

[0019] By adopting the above technical solution and through the sorting index mapping mechanism, a correspondence between continuous numerical space and discrete permutation problem is established, ensuring that any continuous vector can be converted into a legal non-repeating tool arrangement and avoiding the generation of illegal solutions.

[0020] Preferably, the process of calculating the fitness value of the individual gray wolf based on the tool layout mathematical model includes: calculating a target term, which is the absolute value of the sum of the eccentric forces and eccentric moments generated by all tools in the discrete tool arrangement scheme at the current layout position; calculating a penalty term, checking whether the decoded discrete tool arrangement scheme violates the allocation constraint, the structural constraint, or the static balance constraint; if any of the constraints are violated, a preset penalty value is added to the target term, and the accumulated result is determined as the fitness value.

[0021] By adopting the above technical solution and using the penalty function mechanism to handle constraints, the algorithm is forced to automatically eliminate infeasible solutions that violate engineering constraints during the iteration process, thus guiding the population to converge quickly to the feasible solution space.

[0022] Preferably, the process of using the position vectors of the α wolf, β wolf, and δ wolf to guide the continuous position vector update of other gray wolf individuals in the gray wolf population includes: calculating the weighted distances of the other gray wolf individuals relative to the α wolf, β wolf, and δ wolf, respectively; calculating the three potential movement position components of the other gray wolf individuals after being affected by the α wolf, β wolf, and δ wolf, respectively, based on the weighted distances; calculating the arithmetic mean of the three potential movement position components, and determining the arithmetic mean as the final position vector of the other gray wolf individuals after the update.

[0023] By adopting the above technical solution, a social hierarchy is constructed using the three individuals with the best fitness in the population, guiding the entire population to move towards the potential optimal solution region, thus ensuring the convergence trend of the algorithm.

[0024] Preferably, the process of using the position vectors of the α wolf, β wolf, and δ wolf to guide the continuous position vector update of other gray wolf individuals in the gray wolf population further includes: calculating a convergence factor, which decreases linearly from an initial value to zero as the number of iterations increases; calculating a coefficient vector using the convergence factor and a randomly generated vector; and using the coefficient vector to calculate a weighted distance between the α wolf, β wolf, and δ wolf and the other gray wolf individuals, thereby obtaining the weighted distance.

[0025] By adopting the above technical solution and dynamically adjusting the convergence factor, the algorithm's global exploration capability in the early stage of iteration and its local development capability in the later stage are balanced.

[0026] Preferably, the process of performing a local neighborhood search to generate and selectively replace the optimal solution includes: determining whether the current iteration number is a multiple of a preset value; if so, initiating a tabu search and determining the position vector of the α wolf as the current solution; within the inner loop of the tabu search, performing neighborhood operations on the current solution to generate a candidate solution set, the neighborhood operations including swapping and insertion operations; calculating the fitness of each candidate solution in the candidate solution set, selecting the non-tabu optimal candidate solution and updating the current solution, while recording the operation behavior of generating the optimal candidate solution in the tabu table; after the local neighborhood search is completed, comparing the fitness value of the obtained local optimal solution with the fitness value of the original α wolf; if the fitness value of the local optimal solution is better than that of the original α wolf, then replacing the position vector of the original α wolf with the position vector of the local optimal solution and updating the fitness value; if the fitness value of the local optimal solution is not better than that of the original α wolf, then keeping the position of the original α wolf unchanged.

[0027] By adopting the above technical solution, the tabu list's memory function prevents the algorithm from repeatedly searching on the same path, and the swap and insertion operations enhance the fine search capability for local extreme value regions, effectively solving the problem that the Grey Wolf algorithm is prone to getting trapped in local optima in the later stages.

[0028] Preferably, the process of verifying the feasibility of the tool layout scheme using the tool layout mathematical model includes: substituting the final output tool layout scheme into the tool layout mathematical model, recalculating the eccentric force and the eccentric moment, and verifying whether they are within the optimization range; checking whether the tool layout scheme satisfies the allocation constraints and the structural constraints, ensuring that the tool installation position is unique and the model matches; calculating the total moment generated by the mass of all tools in the tool layout scheme, and verifying whether the centroid offset of the total moment in the XY plane and Z-axis direction satisfies the static balance constraint.

[0029] By adopting the above technical solution, a final closed-loop verification mechanism is provided, ensuring the absolute reliability of the output tool layout design scheme at both the physical and engineering levels.

[0030] This invention provides a cutter layout method for a circular cutterhead of a tunnel boring machine. It has the following advantages:

[0031] 1. This invention achieves quantitative control of the stress state of the cutterhead during tunneling by constructing a multi-objective mathematical model based on cutting force data from the CSM model and introducing strict engineering constraints such as static balance and structural matching. This design can reduce the eccentric force and overturning moment of the cutterhead during tunneling, effectively ensuring the mechanical balance of the cutter layout, thereby improving the stability of the tunnel boring machine and extending the service life of the equipment.

[0032] 2. This invention achieves an organic combination of global optimization and local breakthrough by embedding a tabu search strategy into the global iteration of the Grey Wolf optimization algorithm, utilizing neighborhood operations such as swapping and insertion, as well as a tabu list mechanism. This hybrid strategy effectively overcomes the defect of single biomimetic algorithms easily getting trapped in local optima, enhances the algorithm's fine search capability in the solution space, and thus obtains tool layout schemes with higher accuracy and stronger robustness.

[0033] 3. This invention employs a continuous coding strategy to map discrete combinatorial optimization problems into continuous space solutions, coupled with an automated iterative verification process, replacing the traditional manual, experience-based layout method. This not only shortens the calculation cycle of tool holder design and improves design and development efficiency, but also avoids interference risks in subsequent manufacturing and assembly through pre-emptive feasibility verification, effectively reducing trial-and-error costs in production. Attached Figure Description

[0034] Figure 1 This is an overall flowchart of the cutter layout method for a circular cutterhead of a tunnel boring machine according to an embodiment of the present invention;

[0035] Figure 2 This is a schematic diagram of the tool groove layout and coordinate system definition in an embodiment of the present invention. Detailed Implementation

[0036] The technical solutions in 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. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] See attached document Figure 1 This invention provides a method for the cutter layout of a circular cutterhead for a tunnel boring machine. This method achieves automated solution and optimization of the cutter layout by using a hybrid gray wolf optimization algorithm and a tabu search algorithm.

[0038] The method first performs a model building step. Geometric parameters of the tunnel boring machine (TBM) cutterhead and cutting force data are acquired, with the cutting force data based on the CSM model. Combining the physical structural characteristics of the cutterhead, a mathematical model of the cutter layout is established, including an objective function and constraints. This mathematical model aims to minimize the sum of the absolute values ​​of the eccentric forces and the sum of the absolute values ​​of the eccentric moments of the cutterhead in the three spatial coordinate axes. Simultaneously, constraints are set that must be satisfied, including allocation constraints ensuring only one cutter is assigned to each slot, structural constraints matching the cutter type with the slot, and static balance constraints limiting the offset of the cutterhead's center of gravity.

[0039] After model construction, encoding and population initialization are performed. A continuous encoding strategy is adopted to map the discrete tool layout combination optimization problem into position vectors in a continuous numerical space. Multiple position vectors are randomly generated within a predetermined value range to form an initial gray wolf population. According to the decoding rules, the continuous position vector of each gray wolf is converted into a specific physical arrangement scheme of the tool, and substituted into the mathematical model to calculate the fitness value. The population is sorted according to the fitness value, and the three solutions with the best fitness are selected and defined as α wolf, β wolf, and δ wolf, respectively. These three alpha wolves will serve as guiding references for population position updates in subsequent iterations.

[0040] Subsequently, the global optimization iteration process of the Grey Wolf optimization algorithm begins. For example... Figure 1As shown in the main loop on the left, based on the position update formula of the gray wolf optimization algorithm, the continuous position vectors of other gray wolves in the population are calculated and updated using the position vectors of α, β, and δ wolves in the current iteration step. This step drives the entire population to approach the optimal solution region in the solution space by simulating the encirclement and hunting behavior of the gray wolf pack, thus completing the global search.

[0041] During the iterative process of the global search, a tabu search strategy is periodically embedded for local improvement. A trigger mechanism for the tabu search is set, for example, triggering it once every predetermined number of global iterations. When the trigger condition is met, the tabu search is performed using the best α wolf in the current population as the initial solution. For example... Figure 1 As shown in the right-hand branch, this tabu search process employs two neighborhood structure operations: swapping and insertion. Candidate solutions are generated within the neighborhood of the current solution, and a tabu list is used to record these operations to prevent search loops. During the tabu search, the fitness of the candidate solutions is calculated. If a new solution with better fitness than the current α wolf is found, the original α wolf position is directly replaced by this new solution, thereby enhancing the algorithm's ability to escape local optima.

[0042] Finally, the termination judgment and scheme verification steps are executed. It is determined whether the current iteration count has reached the preset maximum iteration count. If not, the loop of position update and local improvement continues; if it has, the algorithm terminates. The cutter layout scheme corresponding to the α wolf at the final moment is output as the optimal solution. Using the mathematical model constructed in step one, the feasibility of the output layout scheme is finally verified, confirming that it strictly satisfies all constraints such as geometric interference, static balance, and structural allocation, thus completing the cutter layout design of the shield machine's circular cutterhead.

[0043] See attached document Figure 2 Establishing a mathematical optimization model for the cutterhead layout is one of the core steps in this embodiment. This embodiment of the invention provides a cutterhead layout method for a circular cutterhead of a tunnel boring machine. When constructing the mathematical model, it first combines the attached... Figure 2 The geometric features of the cutterhead shown establish a spatial rectangular coordinate system. The center of the cutterhead is defined as the origin, the plane containing the cutterhead is the XOY plane, and the tunnel boring machine's advancing direction is... Axial direction. The cutter head contains a quantity of The slots for the tools to be placed. Define the tool set as... , including from 1 to A sequence of integers. Define the candidate slot set as... , including from 1 to The integer sequence, in this embodiment, represents the number of candidate slots. Equal to the number of cutting tools Each slot Its spatial position is uniquely determined by its coordinates in the coordinate system and its installation angle.

[0044] To optimize the cutterhead layout, this mathematical model sets an objective function aimed at minimizing the unbalanced forces and overturning moments of the cutterhead during the tunneling process. This objective function is expressed as minimizing the sum of the absolute values ​​of the resultant forces and resultant moments along the three coordinate axes, and its mathematical expression is as follows:

[0045] ;

[0046] The objective function described above involves calculating the resultant force and resultant moment in various directions. The specific formulas for calculating these components are as follows:

[0047]

[0048] In the above formula, Denotes a decision variable if and only if its number is The cutting tool is installed in the numbered When the slot is filled, the variable takes the value of 1; otherwise, it takes the value of 0. Indicates the number is The normal force acting on the cutting tool; Indicates the number is The lateral force acting on the cutting tool. The values ​​of these two mechanical parameters were calculated using a model from the Colorado School of Mines. (Symbols) Indicates the number is The installation angle of the slot. , , They represent the numbers respectively. The horizontal, vertical, and axial coordinates of the slot in the spatial coordinate system.

[0049] The mathematical model also needs to satisfy a series of strict constraints to ensure the engineering feasibility of the layout scheme. The first is the allocation constraint, ensuring that each tool can only be assigned to one slot, and that each slot can only hold one tool. This constraint is expressed by the following formula:

[0050]

[0051] Secondly, there are structural constraints to ensure the physical compatibility between the tool and the slot. This constraint requires that the tool type must match the slot design type, and the tool cannot be installed in slots defined as prohibited areas. This constraint is expressed by the following formula:

[0052] ;

[0053] ;

[0054] Finally, there is the static balance constraint, used to control the overall centroid offset of the tool turret, ensuring it remains within acceptable design tolerances. This constraint involves calculating the tool mass and positional moments, requiring the system to... plane and The centroidal moment offset in the axial direction does not exceed a preset threshold. This constraint is expressed by the following formula:

[0055]

[0056] In the above formula, Indicates the number is The quality of the cutting tools; Indicates the total mass of the system; express Permissible offset coefficient in the plane; express Allowable offset coefficient in the axial direction.

[0057] See attached document Figure 2 The figure illustrates the multi-helix layout and slot distribution of a circular cutterhead. In this embodiment of the invention, a continuous encoding and decoding strategy is employed to map the physical slots to the algorithm's search space during optimization calculations. This invention provides a cutter layout method for a circular cutterhead of a tunnel boring machine. The encoding and decoding steps in this method transform the discrete cutter layout problem into a continuous domain problem for solution. In this embodiment, the number of cutters and the number of slots are both set to... The geometric relationship between the slots is represented by a Cartesian coordinate system, where the center of the circle is defined as the origin, and the position of each slot corresponds to unique coordinate data.

[0058] During the encoding phase, the state of each individual gray wolf in the algorithm is represented by a position vector. This position vector is defined as a continuous vector. Its mathematical expression is:

[0059] ;

[0060] In this formula, Represents the number of cutting tools; each dimension in the vector Represents a continuous numerical value. This embodiment defines each dimension. The value range is a closed interval [0,1]. Through this encoding method, the algorithm can search and update tool layout schemes within a continuous space.

[0061] During the decoding phase, the method will use the aforementioned continuous position vectors Converted into discrete tool arrangement This arrangement This refers to the specific tool allocation scheme. The decoding process specifically performs the following operations: for the vector All values ​​in the vector are sorted in ascending order. During the sorting process, the system records the position of each value in the original vector. The original indices are used for sorting. After sorting, the sequence of these original indices constitutes the tool installation order.

[0062] This index sequence directly determines the correspondence between tools and slots. The positional order in the index sequence corresponds to the slot number, while the numerical value in the index sequence corresponds to the tool number. For example, if the sorted index sequence is (3,1,2), then this sequence represents the specific installation instructions: tool number 3 is installed in slot 1, tool number 1 is installed in slot 2, and tool number 2 is installed in slot 3. Through this decoding mechanism, any continuous vector generated by the algorithm can be mapped to a valid tool arrangement scheme, thereby ensuring that each slot is assigned one and only one tool.

[0063] See attached document Figure 1 In the fitness calculation step, after initializing and updating the position of the individual gray wolves, this embodiment of the invention uses a fitness function to quantitatively evaluate the quality of each solution. This fitness function comprehensively considers mechanical performance objectives and engineering constraints, and its mathematical expression is:

[0064] ;

[0065] in, Representing the The fitness value of an individual gray wolf; This represents the continuous position vector of the current individual gray wolf; This objective term quantifies the mechanical properties of the tool layout scheme. Specifically, it represents the sum of the eccentric force and eccentric moment generated by the tool at its current position. The values ​​of the eccentric force and eccentric moment are calculated based on the CSM (Computer-Side Module) mechanical model. The system uses the decoded tool position coordinates and substitutes them into the relevant mechanical formulas of the CSM model to calculate the force and moment components of the tool head in three spatial directions, and then uses the sum of the absolute values ​​of these components as the objective term. The calculation results.

[0066] This represents a penalty term used to handle constraint violations. Its function is to check the discrete tool arrangement generated after the decoding step. Does it violate the preset constraints? During the calculation process, the algorithm performs permutations... The algorithm performs a traversal check. Whenever a constraint violation is detected, such as a violation of an assignment constraint, structural constraint, or static equilibrium constraint, the algorithm adds a pre-set large penalty value to the objective term. In this way, infeasible solutions that violate constraints will obtain larger fitness values, thus being naturally eliminated by the algorithm mechanism during subsequent population updates, ensuring that the final optimization direction converges to the feasible solution space.

[0067] See attached document Figure 1 In the initialization step of the gray wolf population, this embodiment of the invention establishes an initial solution set through a random generation strategy, providing a search basis for subsequent iterative optimization. In the initialization phase, the size parameter of the gray wolf population is first set to... This parameter determines the number of individuals the algorithm searches in parallel during a single iteration. The dimension of the solution vector is set to... This dimension value corresponds to the total number of tools to be placed.

[0068] The initialization process uses a nested loop structure. The outer loop iterates through each individual in the population, with the loop variable incrementing from 1 to... The inner loop generates the dimensional components of each individual's position vector, and the loop variable... Increasing from 1 to In each step of the inner loop, the algorithm generates a uniformly distributed random real number within the closed interval [0,1] and assigns this random number to the first iteration. The first gray wolf individual dimensional components When the inner loop finishes, by A vector composed of dimensional components That is, it constitutes the first The initial position of each individual gray wolf. This process is repeated until an individual gray wolf is generated. Each individual, and all generated individuals Add the population to the population set PO to complete the population construction.

[0069] After the population set PO is constructed, the algorithm calls the previously defined fitness function to evaluate each gray wolf individual in the set. The algorithm evaluates and calculates the corresponding fitness values. The fitness value calculation process includes decoding a continuous vector into a discrete permutation and substituting it into a mathematical model to solve for the objective function and penalty term. After obtaining the fitness values ​​of all individuals, the algorithm sorts the individuals in the population set PO according to their fitness values. For the minimization problem in this embodiment, a smaller fitness value indicates a higher quality solution.

[0070] Based on the ranking results, the algorithm selects the top three individuals with the best fitness values ​​from the population as the leaders of the group. Specifically, the individual ranked first with the smallest fitness value is defined as α wolf, and its position vector is recorded as follows: Fitness value Let the second-ranked individual be defined as the β wolf, and record its position vector as... Fitness value Define the third-ranked individual as δ-wolf, and record its position vector as follows: Fitness value The position vectors of these three alpha wolves contain information about the currently discovered optimal region in the knowledge space, and will serve as a guide for subsequent iterations and updates of the gray wolf optimization algorithm. The core reference for wolf movement.

[0071] See attached document Figure 1 As shown in the main loop on the left, after determining the hierarchical structure of the gray wolf population, this embodiment of the invention executes the iterative update step of the gray wolf optimization algorithm. This step simulates the encirclement and hunting behavior of gray wolf packs in nature, using the positional information of the alpha, beta, and delta wolves in leadership positions to guide the position updates of other omega wolves in the population, thereby driving the solution vector to approach the global optimum in the search space.

[0072] At the start of each iteration, the algorithm first updates the convergence factor. The convergence factor controls the transition from global exploration to local development. As the number of iterations increases, the convergence factor decreases linearly from its initial value to zero. The formula for updating the convergence factor is as follows:

[0073] ;

[0074] In the above formula, Indicates the convergence factor; Indicates the current iteration number. (Symbol) This indicates the preset maximum number of iterations.

[0075] After updating the convergence factor, the algorithm calculates the coefficient vector for each individual in the population relative to the α, β, and δ wolves. These coefficient vectors are used to simulate the randomness and encirclement behavior of gray wolves approaching prey during hunting. and The calculation formula is as follows:

[0076] ;

[0077] ;

[0078] In the above formula, and Let represent a random vector whose magnitude lies within the closed interval [0,1]. The algorithm calculates the corresponding coefficient vectors for α wolf, β wolf, and δ wolf, respectively. as well as .

[0079] Subsequently, based on the coefficient vector obtained earlier, the algorithm calculates the distances between the current gray wolf and the three alpha wolves, and thereby infers the potential movement position of the current wolf after being influenced by the three alpha wolves. For the alpha wolf, the distance vector is calculated. and the corresponding positional components The formula is as follows:

[0080] ;

[0081] ;

[0082] For the β wolf, calculate the distance vector. and the corresponding positional components The formula is as follows:

[0083] ;

[0084] ;

[0085] For the δ wolf, calculate the distance vector. and the corresponding positional components The formula is as follows:

[0086] ;

[0087] ;

[0088] In the above three sets of formulas, Indicates the current number The current position vector of each individual gray wolf; , , These represent the position vectors of wolf α, wolf β, and wolf δ in the current iteration, respectively. , , These represent the weighted distances between the current individual and the three alpha wolves.

[0089] Finally, the algorithm comprehensively considers the guiding effects of α, β, and δ wolves on the current individual, and calculates the final updated position vector of the current gray wolf individual by averaging the three potential position components. The position update formula is as follows:

[0090] ;

[0091] In the above formula, This represents the updated position vector of the individual gray wolf. The algorithm will utilize... Replace the original This process completes one position update. This process is repeated for all Omega wolves in the population until the position of all individuals in the current iteration has been updated. After the update is complete, the algorithm recalculates the fitness values ​​of all individuals and updates the identities of the α, β, and δ wolves accordingly, preparing for the next iteration.

[0092] See attached document Figure 1 As shown in the branch flow on the right, this embodiment of the invention embeds a tabu search strategy into the global search framework of the gray wolf optimization algorithm to enhance the algorithm's fine-grained search capability in local regions and avoid getting trapped in local optima. This strategy is not executed in every global iteration, but rather through a specific triggering mechanism. During each iteration of the main loop, the system determines whether the current global iteration number i is a multiple of 5. If the determination result is yes, the position vector of the α wolf with the best fitness in the current population is used as the initial solution, and the tabu search process is initiated.

[0093] The tabu search algorithm first sets the current position vector of the α wolf. Assign the value to the current solution variable The algorithm then enters the inner loop of the tabu search, which performs a predetermined number of iterations. In each inner iteration, the algorithm iterates over the current solution... Neighborhood operations are performed to generate a candidate solution set. This embodiment defines two neighborhood structure operations: a swap operation and an insertion operation. A swap operation involves randomly selecting two different dimension indices in the position vector and exchanging the values ​​at those two dimensions. An insertion operation involves randomly selecting a value from one dimension in the position vector, taking it, and inserting it into another randomly selected position index, shifting the values ​​between the original and new positions sequentially.

[0094] During the generation of candidate solutions, the algorithm incorporates a tabu list mechanism to manage the search path. The system randomly performs the aforementioned swap or insertion operations to generate 10 candidate solutions. When generating each candidate solution, the system checks whether the operation that generated the solution already exists in the tabu list. If the operation is taboo, the candidate solution is discarded and regenerated until 10 non-taboo candidate solutions are obtained. Subsequently, the algorithm uses a fitness function to calculate the fitness values ​​of these 10 candidate solutions and selects the optimal solution with the smallest fitness value, denoted as . .

[0095] Having determined the optimal candidate solution for the current iteration step Then, the algorithm records the operational characteristics of the solution and adds them to the tabu list. The purpose of the tabu list is to record recently executed search actions to prevent the algorithm from repeatedly searching the same path. The algorithm then sets the current solution... Updated to The search process continues until the inner loop of the tabu search reaches its maximum number of iterations, at which point the search ends. This is the optimal solution obtained from tabu search. .

[0096] Finally, the algorithm performs a best-preference replacement step. The system compares the optimized solution obtained from tabu search. fitness value Fitness value of the original α wolf .like Less than This indicates that the tabu search has found a solution of higher quality than the current global optimum in the local neighborhood. At this point, the system will change the position vector of the α wolf. Updated to and will Updated to .like Not better than Then the original position of the α wolf remains unchanged. Through this mechanism, the present invention can improve the accuracy and robustness of the solution while ensuring global convergence.

[0097] See attached document Figure 1 As shown in the intermediate output flow, this embodiment of the invention sets up termination judgment and scheme verification steps during the collaborative operation of the gray wolf optimization algorithm and the tabu search strategy to ensure that the final output tool layout scheme meets the engineering design requirements.

[0098] The algorithm's termination decision is based on a preset iteration count threshold. At the end of each main loop, the system checks the current iteration counter. Has the maximum number of iterations been reached? If the current iteration count has not reached the maximum value, the algorithm will return to the beginning of the main loop and continue executing the next round of gray wolf position update and local tabu search. If the current iteration count has reached the maximum value, the algorithm is considered to have met the termination condition, and all search iterations will stop. At this time, the system outputs the position vector of the alpha wolf that is currently in a dominant position in the population. And based on the aforementioned decoding strategy, it is converted into a specific discrete tool layout scheme.

[0099] To ensure that the tool layout scheme output above is physically feasible and meets design specifications, this embodiment performs a final mathematical model verification step after the algorithm terminates. The verification process is based on the tool set. and candidate location set The solution is verified using the mathematical model established in the aforementioned step of constructing a mathematical optimization model for tool layout.

[0100] The system first verifies whether the proposed solution meets the optimization objective of minimizing eccentric force and eccentric moment. The system then substitutes the finally decoded tool layout scheme into the force calculation formula disclosed in the aforementioned step of constructing the mathematical optimization model for the tool layout, and recalculates the force on the tool head. , , Resultant force of eccentric forces in three directions , , and the resultant force of overturning moment , , This confirms that the objective function value is within the expected optimization range.

[0101] The system then performs a constraint compliance check on the layout scheme. Regarding allocation and structural constraint checks, the system verifies whether the scheme meets the allocation constraints in the aforementioned mathematical optimization model step for constructing the tool layout, ensuring that each slot is allocated only one tool and each tool is allocated only to one slot. It also verifies whether the structural constraints are met, ensuring that the tool type matches the slot type and that the tool is not installed in a prohibited area.

[0102] Regarding static balance constraint checks, the system utilizes the static balance constraint formula disclosed in the aforementioned step of constructing the mathematical optimization model for tool layout to calculate the total torque generated by the mass of all tools in the layout scheme. plane and The components in the axial direction. The system verifies whether these moment components satisfy the inequality constraints, that is, confirms that the centroid offset of the system in each direction does not exceed the product of the allowable offset coefficient and the total mass of the system. Only when the output layout scheme passes all the above objective function calculations and constraint condition verifications is the scheme considered valid and used as the final result to guide the actual manufacturing and assembly of the tunnel boring machine's circular cutterhead.

Claims

1. A method for cutter layout of a circular cutterhead in a tunnel boring machine, characterized in that, Includes the following steps: A mathematical model of tool layout is constructed. Combining the geometric parameters of the tool head and the cutting force data of the CSM model, the optimization objectives of minimizing eccentric force and eccentric moment and the constraints that the mathematical model of tool layout must satisfy are determined. A continuous coding strategy is used to initialize the gray wolf population. The continuous position vectors of individual gray wolves are decoded into discrete tool arrangement schemes, and the fitness value of the individual gray wolves is calculated based on the tool layout mathematical model. The gray wolf optimization algorithm is executed globally iterated. Based on the fitness value, α wolf, β wolf and δ wolf are selected. The position vectors of the α wolf, β wolf and δ wolf are used to guide the continuous position vector update of other gray wolf individuals in the gray wolf population. In the global iteration, a tabu search strategy is embedded. When the triggering condition is met, the α wolf with the best fitness is determined as the initial solution, and a local neighborhood search is performed to generate and replace the optimal solution. The process terminates when the preset maximum number of iterations is reached, outputs the tool layout scheme corresponding to the α wolf at the final moment, and uses the tool layout mathematical model to verify the feasibility of the tool layout scheme.

2. The cutter layout method for a circular cutterhead of a tunnel boring machine according to claim 1, characterized in that, The process of constructing the mathematical model of the tool layout includes: Establish a spatial rectangular coordinate system, define the tool set and the candidate slot set, and determine the spatial coordinates and installation angle of each slot in the candidate slot set; The optimization objective is set as minimizing the sum of the absolute values ​​of the resultant eccentric forces and the sum of the absolute values ​​of the resultant eccentric moments of the cutter head in the three coordinate axis directions; The constraints are defined, and the constraints include: Ensure that each slot is assigned only one tool and each tool is assigned only to one slot; Structural constraints ensure that the tool type matches the slot type and is not installed in prohibited areas; and static balance constraints limit the offset of the tool turret's centroid to within the allowable tolerance range.

3. The cutter layout method for a circular cutterhead of a tunnel boring machine according to claim 2, characterized in that, The process of determining the optimization objective of minimizing eccentric force and eccentric moment includes: Based on the normal force and lateral force in the cutting force data of the CSM model, and combined with the installation angle and decision variables of the slot, the eccentric force resultant components of the cutter head in the three coordinate axis directions are calculated respectively. Based on the normal force, the lateral force, and the spatial coordinates of the slot, and in conjunction with the decision variables, the eccentric torque components of the cutter head in the three coordinate axis directions are calculated respectively. The absolute values ​​of the resultant force components of the eccentric forces in the three directions and the absolute values ​​of the resultant moment components of the eccentric forces in the three directions are accumulated, and the sum is determined as the target value for optimization.

4. The cutter layout method for a circular cutterhead of a tunnel boring machine according to claim 1, characterized in that, The process of initializing the gray wolf population using a continuous coding strategy and decoding the continuous position vectors of individual gray wolves into a discrete tool arrangement scheme includes: Generate the continuous position vector with the same dimension as the number of tools, and limit the value range of each dimension in the continuous position vector to a preset closed interval; The values ​​in the continuous position vector are sorted in ascending order, and the original index of each value in the original vector is recorded. The original index sequence generated after sorting is determined as the installation order of the tools. The tools with corresponding numbers are sequentially assigned to the slots with corresponding numbers, thereby decoding the continuous position vector into the discrete tool arrangement scheme.

5. A method for cutter layout of a circular cutterhead for a tunnel boring machine according to claim 2, characterized in that, The process of calculating the fitness value of the individual gray wolf based on the mathematical model of the tool layout includes: Calculate the target term, which is the absolute value of the sum of the eccentric forces and eccentric moments generated by all tools in the current layout position in the discrete tool arrangement scheme; Calculate the penalty term and check whether the decoded discrete tool arrangement scheme violates the allocation constraint, the structural constraint, or the static balance constraint. If any of the constraints are violated, a preset penalty value is added to the target item, and the accumulated result is determined as the fitness value.

6. The cutter layout method for a circular cutterhead of a tunnel boring machine according to claim 1, characterized in that, The process of using the position vectors of the α wolf, the β wolf, and the δ wolf to guide the continuous position vector updates of other gray wolf individuals in the gray wolf population includes: Calculate the weighted distances of the other gray wolf individuals relative to the α wolf, the β wolf, and the δ wolf, respectively; Based on the weighted distance, calculate the three potential movement position components of the other gray wolf individuals after being affected by the α wolf, the β wolf, and the δ wolf; Calculate the arithmetic mean of the three potential movement position components, and determine the arithmetic mean as the final position vector updated for the other gray wolf individuals.

7. A method for cutter layout of a circular cutterhead for a tunnel boring machine according to claim 6, characterized in that, The process of using the position vectors of the α wolf, the β wolf, and the δ wolf to guide the continuous position vector updates of other gray wolf individuals in the gray wolf population also includes: Calculate the convergence factor, which decreases linearly from its initial value to zero as the number of iterations increases; The coefficient vector is calculated using the convergence factor and the randomly generated vector. The distances between the α wolf, the β wolf, and the δ wolf and the other individual gray wolves are then calculated using the weighted coefficient vectors to obtain the weighted distances.

8. The cutter layout method for a circular cutterhead of a tunnel boring machine according to claim 1, characterized in that, The process of performing a local neighborhood search to generate and selectively replace the optimal solution includes: Determine if the current iteration number is a multiple of a preset value. If so, initiate tabu search and determine the position vector of the α wolf as the current solution. Within the inner loop of the tabu search, a neighborhood operation is performed on the current solution to generate a candidate solution set. The neighborhood operation includes a swap operation and an insertion operation. Calculate the fitness of each candidate solution in the candidate solution set, select the optimal candidate solution that is not taboo and update the current solution, and record the operation behavior of generating the optimal candidate solution in the taboo table.

9. A method for cutter layout of a circular cutterhead for a tunnel boring machine according to claim 8, characterized in that, The process of selectively replacing the optimal solution includes: After the local neighborhood search is completed, the fitness value of the obtained local optimal solution is compared with the fitness value of the original α wolf; If the fitness value of the local optimal solution is better than that of the original α wolf, then the position vector of the original α wolf is replaced by the position vector of the local optimal solution, and the fitness value is updated. If the fitness value of the local optimal solution is not better than that of the original α wolf, then the position of the original α wolf remains unchanged.

10. A method for cutter layout of a circular cutterhead for a tunnel boring machine according to claim 2, characterized in that, The process of verifying the feasibility of the tool layout scheme using the tool layout mathematical model includes: Substitute the final output tool layout scheme into the tool layout mathematical model, recalculate the eccentric force and the eccentric moment, and verify whether it is within the optimization range; Check whether the tool layout scheme meets the allocation constraints and the structural constraints to ensure that the tool installation position is unique and the model is matched; Calculate the total torque generated by the mass of all tools in the tool layout scheme, and verify whether the centroid offset of the total torque in the XY plane and Z-axis direction satisfies the static balance constraint.

Images