Photovoltaic module production control method and device based on swarm intelligence algorithm

Abstract

The invention provides a control method and control equipment for photovoltaic module production based on a swarm intelligence algorithm, and relates to the technical field of intelligent manufacturing of photovoltaic modules. The method comprises the following steps: acquiring a discrete event simulation model and a multi-target optimization model of a photovoltaic module production line; the multi-objective optimization model takes cache layout, cache capacity and equipment speed as joint decision variables; the discrete event simulation model takes equipment operation data in a set time as input; the output yield, the quantity of products in process and the production energy consumption are taken as output parameters; updating the discrete event simulation model to perform simulation operation by using a joint decision variable output by the multi-objective optimization model to obtain a production efficiency evaluation index; and when the production efficiency evaluation index meets the optimization condition, controlling the operation of the photovoltaic module production line by using the corresponding joint decision variable. According to the invention, global optimization control of cache layout and speed dynamic optimization can be realized at the same time.

Description

Technical Field

[0001] The present invention relates to the field of intelligent manufacturing of photovoltaic modules, and particularly to a control method and control device for photovoltaic module production based on swarm intelligence algorithms. Background Art

[0002] In recent years, with the continuous growth of the global demand for green energy, the production scale and automation level of photovoltaic modules have been continuously improved. The photovoltaic module production line includes multiple key processes, and buffer areas need to be set between different processes to temporarily store work in progress (WIP). Whether the buffer area capacity is too large or too small will have an adverse impact on the stable operation of the production line. Reasonably setting the buffer area can cope with production fluctuations. In addition, the speed adjustment range of key process equipment affects production capacity and energy consumption, and existing optimization methods are difficult to solve multi-objective optimization problems. Therefore, a global optimization control scheme that simultaneously realizes cache layout and speed dynamic optimization is needed. Summary of the Invention

[0003] Embodiments of the present invention provide a control method and control device for photovoltaic module production based on swarm intelligence algorithms to solve the problem of how to simultaneously achieve global optimization control of cache layout and speed dynamic optimization.

[0004] In a first aspect, embodiments of the present invention provide a control method for photovoltaic module production based on swarm intelligence algorithms, including:

[0005] Obtaining a discrete event simulation model and a multi-objective optimization model of a photovoltaic module production line; wherein, the multi-objective optimization model includes a multi-objective function and joint constraint conditions; the multi-objective function takes cache layout, cache capacity, and equipment speed as joint decision variables; the joint constraint conditions include process constraints, space constraints, and beat matching constraints; the discrete event simulation model takes equipment operation data within a set time as input; and takes output output, work-in-progress quantity, and production energy consumption as output parameters;

[0006] Updating the discrete event simulation model with the joint decision variables output by the multi-objective optimization model for simulation operation to obtain a production efficiency evaluation index;

[0007] When the production efficiency evaluation index meets the optimization conditions, controlling the operation of the photovoltaic module production line with the corresponding joint decision variables; otherwise, applying a swarm intelligence algorithm to optimize the multi-objective optimization model to obtain new joint decision variables until the production efficiency evaluation index meets the optimization conditions.

[0008] In a possible implementation manner, the applying a swarm intelligence algorithm to optimize the multi-objective optimization model to obtain new joint decision variables includes:

[0009] Apply the swarm intelligence algorithm to solve the multi-objective function and generate a set of initial solutions;

[0010] Use the discrete event simulation model to simulate and run the initial solutions, obtain the corresponding multi-objective metrics and calculate the fitness;

[0011] According to the fitness evaluation results, use the iterative mechanism of the swarm intelligence algorithm to update the solutions until the convergence condition or the preset number of iterations is reached, and output the optimal solution;

[0012] Determine the corresponding joint decision variables according to the optimal solution.

[0013] In a possible implementation, the swarm intelligence algorithm is a genetic algorithm.

[0014] In a possible implementation, the cache layout uses binary encoding, the cache capacity uses integer encoding, and the device speed uses real number encoding or discrete gear encoding.

[0015] In a possible implementation, the device operation data within the set time includes the device failure frequency, the pipeline transmission speed, the switching time, and the device recovery time.

[0016] In a possible implementation, the set time is determined according to the actual order demand, the failure data, and the personnel changes;

[0017] Among them, the large actual order demand, the amount of failure data, and the frequency of personnel changes are all inversely proportional to the set time.

[0018] In a possible implementation, before obtaining the discrete event simulation model and the multi-objective optimization model of the photovoltaic module production line, it further includes:

[0019] Obtain the production parameters of each process of the photovoltaic module production line and the change range of the production parameters; among them, the production parameters include: equipment type, material transfer path, production beat, pipeline transmission speed, switching time, personnel configuration, number of buffer areas, buffer capacity, equipment failure rate, and equipment recovery time; the equipment type includes two or more of: string welding machine, laminating machine, layer press, typesetting machine, framing machine, cutting machine, conveyor, and curing machine;

[0020] Establish the discrete event simulation model according to the production parameters and the change range of the production parameters, and configure the candidate positions of the buffer areas and the upper and lower limits of the buffer capacity in the discrete event simulation model.

[0021] In a possible implementation, the upper limit of the capacity of each buffer area is less than or equal to 50 pieces.

[0022] In a possible implementation, when the production efficiency evaluation index meets the optimization condition, the optimal solution corresponding to the multi-objective optimization model is sent to the Manufacturing Execution System (MES), and MES issues a speed regulation instruction and a cache capacity control instruction to the equipment and / or the management terminal of the photovoltaic module production line.

[0023] In a second aspect, an embodiment of the present invention provides a control device for photovoltaic module production based on a swarm intelligence algorithm, including:

[0024] An acquisition module, configured to acquire a discrete event simulation model and a multi-objective optimization model of the photovoltaic module production line; wherein, the multi-objective optimization model includes a multi-objective function and joint constraint conditions; the multi-objective function takes cache layout, cache capacity, and equipment speed as joint decision variables; the joint constraint conditions include process constraints, space constraints, and cycle time matching constraints; the discrete event simulation model takes the equipment operation data within a set time as input; and takes the output output, the number of work-in-process, and the production energy consumption as output parameters;

[0025] A control module, configured to update the discrete event simulation model for simulation operation with the joint decision variables output by the multi-objective optimization model, and obtain a production efficiency evaluation index;

[0026] When the production efficiency evaluation index meets the optimization condition, control the operation of the photovoltaic module production line with the corresponding joint decision variables; otherwise, apply a swarm intelligence algorithm to optimize the multi-objective optimization model to obtain new joint decision variables until the production efficiency evaluation index meets the optimization condition.

[0027] In a third aspect, an embodiment of the present invention provides a control device for photovoltaic module production based on a swarm intelligence algorithm, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the method in the first aspect or any possible implementation manner of the first aspect is implemented.

[0028] In the embodiments of the present invention, a multi-objective optimization-discrete event simulation dual-engine collaborative mechanism is constructed to achieve global dynamic joint optimization control of the buffer layout, buffer capacity configuration, and equipment speed in the photovoltaic module production line. Based on the swarm intelligence algorithm, global optimization is performed on the multi-objective functions (maximizing output, minimizing work-in-process, and energy consumption) to generate a set of joint decision variables including buffer activation positions, capacity thresholds, and equipment speed regulation parameters. The discrete event simulation model is used to simulate the operation of the actual production line, dynamically verify the actual effectiveness of the solution set, and feedback the production efficiency indicators. When the indicators do not meet the optimization conditions, the algorithm is triggered to iteratively update the solution set until a global optimal solution that simultaneously satisfies process constraints, space limitations, and cycle time matching is output. This application breaks through the limitations of traditional single-variable static optimization, improves the utilization rate of the buffer area, enhances the equipment collaboration efficiency, and in the scenarios of order fluctuations or sudden failures, shortens the production interruption time through dynamic optimization, realizing the three-in-one adaptive control of "layout-speed-capacity", and providing a comprehensive and stable control method for the photovoltaic module production line. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 is a schematic diagram of the process flow of a photovoltaic module production line provided by an embodiment of the present invention;

[0030] Figure 2 is an example diagram of the relationship between candidate buffer positions and buffer capacity provided by an embodiment of the present invention;

[0031] Figure 3 is a flowchart of the implementation of a control method for photovoltaic module production based on the swarm intelligence algorithm provided by an embodiment of the present invention;

[0032] Figure 4 is a schematic diagram of the Pareto front of the multi-objective optimization results shown in an embodiment of the present application;

[0033] Figure 5 is a schematic diagram of the optimization process of the multi-objective optimization model by the swarm intelligence algorithm provided by an embodiment of the present application;

[0034] Figure 6 is a schematic diagram of the discrete event simulation model of a photovoltaic module production line provided by an embodiment of the present application;

[0035] Figure 7 is a flowchart of model establishment and simulation provided by an embodiment of the present invention;

[0036] Figure 8 is a schematic diagram of the structure of a control device for photovoltaic module production based on the swarm intelligence algorithm provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] Typical photovoltaic module production lines include multiple key processes, such as cell string soldering, overlapping soldering, lamination, framing, testing, and packaging. These processes not only have different production rhythms and failure modes, but also buffer areas need to be set between different processes to temporarily store work-in-progress. Reasonable buffer layout and capacity design can play a buffering role when fluctuations occur in upstream and downstream processes, ensuring the smooth operation of the production line.

[0038] However, if the buffer capacity is too large, it will lead to the backlog of work-in-progress, capital occupation, and site waste; while if the buffer capacity is too small, it will not be able to effectively cope with the production capacity imbalance caused by equipment failures, process switching, or order batch changes. In addition, some key process equipment (such as string soldering machines, laminators, film pasting machines, or automatic conveyor belts, etc.) has a certain speed adjustment range, and its speed change will directly affect production capacity and energy consumption. Existing optimization methods often rely on traditional empirical management or linear models and cannot effectively solve multi-objective optimization problems, especially the balance among multiple objectives such as equipment speed, buffer layout, energy consumption, and output. Therefore, how to optimize under multi-objective and multi-constraint conditions has become the key to improving the production efficiency and flexibility of photovoltaic modules.

[0039] Swarm intelligence algorithms (such as genetic algorithms, particle swarm algorithms, ant algorithms, firefly algorithms, etc.) are widely used in complex optimization problems such as production scheduling, path planning, and parameter optimization, and have the advantages of distributed parallel search and jumping out of local optima. Combining swarm intelligence algorithms with discrete event simulation of the production line can iteratively solve multi-objective and multi-constraint optimization problems, dynamically generate feasible combinations of "buffer layout, buffer capacity, and equipment speed", and achieve global optimization of the photovoltaic module production line.

[0040] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0041] Figure 1 FIG. is a schematic diagram of the process flow of a photovoltaic module production line provided by an embodiment of the present invention. The objective of the present application is to optimize the conversion process between various processes in the process flow, ensure the smooth operation of the production line, and adjust the equipment speed of each process to ensure the production efficiency and energy consumption of the production line. As Figure 1 shown, the process flow includes soldering, EL detection, cell layout, overlapping soldering, placing the backsheet, appearance detection, lamination, curing, and safety detection. Figure 1 FIG. is an exemplary process flow of a photovoltaic module production line. In other embodiments, it also includes processes such as framing, testing, and packaging. In the actual production process, according to the photovoltaic module production standards or process improvements, the processes of the photovoltaic module production line include any two or more of the foregoing processes.

[0042] In addition, the cache layout mainly refers to the setting position of the buffer area. During the implementation of the solution of this application, a buffer area can be set for each process. Or, a buffer area is only set for some important processes. For example Figure 2 As shown, a buffer area 1 is set for the cell string soldering process, a buffer area 2 is set for the stack soldering, a buffer area 3 is set for the lamination, and no buffer area is set between the processes of framing, testing and packaging.

[0043] Figure 3 FIG. is a flowchart of the implementation of a control method for photovoltaic module production based on a swarm intelligence algorithm provided by an embodiment of the present invention. As shown in Figure 3 the following steps are included:

[0044] S301, obtaining a discrete event simulation model and a multi-objective optimization model of a photovoltaic module production line; wherein, the multi-objective optimization model includes a multi-objective function and joint constraint conditions; the multi-objective function takes the cache layout, cache capacity and equipment speed as joint decision variables; the joint constraint conditions include process constraints, space constraints and beat matching constraints; the discrete event simulation model takes the equipment operation data within a set time as input; and takes the output output (Throughput), the number of work-in-process and production energy consumption as output parameters.

[0045] The execution subject of the control method for photovoltaic module production based on the swarm intelligence algorithm provided by the embodiment of this application can be a device with data processing functions such as a server, a processor, a microprocessor, etc. In the actual implementation process, the specific implementation method of the execution subject can be selected according to actual needs, and this embodiment does not make special restrictions on this, as long as it is a device with data processing functions.

[0046] Among them, as a dynamic process simulator, the discrete event simulation model focuses on high-fidelity restoration of timing events and random disturbances during the operation of the photovoltaic module production line. It takes equipment operation data (such as failure frequency, transmission speed) as input, and simulates processes such as material flow and equipment state switching through an event-driven mechanism, and quantitatively outputs dynamic indicators such as output, the number of work-in-process and production energy consumption, which are mainly used to verify the feasibility of the production line configuration, predict bottleneck links and evaluate the impact of sudden failures.

[0047] In the specific implementation process, the discrete event simulation model takes the equipment operation data within a set time as input. In order to ensure that the discrete event simulation model is updated regularly to adapt to the real-time needs and optimization of the photovoltaic module production line.

[0048] In a possible implementation manner, the equipment operation data within a set time includes equipment failure frequency, pipeline transmission speed, switching time and equipment recovery time.

[0049] Among them, the equipment failure frequency in the equipment operation data reflects the equipment reliability, directly affects the production continuity, and leads to a decrease in output, accumulation of work-in-process, and an increase in energy consumption; the pipeline transmission speed determines the production rhythm and needs to be matched with the equipment processing speed to balance the quantity of work-in-process and energy consumption; the switching time reflects the flexibility of the production line and affects the changeover efficiency and buffer requirements; the equipment recovery time determines the resumption speed after a failure and affects the downtime loss and production scheduling. The four types of data jointly quantify the impacts of different combinations on output, work-in-process quantity, and energy consumption through discrete event simulation, and can guide the multi-objective optimization model to find the optimal solution.

[0050] The multi-objective optimization model, as a static decision-making optimizer, focuses on finding the optimal combination of buffer layout, buffer capacity allocation, and equipment speed adjustment through mathematical modeling under multiple conditions such as process constraints, space limitations, and rhythm matching. It takes decision variable encoding and constraint conditions as inputs, generates a candidate solution set through swarm intelligence algorithms, and outputs optimization results that can be mapped to production line control parameters.

[0051] The discrete event simulation model and the multi-objective optimization model complement each other. The simulation model provides the calculation of the objective function and the verification of constraint feasibility for the optimization model, while the optimization model iteratively generates decision-making schemes to be evaluated for the simulation model. Finally, the coordinated improvement of production line efficiency is achieved through the "dynamic simulation - static optimization" closed-loop.

[0052] S302. Update the discrete event simulation model with the joint decision variables output by the multi-objective optimization model for simulation operation to obtain production efficiency evaluation indicators.

[0053] Updating the discrete event simulation model with the joint decision variables output by the multi-objective optimization model for simulation operation aims to conduct closed-loop verification and achieve the deep coupling of static mathematical optimization and dynamic production simulation. The multi-objective optimization model provides the global optimal solution (such as buffer layout, equipment speed combination), and the discrete event simulation model verifies the actual effectiveness of the solution by simulating random events (equipment failures, rhythm fluctuations) in the real production line, exposes the defects of the theoretical solution in the dynamic environment (such as work-in-process accumulation, reduced output, increased production energy consumption caused by rhythm mismatch), and then drives the optimization model to iteratively modify the constraint conditions and objective weights, and finally outputs a solution with both theoretical optimality and engineering feasibility.

[0054] S303. When the production efficiency evaluation indicators meet the optimization conditions, control the operation of the photovoltaic module production line with the corresponding joint decision variables; otherwise, apply swarm intelligence algorithms to optimize the multi-objective optimization model to obtain new joint decision variables until the production efficiency evaluation indicators meet the optimization conditions.

[0055] By establishing a real-time closed-loop feedback mechanism of "evaluation - optimization - execution", the dynamic control ability of the photovoltaic module production line is significantly improved. When the production efficiency indicators (such as output, work-in-progress, energy consumption) generated by the simulation reach the preset optimization threshold, the current joint decision variables (buffer layout / capacity / equipment speed combination) are directly used to drive the operation of the production line. If the standard is not met, a swarm intelligence algorithm is triggered to perform dynamic optimization on the multi-objective optimization model, and new solutions are iteratively generated and verified based on the simulation feedback data until the optimal parameters that meet the actual production requirements are output. This mechanism breaks through the limitation of the disconnection between traditional static optimization and the dynamic characteristics of actual production. Through continuous adaptive parameter tuning, it realizes the improvement of production efficiency and the shortening of the production cycle, and can still maintain the stable operation of the production line under disturbances such as order fluctuations or equipment aging, effectively reducing the frequency of manual intervention.

[0056] In this embodiment, a dual-engine collaborative mechanism of multi-objective optimization - discrete event simulation is constructed to achieve global dynamic joint optimization control of the buffer layout, buffer capacity configuration, and equipment speed of the photovoltaic module production line. Based on the swarm intelligence algorithm, global optimization is performed on the multi-objective functions (maximizing output, minimizing work-in-progress and energy consumption), and a set of joint decision variables including buffer activation positions, capacity thresholds, and equipment speed regulation parameters is generated. The discrete event simulation model is used to simulate the operation of the real production line, dynamically verify the actual effectiveness of the solution set, and feedback the production efficiency indicators. When the indicators do not meet the optimization conditions, the algorithm is triggered to iteratively update the solution set until the global optimal solution that simultaneously meets the process constraints, space limitations, and beat matching is output. This application breaks through the limitation of traditional single-variable static optimization, improves the utilization rate of the buffer area, enhances the equipment cooperation efficiency, and in the scenario of order fluctuations or sudden failures, shortens the production interruption time through dynamic optimization, realizing the three-in-one adaptive control of "layout - speed - capacity", providing a comprehensive and stable control method for the photovoltaic module production line.

[0057] In a possible implementation, the swarm intelligence algorithm is applied to optimize the multi-objective optimization model to obtain new joint decision variables, including:

[0058] The swarm intelligence algorithm is applied to solve the multi-objective function to generate a set of initial solutions;

[0059] The discrete event simulation model is used to perform simulation operation on the initial solutions to obtain the corresponding multi-objective indicators and calculate the fitness;

[0060] According to the fitness evaluation results, the iterative mechanism of the swarm intelligence algorithm is used to update the solutions until the convergence condition or the preset number of iterations is reached, and the optimal solution is output;

[0061] The corresponding joint decision variables are determined according to the optimal solution.

[0062] When constructing the objective function, a weighted comprehensive objective function based on normalization processing is adopted, and a multi-dimensional constraint condition system is set. The specific processing method is as follows:

[0063] The objective function fuses three major objectives of maximizing output, minimizing work-in-process, and minimizing energy consumption through a normalization weighting method. Its mathematical expression is:

[0064]

[0065] In the formula, TP is the actual output (pieces / hour), WIP is the number of work-in-process (pieces), E is the production energy consumption (kWh), and ω is the weight coefficient (∑ω = 1). Each objective item is normalized by the boundary value of the feasible region (subscript max / min) to eliminate the dimension difference. The weight coefficient ω1 + ω2 + ω3 = 1 reflects the optimization tendency; the constraint conditions are defined from three dimensions: process feasibility, space limitation, and production rhythm matching:

[0066] Process constraint: The equipment speed needs to satisfy υ i min ≤υ i ≤υ i max , ensuring that the equipment adjusts its speed within the process allowable range;

[0067] Space constraint: The buffer capacity needs to satisfy C j min ≤C j ≤C j max , avoiding exceeding the physical space limitation of the site

[0068] Rhythm matching constraint: The total time of each process ∑T k ≤T cycLE , ensuring the coordination of processes within the production cycle.

[0069] The above function and constraints are synergistically optimized through a swarm intelligence algorithm to finally obtain a global optimal solution that satisfies the multi-objective balance.

[0070] Among them, the normalization weighting method eliminates the dimension difference in this multi-objective optimization model and balances the influence of each objective. The optimization direction is flexibly adjusted through the weight coefficient, and the feasibility and practical applicability of the solution are ensured in combination with the constraint conditions. These processing methods jointly improve the practicality and effectiveness of the optimization model, enabling the generated solution to be applied in the actual production environment while satisfying multiple conflicting objectives.

[0071] In this embodiment, first, a swarm intelligence algorithm is used to globally search for the multi-objective function to generate an initial solution set covering multi-dimensional variables such as cache layout and device speed. Subsequently, each solution is injected into a discrete event simulation model to simulate the operation of the real production line, dynamically calculate output, work-in-process, and energy consumption indicators, and quantify the fitness. Based on the fitness evaluation results, the solution set is continuously optimized through iterative operations such as selection, crossover, and mutation until the convergence condition or the maximum number of iterations is reached. Finally, the global optimal solution is output and decoded into executable joint decision variables. This embodiment breaks through the limitations of traditional single-objective optimization. While ensuring the diversity of the solution set, it effectively avoids the risk of the disconnection between the theoretical solution and actual production through simulation verification, improves the resource allocation efficiency of the production line, shortens the optimization cycle, and demonstrates strong robustness in dealing with complex scenarios such as sudden equipment failures and sudden changes in order beats.

[0072] In a possible implementation, the swarm intelligence algorithm is a genetic algorithm.

[0073] The genetic algorithm supports Pareto front generation and the linear weighted method, flexibly balancing the conflicting objectives of output, work-in-process, and energy consumption, and is particularly suitable for solving the comprehensive objective function after normalization and weighting.

[0074] Figure 4 It is a schematic diagram of the Pareto front of the multi-objective optimization result shown in an embodiment of the present application. It can be seen that the Pareto front obtained based on the genetic algorithm is evenly distributed, facilitating the clear identification of the trade-off relationship between work-in-process and output.

[0075] In this embodiment, as the core of the swarm intelligence algorithm, the genetic algorithm conducts a wide-area search in the solution space through the population evolution mechanism (selection, crossover, mutation), avoids local optima, and supports the Pareto front or the linear weighted method to flexibly coordinate multi-objective conflicts. The algorithm is embedded with a constraint processing mechanism to automatically exclude infeasible solutions and ensure the engineering feasibility of the solutions. Combined with the closed-loop feedback of the discrete event simulation model, it dynamically corrects the influence of random interference on fitness and enhances the robustness of the solution set. With the help of a parallel computing architecture, it can synchronously simulate and evaluate 200 groups of initial solutions, compress the parameter optimization time, solve the problem of multi-variable coupling optimization of the photovoltaic production line, and achieve the global optimum.

[0076] In a possible implementation, the cache layout uses binary coding, the cache capacity uses integer coding, and the device speed uses real number coding or discrete gear coding.

[0077] Among them, the three types of variables act on the global optimization objective of the production line through a hybrid coding mechanism to jointly achieve the joint decision-making of buffer layout, buffer capacity allocation, and equipment speed. Optionally, when using binary coding to represent the enabling status of each candidate buffer, 0 indicates disabled and 1 indicates enabled; when setting the capacity configuration value of the enabled buffer through integer coding, the capacity configuration value takes discrete integer values within the range of 5 - 80 pieces; when dynamically adjusting the operating speed of key production equipment based on continuous variable coding, it takes values within the allowable range of each equipment's process parameters.

[0078] In this embodiment, the three types of decision variables of buffer layout, buffer capacity, and equipment speed are synergistically optimized through a hybrid coding mechanism. Binary coding is used to quickly screen the buffer enabling status, integer coding is used to accurately allocate buffer capacity, and real number / discrete gear coding is used to dynamically adjust equipment speed, realizing the global optimization of production rhythm matching and resource allocation, thereby effectively improving the production line throughput, reducing work-in-process inventory, and optimizing energy consumption distribution.

[0079] In actual operation, in order to ensure the stable operation of photovoltaic module production, it is necessary to dynamically optimize the discrete event simulation model and the multi-objective optimization model. However, continuous dynamic optimization will increase the computing pressure. Therefore, according to the different changes that occur during the production process of photovoltaic modules, it is necessary to start the optimization process in a timely manner.

[0080] Figure 5 It is a schematic diagram of the optimization process of the multi-objective optimization model by the swarm intelligence algorithm provided in an embodiment of the present application. As Figure 5 shown, it includes the following steps:

[0081] Initialize the population: Randomly generate solutions, including parameters such as buffer layout, buffer capacity, and equipment speed;

[0082] Simulation evaluation: Substitute the solution into the discrete event simulation model to simulate the production process;

[0083] Fitness calculation: Evaluate the solution according to the objective function; calculate indicators such as output, work-in-process quantity, and energy consumption;

[0084] Iterative update: Use operators such as Genetic Algorithm (GA), Particle Swarm Optimization (PSO), etc. to generate new solutions and repeat the simulation;

[0085] When it is judged whether the convergence condition or the preset number of iterations is reached, output the optimal solution: Obtain the optimal or near-optimal solution; otherwise, continuously iterate the update step until the convergence condition or the preset number of iterations is reached;

[0086] Output the optimal solution: Obtain the optimal or near-optimal solution.

[0087] In a possible implementation, the set time is determined according to actual order requirements, fault data, and personnel changes;

[0088] Among them, the large actual order demand, the amount of fault data, and the frequency of personnel changes are all inversely proportional to the set time.

[0089] When the order demand surges, shortening the set time can accelerate the optimization cycle, enabling parameters such as cache layout and device speed to quickly respond to order demands, avoiding production plan delays caused by overly long simulation cycles, and thus enhancing order delivery capabilities.

[0090] In the face of high-frequency equipment failures, shortening the set time can increase the sampling density of fault data, enabling the optimization model to more promptly respond to downtime impacts through cache capacity adjustment and device rate compensation, reducing work-in-process backlogs and energy consumption losses.

[0091] When personnel changes are frequent, dynamically shortening the set time can enhance the system's adaptability to changes in operation proficiency, maintaining stable product quality and production efficiency through real-time optimization of production rhythms and resource allocation.

[0092] In this embodiment, through flexible adjustment in the time dimension, dynamic optimization and allocation of production resources are achieved, effectively balancing the contradictions among production efficiency, stability, and energy consumption costs in complex scenarios such as order fluctuations, equipment failures, and personnel flows.

[0093] In the foregoing embodiment, the discrete event simulation model is mainly utilized. The following introduces the construction process of the discrete event simulation model.

[0094] In a possible implementation, before obtaining the discrete event simulation model and multi-objective optimization model of the photovoltaic module production line, it further includes:

[0095] Obtain the production parameters of each process of the photovoltaic module production line and the change ranges of the production parameters; among them, the production parameters include: equipment type, material transfer path, production rhythm, pipeline transfer speed, switching time, personnel configuration, number of buffer areas, buffer capacity, equipment failure rate, and equipment recovery time; the equipment type includes two or more of: string welding machine, laminating machine, layer press, typesetting machine, framing machine, cutting machine, conveyor, and curing machine;

[0096] Establish a discrete event simulation model based on the production parameters and the change ranges of the production parameters, and configure the candidate positions of the buffer areas and the upper and lower limits of the buffer capacity in the discrete event simulation model.

[0097] Among them, the discrete event simulation model is specifically a 1:1 digital twin model of the photovoltaic module production line constructed using the Plantsimulation engineering simulation software according to the process flow and factory layout of the photovoltaic module production line. In the specific implementation process of the solution of this application, the Plantsimulation software is interconnected with the PLC control system to achieve real-time simulation and optimization of the transient implementation on the production line. Figure 6 It is a schematic diagram of the discrete event simulation model of the photovoltaic module production line provided by an embodiment of this application.

[0098] For different processes, the production parameters of the corresponding different processes are different. For example: String welding machine: 70 - 120 pieces / min; Laminating welding machine: 15 - 20 s / component; Laminator: 60 - 90 s / component; Layout machine: 50 - 80 s / component; Framing machine: 20 - 40 s / component; Cutting machine: 15 - 20 s / component; Conveyor speed: 1 - 1.6 m / s.

[0099] In order to standardize the discrete event simulation model, the candidate positions of the buffer area and the upper and lower limits of the buffer area capacity can be pre-configured in the initial discrete event simulation model. For example: The candidate positions include welding - EL detection, cell layout - laminating welding, laminating welding - positioning glue, positioning glue - placing cushion blocks, placing back glue film - placing back glass / back plate, tape edge sealing - lamination, lamination - removing tape, appearance inspection - framing, etc.; According to the site space limitation, the upper and lower limits of the buffer capacity (1 - 80 pieces) are set. The discrete event simulation model can be adapted to multiple photovoltaic module production lines with different process flows, and the buffer area positions are set correspondingly for different photovoltaic module production lines.

[0100] In a possible implementation manner, the upper limit of the capacity of each buffer area is less than or equal to 50 pieces.

[0101] In this embodiment, the upper limit of the buffer area capacity is set to 50 to achieve efficient allocation of production resources. When the uneven material flow is caused by the difference in equipment speed, the buffer capacity limit prompts the model to find a balance between equipment rate adjustment and buffer layout selection, such as enabling more buffer areas or optimizing the equipment beat matching degree to compensate for the insufficient capacity, maintaining production continuity and avoiding inventory backlog. This mechanism reduces the ineffective energy consumption of equipment, reduces the factory space cost, and optimizes the production volume, energy consumption and space utilization rate.

[0102] In the actual implementation process, the discrete event simulation model is not put into application immediately after being established. During the establishment process of the discrete event simulation model, it is necessary to conduct trial operation and optimization on the discrete event simulation model to ensure that the discrete event simulation model can correctly simulate the events of the photovoltaic module production line.

[0103] Figure 7 It is a flowchart of modeling and simulation provided by an embodiment of the present invention, as Figure 7As shown, data collection is carried out to obtain the production parameters of each process of the photovoltaic module production line and the change range of the production parameters. A discrete event simulation model of the photovoltaic module production line is established, the model is run, and the correctness of the model is judged. When the model can correctly reflect the events of the photovoltaic module production line, the simulation results are analyzed, and an optimization plan is formulated for the simulation model. The simulation model is modified according to the optimization plan, and the modified simulation model is run again. When it is judged that the optimization performance is effective (optimizing the output, the number of work-in-process, and the production energy consumption), it is applied to the photovoltaic module production factory for trial production; otherwise, the optimization plan is re-formulated.

[0104] The core of the control method for photovoltaic module production based on the swarm intelligence algorithm mentioned in the foregoing embodiment lies in determining the joint decision variables for controlling the operation of the photovoltaic module production line. Currently, the control of the photovoltaic module production line mainly relies on the MES system. Therefore, after determining the joint decision variables for controlling the operation of the photovoltaic module production line, communication with the MES system is also required to achieve comprehensive control of the photovoltaic module production line.

[0105] In a possible implementation manner, when the production efficiency evaluation index meets the optimization conditions, the optimal solution corresponding to the multi-objective optimization model is sent to the MES, and the MES issues the speed regulation instruction and the buffer capacity control instruction to the equipment and / or the management personnel terminal of the photovoltaic module production line.

[0106] Among them, the optimal solution corresponding to the multi-objective optimization model is sent to the MES, and then the MES is responsible for issuing the speed regulation instruction and the buffer capacity control instruction to the equipment of the photovoltaic module production line to achieve precise control of the operation of the production line equipment.

[0107] In addition, in the actual implementation process, in order to facilitate the management personnel to conduct inspections and troubleshoot faults in a timely manner, a mobile terminal management interface is established to adapt to the MES system. When the production efficiency evaluation index reaches the optimization conditions, the MES will send the speed regulation instruction and the buffer capacity control instruction to the terminal equipment of the management personnel so that they can make manual judgments and supervise the equipment in a timely manner.

[0108] In this embodiment, when the optimal solution generated by the multi-objective optimization model is automatically sent to the production line equipment through the MES, the equipment speed adjustment and buffer capacity configuration can be quickly completed, ensuring that the production parameters are always in an optimized state and reducing the delay and error of manual intervention. At the same time, the mobile terminal interface synchronously pushes the key instructions to the management personnel, enabling them to monitor the optimization process in real time and conduct manual review. Especially when complex working conditions or abnormal events occur, they can quickly intervene in the decision-making through the terminal interface, retaining the high efficiency of the automated system and endowing the production management with the necessary flexibility.

[0109] It should be understood that the sequence numbers of the steps in the above embodiments do not indicate the order of execution, and the order of execution of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0110] To verify the solution of this application, experiments on offline optimization and dynamic rolling optimization were carried out respectively. The details are as follows:

[0111] Embodiment 1: Offline optimization

[0112] 1. Production line conditions

[0113] Target output: 500 components per day on average;

[0114] Maximum equipment speed: String welding machine: 70 - 120 pieces / min; Laminator: 60 - 90 seconds / piece; Framing machine: fixed at 30 pieces / h; Current buffer areas: between string welding and stacking welding (capacity 10 pieces), between lamination and framing (capacity 8 pieces).

[0115] 2. Algorithm and simulation configuration

[0116] Algorithm and coding: Genetic algorithm (GA), population size 50, maximum number of iterations 100;

[0117] Buffer layout: An additional buffer area can be added between stacking welding and lamination (0 means not used; 1 means used);

[0118] Buffer capacity: The capacity range of each buffer area is 0 - 50;

[0119] Equipment speed: The speed of the string welding machine is discretized as [70, 80, 90, 100, 110, 120]; The lamination machine cycle can take continuous values from 60 to 90 seconds; The framing machine remains fixed.

[0120] Multi-objective function:

[0121] Simulation running time: Simulate for 2 days (16 hours / day, two shifts), and take the average result multiple times.

[0122] 3. Optimization process and results

[0123] After about 80 generations of iteration, convergence was achieved, and an example of the optimal solution is as follows:

[0124] Enable the buffer area between stacking welding and lamination, with a capacity of 20 pieces; Expand the buffer between string welding and stacking welding to 25 pieces, and adjust the buffer between lamination and framing to 6 pieces; The speed of the string welding machine is 95 pieces / min; The lamination machine cycle is set to 70 seconds / piece.

[0125] Compared with the original plan: The output increased by about 6%; The work-in-progress decreased by about 12%; The energy consumption decreased by about 4%.

[0126] Embodiment 2: Dynamic Rolling Optimization

[0127] 1. Production Line Operation Monitoring

[0128] In actual production, the order volume change is statistically analyzed weekly (such as a 20% increase in peak season demand), and the equipment failure rate (such as the shortening of the fault interval of the string welding machine);

[0129] The latest equipment data (such as fault frequency, yield rate, operation time) is input into the updated simulation model.

[0130] 2. Regular or Real-time Update

[0131] Every Sunday, rapid iteration is performed through the swarm intelligence algorithm (which can be shortened to 30 generations) to output new cache and speed configurations; speed adjustment instructions (such as the speed of the automatic transmission line and the string welding machine) are issued through the production line MES system and actually executed next Monday; when data is collected again on Friday, the actual output, work-in-process, and energy consumption are compared and analyzed to evaluate the optimization effect.

[0132] 3. Continuous Effect

[0133] The production line after dynamic adjustment still maintains high efficiency in the case of large batch orders and fault fluctuations; there is no significant accumulation of work-in-process, and the land occupation and capital pressure decrease; the overload situation of the equipment is reduced, and the failure rate and maintenance cost are further reduced.

[0134] The following is the device embodiment of the present invention. For the details not described in detail, reference can be made to the corresponding method embodiment above.

[0135] Figure 8 The structural schematic diagram of the control device for photovoltaic module production based on the swarm intelligence algorithm provided by the embodiment of the present invention is shown. For the convenience of description, only the parts related to the embodiment of the present invention are shown and are described in detail as follows:

[0136] As Figure 8 shown, the control device for photovoltaic module production based on the swarm intelligence algorithm includes:

[0137] An acquisition module 801, configured to acquire a discrete event simulation model and a multi-objective optimization model of a photovoltaic module production line; wherein, the multi-objective optimization model includes a multi-objective function and joint constraint conditions; the multi-objective function uses cache layout, cache capacity, and equipment speed as joint decision variables; the joint constraint conditions include process constraints, space constraints, and beat matching constraints; the discrete event simulation model uses the equipment operation data within a set time as input; and uses output output, work-in-process quantity, and production energy consumption as output parameters;

[0138] A control module 802, configured to update the discrete event simulation model for simulation operation with the joint decision variables output by the multi-objective optimization model to obtain a production efficiency evaluation index;

[0139] When the production efficiency evaluation index meets the optimization conditions, the operation of the photovoltaic module production line is controlled by the corresponding joint decision variables; otherwise, the multi-objective optimization model is optimized by using the swarm intelligence algorithm to obtain new joint decision variables until the production efficiency evaluation index meets the optimization conditions.

[0140] In a possible implementation, the control module 802 is specifically configured to:

[0141] Use the swarm intelligence algorithm to solve the multi-objective function and generate a set of initial solutions;

[0142] Use the discrete event simulation model to simulate the operation of the initial solutions, obtain the corresponding multi-objective indicators and calculate the fitness;

[0143] According to the fitness evaluation results, use the iterative mechanism of the swarm intelligence algorithm to update the solutions until the convergence condition or the preset number of iterations is reached, and output the optimal solutions;

[0144] Determine the corresponding joint decision variables according to the optimal solutions.

[0145] In a possible implementation, the acquisition module 801 is further configured to acquire the production parameters of each process of the photovoltaic module production line and the change ranges of the production parameters; wherein, the production parameters include: equipment type, material transfer path, production beat, pipeline transfer speed, switching time, personnel configuration, buffer number, buffer capacity, equipment failure rate and equipment recovery time; the equipment type includes two or more of: string welding machine, laminating machine, layer press, typesetting machine, framing machine, cutting machine, conveyor and curing machine;

[0146] The control module 802 is further configured to establish a discrete event simulation model according to the production parameters and the change ranges of the production parameters, and configure the candidate positions of the buffer areas and the upper and lower limits of the buffer capacity in the discrete event simulation model.

[0147] In a possible implementation, the control module 802 is further configured to, when the production efficiency evaluation index meets the optimization conditions, send the optimal solutions corresponding to the multi-objective optimization model to the MES, and the MES issues the speed regulation instruction and the buffer capacity control instruction to the equipment and / or the management personnel terminal of the photovoltaic module production line.

[0148] In this embodiment, a global dynamic joint optimization control of the buffer layout, buffer capacity configuration and equipment speed of the photovoltaic module production line is realized by constructing a multi-objective optimization-discrete event simulation dual-engine collaborative mechanism. Based on the swarm intelligence algorithm, a global optimization of the multi-objective function (maximizing output, minimizing work-in-process and energy consumption) is carried out to generate a set of joint decision variables including buffer enabling positions, capacity thresholds and equipment speed regulation parameters. The discrete event simulation model is used to simulate the operation of the actual production line, dynamically verify the actual effectiveness of the solution set and feedback the production efficiency index. When the index does not meet the optimization conditions, the algorithm is triggered to iteratively update the solution set until a global optimal solution that simultaneously meets the process constraints, space limitations and beat matching is output. This application breaks through the limitations of traditional single-variable static optimization, improves the utilization rate of the buffer area, enhances the equipment collaboration efficiency, and in the scenario of order fluctuations or sudden failures, shortens the production interruption time through dynamic optimization, realizing the three-in-one adaptive control of "layout-speed-capacity", and providing a comprehensive and stable control method for the photovoltaic module production line.

[0149] An embodiment of the present invention also provides a control device for photovoltaic module production based on the swarm intelligence algorithm, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the method in the above method embodiment is implemented. Exemplarily, the control device for photovoltaic module production based on the swarm intelligence algorithm can be a computer, a controller, etc., which is not limited herein.

[0150] In the above embodiments, the descriptions of the respective embodiments have their own emphases. For parts not detailed or recorded in a certain embodiment, reference may be made to the relevant descriptions of other embodiments. Without special instructions and logical conflicts, the terms and / or descriptions between different embodiments are consistent and can be mutually referred to, and the technical features in different embodiments can be combined to form new embodiments according to their internal logical relationships.

[0151] The above-described embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A control method for photovoltaic module production based on swarm intelligence algorithm, characterized in that: include: Obtain a discrete event simulation model and a multi-objective optimization model for a photovoltaic module production line; wherein the multi-objective optimization model includes a multi-objective function and joint constraints; the multi-objective function uses cache layout, cache capacity and equipment speed as joint decision variables; the joint constraints include process constraints, space constraints and beat matching constraints; the discrete event simulation model uses equipment operation data within a set time as input; and uses output output, number of work-in-progress and production energy consumption as output parameters; Using the joint decision variables output by the multi-objective optimization model, the discrete event simulation model is updated to perform simulation operation to obtain a production efficiency evaluation index; When the production efficiency evaluation index meets the optimization conditions, the operation of the photovoltaic module production line is controlled by the corresponding joint decision variables; otherwise, the swarm intelligence algorithm is applied to optimize the multi-objective optimization model to obtain new joint decision variables until the production efficiency evaluation index meets the optimization conditions.

2. The control method for photovoltaic module production based on swarm intelligence algorithm according to claim 1, characterized in that: The application of a swarm intelligence algorithm to optimize the multi-objective optimization model to obtain new joint decision variables includes: Applying a swarm intelligence algorithm to solve the multi-objective function and generate a set of initial solutions; The discrete event simulation model is used to simulate the initial solution, obtain the corresponding multi-objective indicators and calculate the fitness; According to the fitness evaluation results, the iterative mechanism of the swarm intelligence algorithm is used to update the solution until the convergence condition or the preset iteration number is reached, and the optimal solution is output; The corresponding joint decision variables are determined according to the optimal solution.

3. The control method for photovoltaic module production based on swarm intelligence algorithm according to claim 2 is characterized in that: The swarm intelligence algorithm is a genetic algorithm.

4. The control method for photovoltaic module production based on swarm intelligence algorithm according to claim 1 or 2, characterized in that: The cache layout is coded in binary, the cache capacity is coded in integer, and the device speed is coded in real number or discrete gear.

5. The control method for photovoltaic module production based on swarm intelligence algorithm according to claim 1, characterized in that: The equipment operation data within the set time includes equipment failure frequency, pipeline transmission speed, switching time and equipment recovery time.

6. The control method for photovoltaic module production based on swarm intelligence algorithm according to claim 1, characterized in that: The setting time is determined according to actual order requirements, fault data and personnel changes; Among them, the actual order demand, the amount of fault data, and the frequency of personnel changes are all inversely proportional to the set time.

7. The control method for photovoltaic module production based on swarm intelligence algorithm according to claim 1, characterized in that: Before obtaining the discrete event simulation model and the multi-objective optimization model of the photovoltaic module production line, the method further includes: Obtaining the production parameters of each process of the photovoltaic module production line and the range of variation of the production parameters; wherein the production parameters include: equipment type, material transmission path, production beat, assembly line transmission speed, switching time, staffing, number of buffer areas, buffer capacity, equipment failure rate and equipment recovery time; equipment types include: two or more of stringing machine, stitch welding machine, laminating machine, typesetting machine, framing machine, cutting machine, conveyor and curing machine; The discrete event simulation model is established according to the production parameters and the variation range of the production parameters, and the candidate positions of the buffer area and the upper and lower limits of the buffer area capacity in the discrete event simulation model are configured.

8. The control method for photovoltaic module production based on swarm intelligence algorithm according to claim 7, characterized in that: The upper limit of each buffer area is less than or equal to 50 items.

9. The control method for photovoltaic module production based on swarm intelligence algorithm according to claim 1, characterized in that: When the production efficiency evaluation index meets the optimization conditions, the optimal solution corresponding to the multi-objective optimization model is sent to the manufacturing execution system MES, and the MES sends the speed regulation instruction and the cache capacity control instruction to the equipment and / or management personnel terminal of the photovoltaic module production line.

10. A control device for photovoltaic module production based on swarm intelligence algorithm, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the method according to any one of claims 1 to 9 when executing the computer program.

Images