A scheduling method for transformer winding production workshop

Abstract

This invention provides a scheduling method for a transformer winding production workshop, comprising: selecting the corresponding winding machine and determining parameters and decision variables based on the technological characteristics of the transformer winding stage; establishing a scheduling model for the transformer production workshop; inputting product information and equipment information, and setting antibody scale, antibody concentration threshold, and basic coefficients; setting a three-layer real number encoding and scheme based on the commonalities of equipment in high-voltage and low-voltage winding, and randomly generating an initial antibody population; calculating affinity, antibody concentration, and antibody activation degree; selecting immune cells based on antibody activation degree; antibody mutation; adding individuals with higher affinity to update the antibody population; determining whether the termination condition is met; if so, outputting the current scheduling scheme; otherwise, jumping to the calculation step. Compared with the prior art, this invention enables transformer manufacturing enterprises to formulate reasonable production strategies under given resources and constraints, thereby improving scheduling efficiency.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing technology, and in particular to a scheduling method for a transformer winding production workshop. Background Technology

[0002] Transformers, as one of the most crucial pieces of equipment in power systems, undergo a manufacturing process that includes multiple stages: high-voltage molding, winding, high-voltage lead-out, high-voltage molding assembly, high-voltage injection, high-voltage demolding, coil processing, low-voltage drying, low-voltage end sealing, low-voltage demolding, and assembly. Among these, the winding process is the core of transformer production. Winding involves winding insulated conductors onto the winding frame according to a specific pattern, number of layers, and layout. Winding methods are categorized into wire winding, strip winding, and foil winding. Wire winding refers to the traditional winding of round conductors. Foil winding uses flat conductors, while strip winding uses strip-shaped conductors, which can be flat but may be narrower and longer than foil winding. Different models can be produced based on the winding gauge, number of parallel windings, winding width, and the molds used in the production process, all undergoing the same procedures. The winding stage can be divided into high-voltage winding and low-voltage winding, primarily involving winding conductors onto the high-voltage and low-voltage windings before assembling them into a transformer. The high-voltage winding undergoes high-voltage molding, high-voltage winding, high-voltage lead-out, high-voltage assembly, high-voltage injection, high-voltage demolding, and coil processing. The low-voltage winding undergoes low-voltage winding, low-voltage drying, low-voltage end sealing, and low-voltage demolding. After high-voltage and low-voltage winding, the windings proceed through assembly, testing, painting, and final assembly. Except for the high-voltage and low-voltage winding processes, each process has its own dedicated processing machine, while the high-voltage and low-voltage winding processes share a winding machine. Therefore, the scheduling scheme for the winding stage has a significant impact on the workshop's production efficiency. Unlike typical flexible assembly line scheduling problems, this issue is complex because it requires considering the winding characteristics of different products at each stage. The complexity lies in simultaneously considering winding characteristics and machine sharing, satisfying process constraints, and ensuring that the overall production scheduling scheme is both efficient and compliant with process standards to guarantee the smooth operation of the production process.

[0003] Currently, the workshop mainly relies on manual scheduling for order placement. However, due to the large number of production bills of materials (BOMs), the limited number of machines in the winding stage, significant equipment constraints, long changeover times, and the susceptibility to errors in manual scheduling, inefficiency and potential capacity losses are inevitable. Furthermore, with the increasing demand for transformer products, production volume and order numbers will increase significantly. Frequent changes in customer demand necessitate frequent rescheduling of all related projects, posing a significant challenge to production scheduling and the supply chain. Therefore, a high-performance, high-efficiency scheduling solution is urgently needed to enable the workshop to complete orders promptly with low inventory backlog, coping with rapidly growing business demands and frequently changing customer needs.

[0004] Artificial immune algorithms are intelligent algorithms based on the biological immune system. They introduce the concept of immunity from life sciences into the engineering field, using relevant knowledge and theories to find the optimal solution through an immune response mechanism. The biological immune system is a highly complex natural immune system in the human body, capable of recognizing and eliminating invading pathogens. If cells in any cell line are activated and begin to multiply due to antigen stimulation, other cell lines that can recognize this gene type are also activated and begin to multiply. If this process continues continuously, it constitutes self-immunity, and the regulatory mechanism is achieved through the action of all lymphocytes. The basic idea of ​​artificial immune algorithms is to introduce concepts such as immune cells and antigens from the immune system into the optimization problem, seeking the optimal solution through processes such as the growth, mutation, and selection of immune cells. Due to its relatively few parameters and simple logic, artificial immune algorithms have been widely used in many fields. However, artificial immune algorithms suffer from drawbacks such as high computational cost, low convergence accuracy, and slow convergence speed, requiring optimization and improvement based on the specific scenarios and characteristics of scheduling problems. Summary of the Invention

[0005] The purpose of this invention is to address the aforementioned shortcomings of existing technologies by providing a scheduling method and system for transformer production workshops that considers winding characteristics, based on an artificial immune algorithm. Considering the characteristics of the winding process in the transformer production workshop, a scheduling model is constructed with the goal of minimizing the maximum completion time. A directed mutation artificial immune algorithm is designed, incorporating a three-layer encoding scheme, adaptive excitation degree operators, directed mutation, and prior mutation operators, among other strategies, to solve the model, thereby improving computational accuracy and efficiency.

[0006] According to one aspect of the present invention, a scheduling method for a transformer winding production workshop is provided, comprising the following steps:

[0007] S1: Based on the process characteristics of the transformer winding stage, combined with the winding wire gauge, the number of windings and the width of the winding foil, select the corresponding winding machine and set the parameters and decision variables;

[0008] S2: Based on the processing routes of all types of transformers, and with the goal of minimizing the maximum completion time of the workpiece, establish a scheduling model for the transformer production workshop.

[0009] S3: Input workpiece information, processing stage information, equipment information and workpiece processing duration, and set antibody scale, mutation probability, antibody concentration threshold, activation coefficient and antibody concentration base coefficient;

[0010] S4: Based on the equipment commonality of the two processes of high-voltage winding and low-voltage winding, and the need to select a processing winding machine according to the winding wire gauge, winding count, and winding width, a three-layer real number encoding and decoding scheme is set up to randomly generate the initial antibody population.

[0011] S5: Calculate affinity, antibody concentration, and antibody activation level;

[0012] S6: Select immune cells based on the antibody activation level, so that the best immune cells are selected for the next generation;

[0013] S7: Perform antibody mutation to achieve local search by adopting antibody mutation;

[0014] S8: Add individuals with higher fitness to update the antibody population;

[0015] S9: When the stopping criterion is met, output the current scheduling scheme; when the stopping criterion is not met, jump to step S5.

[0016] In a specific embodiment, according to step S4, in the three-layer real number encoding and decoding scheme, the first layer is the workpiece code, where the number represents the workpiece number, and repetition indicates that multiple of this type of workpiece need to be produced; the second and third layers are both machine codes, which together indicate the winding machine that can be used for the workpiece; the second layer is the winding machine type code, where winding machines are divided into multiple categories according to different winding wire gauges, maximum number of parallel windings, and winding widths, with letters representing the winding machine type; the third layer is the corresponding winding machine number, which is allocated according to the number of machines, taking into account whether the machine is already occupied, and giving priority to allocating unoccupied machines.

[0017] In a specific embodiment, according to step S4, in the three-layer real number encoding and decoding scheme, decoding is the reverse processing of encoding. Based on the encoding, the machine to which the workpiece is assigned and the processing order of the workpiece on the assigned machine are obtained. The decoding operation includes the following steps: Step S41: According to the workpiece code and machine code, the workpiece is assigned to its corresponding machine for processing. The corresponding winding machine type is determined according to the winding machine type code, and the machine to be selected is determined according to the winding machine type number in the third row; Step S42: For workpieces assigned to the same winding machine for processing, the processing order is determined sequentially according to the order in which the workpieces appear; Step S43: For each workpiece, the high-voltage follow-up stage and the low-voltage follow-up stage begin processing after the corresponding winding stage is completed. The integration stage begins processing after the high-voltage follow-up stage and the low-voltage follow-up stage are completed.

[0018] In one specific embodiment, according to step S5, the activation degree of the antibody is calculated. ,in The affinity between antigen and antibody. The concentrations of antibodies, This is the affinity coefficient. Let be the antibody concentration coefficient, where The calculation of the coefficients satisfies the expression , It is the baseline coefficient of antibody concentration. This represents the current iteration number.

[0019] In one specific embodiment, according to step S7, the workpiece code and machine code in the existing antibody are changed by combining two mutation operators. When the machine code is changed, the load of each machine is balanced. When the workpiece code is changed, the usable machine of the workpiece remains unchanged. The occurrence of infeasible solutions is avoided during the mutation process.

[0020] In one specific embodiment, for machine code mutation, a directional mutation operator is used that prioritizes mutation towards the machine with the lowest current workload. Workpieces with the same available machine set are divided into the same category, and for each category, mutation is performed according to the following rules:

[0021] If the number of available machines is even, the machine code with the highest number of workpieces is changed to the machine code with the lowest number of workpieces, the machine code with the second highest number of workpieces is changed to the machine code with the second lowest number of workpieces, and so on; if the randomly selected workpiece corresponds to a machine with a low number of workpieces, no mutation is performed.

[0022] If the number of available machines is odd, the machine code in the middle is not mutated, and the remaining machine codes are mutated according to the rule for an even number of machines.

[0023] In one specific embodiment, for the mutation of workpiece codes, a prior mutation operator is used. Before mutation, the selected workpiece codes satisfy the condition that the processing time of the workpieces in the corresponding high-voltage winding or low-voltage winding processes is different before and after mutation; if the processing time is the same, no mutation is performed.

[0024] The effectiveness of this invention lies in its ability to address the actual scheduling needs of a transformer workshop, taking into account the technological characteristics of the winding stage (i.e., high-voltage and low-voltage winding stages sharing multiple parallel machines, and the need to select suitable winding machines based on characteristics such as winding gauge, number of parallel windings, and winding foil width). A mixed-integer programming model is constructed with the goal of minimizing the maximum completion time. To avoid the shortcomings of artificial immune algorithms, such as high computational cost, low convergence accuracy, and slow convergence speed, a directed mutation artificial immune algorithm (AIADM) is proposed based on the artificial immune algorithm. According to the characteristics of this special scheduling problem, a three-layer encoding and decoding scheme is designed, an adaptive excitation degree operator is proposed, and corresponding directed mutation and prior mutation operators are designed for the workpiece code and machine code respectively during the antibody mutation process. This satisfies the machine constraints of the workpiece and avoids infeasible solutions. This invention enables transformer manufacturers or factories to formulate reasonable production strategies and improve scheduling efficiency under given resources and constraints. Attached Figure Description

[0025] After reading the detailed embodiments of the present invention with reference to the accompanying drawings, the reader will gain a clearer understanding of various aspects of the present invention.

[0026] Figure 1 A flowchart illustrating a scheduling method for a transformer winding production workshop according to one aspect of the present invention is shown.

[0027] Figure 2 This is a three-layer coding scheme provided in an illustrative embodiment of this application;

[0028] Figure 3 This is a schematic diagram of a machine code mutation operator provided in an exemplary embodiment of this application;

[0029] Figure 4 The following is shown: Figure 1 The scheduling method of this application is shown in the schematic diagram of the convergence curve of a 5×10 example; and

[0030] Figure 5 The following is shown: Figure 1 The diagram shows the convergence curve of the scheduling method of this application for a 6×15 example. Detailed Implementation

[0031] To make the technical content disclosed in this application more detailed and complete, reference can be made to the accompanying drawings and the various specific embodiments of the present invention described below, in which the same reference numerals represent the same or similar components. However, those skilled in the art should understand that the embodiments provided below are not intended to limit the scope of the present invention. Furthermore, the drawings are for illustrative purposes only and are not drawn to their original dimensions.

[0032] The specific embodiments of various aspects of the present invention will now be described in further detail with reference to the accompanying drawings.

[0033] Figure 1This diagram illustrates a scheduling method for a transformer winding production workshop according to one aspect of the present invention. Specifically, the scheduling method is implemented through steps S1 to S9. More specifically, firstly, based on the technological characteristics of the transformer winding stage, combined with the winding gauge, the number of parallel windings, and the winding foil width, a corresponding winding machine is selected, and parameters and decision variables are set. In step S2, based on the processing routes of all types of transformers, a scheduling model for the transformer production workshop is established with the goal of minimizing the maximum completion time of the workpiece. In step S3, workpiece information, processing stage information, equipment information, and the processing duration of the workpiece are input, and antibody scale, mutation probability, antibody concentration threshold, excitation coefficient, and antibody concentration base coefficient are set. In step S4, based on the equipment commonality between high-voltage and low-voltage winding processes, and the need to select a suitable winding machine based on the winding gauge, the number of parallel windings, and the winding width, a three-layer real number encoding and decoding scheme is set, and an initial antibody population is randomly generated. In step S5, affinity, antibody concentration, and antibody excitation are calculated. Step S6: Select immune cells based on antibody activation level, choosing superior immune cells for the next generation. Step S7: Perform antibody mutation to achieve local search. Step S8: Add individuals with higher fitness, updating the initial antibody population. Finally, determine the termination condition. If the stopping criterion is met, output the current scheduling plan; otherwise, jump to step S5.

[0034] In this illustrative embodiment, step S1 involves considering the actual scheduling needs of the transformer workshop, the technological characteristics of the winding stage, and the need to select a suitable winding machine based on factors such as the winding wire gauge, the number of parallel windings, and the width of the winding foil. Based on these considerations, parameters and decision variables are determined.

[0035] Step S2: Consider the processing routes of all types of transformers and establish a transformer production workshop scheduling model that takes into account winding characteristics, with the goal of minimizing the maximum completion time.

[0036] Set the following parameters and variables in sequence to establish the objective function to be optimized and the constraints:

[0037] (1) Parameters

[0038] : A set of workpieces, with index , ,in For the number of workpieces;

[0039] : The stage set, whose index is, ,in This refers to the quantity at each processing stage.

[0040] : Winding stage, ;

[0041] Integration phase ;

[0042] The final process in the high-pressure stage. ;

[0043] The final process in the low-pressure stage. ;

[0044] : No. Available machines for each stage Its index is , ,in for Quantity;

[0045] The set of available winding machines corresponding to the winding stage. ;

[0046] Available winding machines for the winding stage The winding gauge, its index is , ;Regulation The time indicates that the winding gauge is a winding gauge, in which The number of different types of wire gauges used by the winding machine during the winding stage;

[0047] Available winding machines for the winding stage The number of windings and the number of windings, its index is . , ,in The number and variety of windings produced by the winding machine during the winding stage;

[0048] Available winding machines for the winding stage The winding width, its index is , ,in The types and quantities of winding widths for the winding machine during the winding stage;

[0049] , , : respectively workpiece The winding gauge, number of windings, and winding width;

[0050] : in the Each processing stage, workpiece Processing time;

[0051] During the winding stage, the workpiece Processing time;

[0052] (2) Variables

[0053] : Apart from the winding stage, if the workpiece In the Machines at each stage If the value is above, it is set to 1; otherwise, it is set to 0.

[0054] : During the winding stage, if the workpiece The winding gauge is On the machine, it is processed as 1, otherwise as 0;

[0055] : During the winding stage, if the workpiece After winding and the number of windings is On the machine, it is processed as 1, otherwise as 0;

[0056] : During the winding stage, if the workpiece With a winding width of On the machine, it is processed as 1, otherwise as 0;

[0057] Except for the winding stage, in the first stage Each processing stage, workpiece Start processing time;

[0058] Except for the winding stage, in the first stage Each processing stage, workpiece The completion time of the processing;

[0059] Workpiece during the winding stage Start processing time;

[0060] Workpiece during the winding stage The completion time of the processing;

[0061] Maximum completion time of the workpiece;

[0062] Winding stage Winding machine can be used The processing sequence of the workpiece on the surface. , For winding machine The number of workpieces to be processed.

[0063] (3) Objective function

[0064] Considering the early completion of machining for multiple workpieces, with the objective of minimizing the maximum completion time, this can be expressed as:

[0065] (1)

[0066] (4) Constraints

[0067] workpiece During the processing stage (Except for the winding stage) it can only be processed by one machine.

[0068] (2)

[0069] machine During the processing stage (Except for the winding stage) a maximum of one workpiece can be processed.

[0070] (3)

[0071] Except for the winding stage, the workpiece The completion time of a process is equal to the start time of the process plus the processing duration.

[0072] (4)

[0073] During the winding stage workpiece The completion time of a process is equal to the start time of the process plus the processing duration.

[0074] (5)

[0075] The sequence between processing stages is fixed, except for the winding stage and the integration stage, the next processing stage... The start time of processing is the end time of the previous stage.

[0076] (6)

[0077] workpiece During the winding stage The start time of processing is the latest time between the end time of the previous stage and the completion time of the previous workpiece on the winding machine.

[0078] (7)

[0079] workpiece During the integration phase The start time of processing is the latest time between the end time of the high-pressure stage and the end time of the low-pressure stage.

[0080] (8)

[0081] workpiece During the winding stage Only winding machines with the same wire gauge as its own can be selected for processing.

[0082] (9)

[0083] workpiece During the winding stage Only winding machines with a winding count greater than or equal to the number of windings it can be used for processing.

[0084] (10)

[0085] workpiece During the winding stage In this case, if the wire gauge is tape winding or foil winding, only a winding machine with a width greater than or equal to its own winding width can be selected for processing.

[0086] (11)

[0087] Step S3: Input necessary information such as order details and workpiece details. Table 1 shows partial data for order information, workpiece information, etc. The order number is a unique identifier, and the workpiece ID represents the transformer model. An order can contain multiple products, each with one or more quantities. The delivery date indicates the scheduled delivery date of the order, and production must be completed before that date.

[0088] Table 1

[0089]

[0090] Table 2 shows that the five winding machines belong to four categories, and lists their corresponding winding machine type codes and the number of winding machines.

[0091] Table 2

[0092]

[0093] Table 3 lists the types of winding machines available for the high-pressure and low-pressure winding stages of the workpieces in Table 1.

[0094] Table 3

[0095]

[0096] Step S4: Based on the shared equipment between the high-voltage and low-voltage winding processes, and considering the need to select a suitable winding machine according to the winding gauge, number of windings, and winding width, a three-layer real-number coding scheme was designed to randomly generate the initial antibody population; the specific coding details are as follows... Figure 2 As shown in the diagram. The first layer is the workpiece code, with numbers representing different workpieces. The second and third layers are machine codes. The second layer is the winding machine type code, which is divided into multiple categories according to different winding gauges, maximum number of parallel windings, and winding widths. The letters represent the winding machine types; here, there are two categories: G and H. The third layer is the winding machine number for the corresponding type, indicating which winding machine is which in each category.

[0097] Step S5: Calculate affinity, antibody concentration, and antibody activation degree. First, calculate the affinity between the antibody and the antigen, as well as the similarity between antibodies. Then, calculate the antibody activation degree. Antibody activation degree refers to the overall ability of an antibody in an antibody group to respond to the antigen and be activated by other antibodies. Generally, antibodies with high affinity and low concentration will have a higher activation degree.

[0098] In steps S6 and S7, antibody selection and antibody mutation are performed sequentially. Immune cells are selected based on the antibody's activation level. Regarding antibody mutation, considering the characteristics of this scheduling problem, a combination of two mutation operators is proposed to modify the workpiece code and machine code in the existing antibodies. When changing the machine code, the load level of each machine is balanced; when changing the workpiece code, the available machines for the workpiece remain unchanged. Specifically, for machine code mutation, a directional mutation operator is designed that prioritizes mutation to the machine with the lowest current load, i.e., workpieces with the same available machine set are grouped into the same category. For each category, mutation is performed according to the following rules:

[0099] (1) If the number of available machines is even, then the machine code with the highest workpiece load is changed to the machine code with the lowest workpiece load, the machine code with the second highest workpiece load is changed to the machine code with the second lowest workpiece load, and so on. For example... Figure 3 As shown, the usable machines are G-1, G-2, H-1, and H-2. The workpiece loads of the machines, ordered from highest to lowest, are G-2, G-1, H-2, and H-1. If randomly selected workpieces p1 and p2 are processed on machines G-2 and G-1 respectively, and denoted as m1 and m2 respectively, then m1 should mutate into H-1, and m2 should mutate into H-2, thus generating new antibodies. If the machine load corresponding to the randomly selected workpiece is low, no mutation occurs.

[0100] (2) If the number of available machines is odd, the machine code in the middle is not mutated, and the remaining machine codes are mutated according to the rules in (1).

[0101] A priori mutation operator was designed for workpiece code mutation. Since multiple workpieces in the winding process have the same processing time, directly performing mutation increases the probability that the maximum completion time will be the same before and after mutation, reducing the algorithm's efficiency. Therefore, before mutation, the selected workpiece codes should satisfy the condition that the processing time of the workpiece in the corresponding high-voltage or low-voltage winding process is different before and after mutation. If they are the same, mutation is not performed.

[0102] Step S8: Add individuals with higher fitness and update the initial antibody population.

[0103] Step S9: Determine the termination condition. If the stopping criterion is met, output the current optimal solution; otherwise, jump to step S5 to continue scheduling optimization.

[0104] To test the performance of the Artificial Immune Algorithm with Directed Mutation (AIADM), simulations were performed on scheduling problem test cases of different scales. For example, the Artificial Immune Algorithm (AIA), Flexible Flowshop Scheduling Genetic Algorithm (FFSGA), Artificial Bee Colony Algorithm (ABC), and Improved Cuckoo Algorithm (ICS) were selected as comparison algorithms, with each experiment running independently 10 times. To ensure that the initial solution was feasible, the three-layer encoding method in AIADM was used, and the simulation results were analyzed for the following metrics: Relative Standard Deviation (RSD), Best Relative Percentage Deviation (BRPD), and Mean Relative Percentage Deviation (MRPD).

[0105] Figure 4 The following is shown: Figure 1 The scheduling method of this application is shown in the schematic diagram of the convergence curve of a 5×10 example. Figure 5 The following is shown: Figure 1 The diagram shows the convergence curve of the scheduling method of this application for a 6×15 example. Specifically, Figure 4 and 5The convergence curves for different scales of computational examples show that the AIADM algorithm converges faster. In the 5×10 example, all algorithms except AIA achieved optimal values. However, in the 6×15 example, the AIADM algorithm performed worse than AIA. For medium and large-scale examples, the AIADM algorithm outperformed other algorithms. In medium-scale examples, the AIADM algorithm showed a rapid initial descent rate, with a continued downward trend in the mid-term. In large-scale examples, the ABC algorithm performed poorly, easily getting trapped in local optima, while the AIADM algorithm, although not fully converging within the specified running time, yielded a significantly smaller solution than other algorithms. Overall, the AIADM algorithm performed better because the adaptive excitation degree operator increased the penalty for similar solutions, improving its ability to escape local optima; the directional mutation operator quickly balanced the machine load; and the prior mutation operator improved the efficiency of workpiece code mutation.

[0106] Compared to existing technologies, this application addresses the actual scheduling needs of transformer production workshops, considering the characteristics of the core stages of transformer production processes. Specifically, the high-voltage and low-voltage winding stages share multiple parallel machines, and the winding machine selection must be based on factors such as winding gauge, number of parallel windings, and winding foil width. A mixed-integer programming model is constructed with the goal of minimizing the maximum completion time. To overcome the shortcomings of low convergence accuracy and slow convergence speed of basic artificial immune algorithms, a directed mutation artificial immune algorithm is proposed. Based on the characteristics of this specific scheduling problem, a three-layer encoding and decoding scheme is designed, an adaptive excitation degree operator is proposed, and corresponding directed mutation and prior mutation operators are designed for the workpiece code and machine code respectively during the antibody mutation process. The algorithm's advancement and stability are verified by comparing simulation results with other algorithms. Compared to existing technologies, the scheduling method of this invention enables transformer manufacturing enterprises or factories to formulate reasonable production strategies under given resources and constraints, thereby improving scheduling efficiency.

[0107] Specific embodiments of the present invention have been described above with reference to the accompanying drawings. However, those skilled in the art will understand that various modifications and substitutions can be made to the specific embodiments of the present invention without departing from the spirit and scope of the invention. All such modifications and substitutions fall within the scope defined by the claims of the present invention.

Claims

1. A scheduling method for a transformer winding production workshop, characterized in that, The scheduling method includes the following steps: S1: Based on the process characteristics of the transformer winding stage, combined with the winding wire gauge, the number of windings and the width of the winding foil, select the corresponding winding machine and set the parameters and decision variables; S2: Based on the processing routes of all types of transformers, and with the goal of minimizing the maximum completion time of the workpiece, establish a scheduling model for the transformer production workshop. S3: Input workpiece information, processing stage information, equipment information and workpiece processing duration, and set antibody scale, mutation probability, antibody concentration threshold, activation coefficient and antibody concentration base coefficient; S4: Based on the equipment commonality of the two processes of high-voltage winding and low-voltage winding, and the need to select a processing winding machine according to the winding wire gauge, winding count, and winding width, a three-layer real number encoding and decoding scheme is set up to randomly generate the initial antibody population. S5: Calculate affinity, antibody concentration, and antibody activation level; S6: Based on the antibody activation level, select immune cells so that the superior immune cells are selected as the next generation; S7: Perform antibody mutation to achieve local search by adopting antibody mutation; S8: Add individuals with higher fitness to update the antibody population; S9: Determine if the termination condition is met. If it is met, output the current scheduling scheme; otherwise, jump to step S5.

2. The scheduling method according to claim 1, characterized in that, According to step S4, in the three-layer real number encoding and decoding scheme... The first layer is the workpiece code, where the numbers represent the workpiece number. Repeated occurrences indicate that multiple of this type of workpiece need to be produced. The second and third layers are both machine codes, which together indicate the winding machine that the workpiece can be used with. The second layer is the winding machine type code. Winding machines are divided into multiple categories according to different winding wire gauges, maximum number of parallel windings, and winding widths. The letters represent the winding machine types. The third layer consists of the corresponding winding machine number, which is allocated according to the number of machines. When allocating, it is taken into account whether the machine is already occupied, and unoccupied machines are given priority.

3. The scheduling method according to claim 2, characterized in that, According to step S4, in the three-layer real number encoding and decoding scheme, decoding is the reverse processing of encoding. Based on the encoding, the machine to which the workpiece is assigned and the processing order of the workpiece on the assigned machine are obtained. The decoding operation includes the following steps: Step S41: Based on the workpiece code and machine code, assign the workpiece to its corresponding machine for processing. Determine the corresponding winding machine type based on the winding machine type code, and determine which machine in that type to select based on the winding machine type number in the third row. Step S42: For workpieces assigned to the same winding machine, determine the processing sequence according to the order in which the workpieces appear; Step S43: For each workpiece, the high-pressure follow-up stage and the low-pressure follow-up stage begin processing after the corresponding winding stage is completed, and the integration stage begins processing after the high-pressure follow-up stage and the low-pressure follow-up stage are completed.

4. The scheduling method according to claim 1, characterized in that, Calculate the activation level of the antibody according to step S5. ,in The affinity between antigen and antibody. The concentrations of antibodies, This is the affinity coefficient. Let be the antibody concentration coefficient, where The calculation of the coefficients satisfies the expression , It is the baseline coefficient of antibody concentration. This represents the current iteration number.

5. The scheduling method according to claim 1, characterized in that, According to step S7, the workpiece code and machine code in the existing antibody are changed by combining two mutation operators. When the machine code is changed, the load of each machine is balanced. When the workpiece code is changed, the usable machine of the workpiece remains unchanged. The occurrence of infeasible solutions is avoided during the mutation process.

6. The scheduling method according to claim 5, characterized in that, For machine code mutation, a directional mutation operator is used that prioritizes mutation towards the machine with the lowest current workload. Workpieces with the same available machine set are grouped into the same category, and mutation is performed according to the following rules: If the number of available machines is even, the machine code with the highest number of workpieces is changed to the machine code with the lowest number of workpieces, the machine code with the second highest number of workpieces is changed to the machine code with the second lowest number of workpieces, and so on; if the randomly selected workpiece corresponds to a machine with a low number of workpieces, no mutation is performed. If the number of available machines is odd, the machine code in the middle is not mutated, and the remaining machine codes are mutated according to the rule for an even number of machines.

7. The scheduling method according to claim 5, characterized in that, For workpiece code mutation, a prior mutation operator is used. Before mutation, the selected workpiece code satisfies that the processing time of the workpiece in the corresponding high-voltage winding or low-voltage winding process is different before and after mutation; if the processing time is the same, mutation is not performed.

Images