Waste plastic regeneration production process energy-saving method based on tabu search algorithm

Abstract

The invention belongs to the technical field of information, and particularly relates to a taboo search algorithm-based waste plastic regeneration production process energy-saving method, which comprises the following steps of: firstly, carrying out detailed modeling and energy consumption analysis on a production process, covering links such as raw material pretreatment, melt extrusion and the like, and constructing an energy consumption model; secondly, designing a tabu search algorithm, representing an optimization scheme by a solution vector, and searching an optimal solution by taking the minimum energy consumption as a target function through the steps of initial solution generation, neighborhood search, tabu table operation and the like; and finally, making an optimization scheme according to the optimal solution and implementing, monitoring and evaluating in real time, and adjusting and perfecting according to a result. The scheme aims at improving the energy utilization efficiency and reducing the cost and environmental pollution.

Description

Technical Field

[0001] The present invention belongs to the field of recycled plastic blow molding manufacturing, specifically an energy-saving method for the waste plastic recycling production process based on a taboo search algorithm Background Art

[0002] With the widespread use of plastic products, the amount of waste plastics is increasing. Improper disposal of waste plastics will not only cause serious environmental pollution, but also waste a lot of resources. Therefore, the recycling of waste plastics has become an important research direction in the current environmental protection field.

[0003] The waste plastic recycling process involves multiple links, such as raw material sorting, cleaning, melting, molding, etc. Each link consumes a lot of energy, such as electricity, heat, etc. At present, the production processes of many waste plastic recycling companies have the problem of low energy efficiency, resulting in high production costs and increasing negative impacts on the environment.

[0004] Traditional production process optimization methods are mostly based on experience and trial and error, lacking in systematicness and scientificity. With the development of artificial intelligence technology, some intelligent optimization algorithms have been introduced into production process optimization, but there are still relatively few studies on energy-saving optimization of waste plastic recycling production processes. Therefore, it is of great practical significance to develop an efficient energy-saving optimization solution for waste plastic recycling production processes based on intelligent algorithms. Summary of the invention

[0005] The present invention provides an energy-saving optimization scheme for a waste plastic recycling production process based on a taboo search algorithm, comprising the following steps:

[0006] Model the waste plastic recycling production process in detail, divide it into multiple sub-processes, analyze the sources and influencing factors of energy consumption in each sub-process, and establish an energy consumption model. The energy consumption model is constructed based on factors such as equipment power parameters, operating time, and production output to quantify the energy consumption of each sub-process and the entire production process;

[0007] The design of a taboo search algorithm includes representing an optimization scheme of a waste plastic recycling production process as a solution vector, wherein each element in the solution vector corresponds to a production process parameter; generating an initial solution by random generation or an experience-based method; defining a neighborhood structure, and generating a neighborhood solution by making a slight adjustment to the elements in the solution vector; establishing a taboo table, and recording the most recently visited solutions to prevent the algorithm from falling into a local optimum; and taking the minimum energy consumption of the entire waste plastic recycling production process as an objective function, wherein the objective function is expressed as:

[0008] Where E represents the total energy consumption, ei represents the energy consumption of the i-th sub-process, wi represents the energy consumption weight of the i-th sub-process, n represents the number of sub-processes, and the objective function comprehensively considers the energy consumption of each sub-process, and other constraints can be added according to actual production needs; set the search termination condition. When the search termination condition is met, the algorithm stops searching and outputs the optimal solution currently found;

[0009] According to the optimal solution obtained by the taboo search algorithm, a detailed energy-saving optimization plan for the waste plastic recycling production process is formulated, and the plan is implemented in actual production. The energy consumption and product quality indicators in the production process are monitored in real time. The actual monitoring data are compared with the data before optimization to evaluate the effect of the optimization plan. If the expectation is not met, the energy consumption model, the parameters of the taboo search algorithm or the neighborhood search strategy are adjusted and improved, and the optimization calculation is performed again until a satisfactory energy-saving effect is achieved.

[0010] Furthermore, the waste plastic recycling production process is modeled in detail, specifically including modeling of sub-processes such as raw material pretreatment, melt extrusion, and molding processing. The raw material pretreatment stage considers the power of the cleaning equipment, the operating time, and the relationship between the cleaning time and the raw material characteristics; the melt extrusion stage considers factors such as the power of the heating device, the temperature setting, the specific heat capacity of the material, the heating requirements, and the production efficiency; the molding processing stage considers the type of molding equipment and the molding process parameters.

[0011] Furthermore, the solution vector encoding includes key parameters in the production process, such as the speed and temperature of the cleaning equipment, the screw speed, heating temperature, and extrusion pressure of the melt extruder, the injection pressure and holding time of the molding equipment, etc. Each parameter has a corresponding range of values, and the range of values ​​is determined according to the technical specifications of the equipment and the production process requirements.

[0012] Furthermore, the generation of the initial solution adopts a method that combines random generation and empirical values. First, the approximate reasonable range of some parameters is determined based on production experience, and the parameter values ​​in the initial solution vector are randomly generated within this range. After generating multiple initial solutions, the energy consumption calculation value is relatively high. Further, the neighborhood search strategy adopts a variety of neighborhood search operations, including parameter fine-tuning operations, which increase or decrease a certain value for a parameter in the solution vector; parameter exchange operations, which exchange the values ​​of two parameters in the solution vector; and parameter insertion operations, which insert a parameter in the solution vector into another position.

[0013] Furthermore, the taboo table stores key features of neighborhood solutions and corresponding taboo periods. When a new neighborhood solution is generated, it is checked whether the solution is in the taboo table. If it is and the taboo period has not expired, the solution is regarded as a taboo solution unless the unbanning condition is met, that is, the objective function value of the solution is better than the current optimal solution. Otherwise, the solution is not considered. The taboo table is updated after each iteration, the newly accessed solution is added to the taboo table, and the taboo period of the existing taboo solution is shortened. When the taboo period is 0, the solution is removed from the taboo table.

[0014] Furthermore, the search termination condition includes reaching the maximum number of iterations, such as setting the maximum number of iterations to 1000. When the number of iterations reaches 1000, the algorithm stops; or the objective function value does not improve significantly within a certain number of iterations, such as setting the objective function value of the optimal solution to change by less than a certain threshold value, such as 0.1%, in 50 consecutive iterations, the objective function value is considered to have converged and the algorithm stops.

[0015] Furthermore, when implementing the optimization plan in actual production, the operating status of the production equipment, energy consumption, product quality and other indicators are monitored in real time, and relevant data are collected through energy consumption monitoring equipment and quality inspection instruments installed on the equipment. The power consumption and heat energy consumption data of each device are recorded once an hour. At regular intervals, such as half an hour, the quality of the recycled plastic products produced is inspected, including dimensional accuracy, strength, appearance and other aspects.

[0016] Furthermore, the energy-saving optimization scheme for the waste plastic recycling production process based on the taboo search algorithm is characterized in that it includes collecting detailed energy consumption data of various equipment in each production line within the enterprise, with a time span of one quarter, and establishing a detailed energy consumption model covering multiple links such as raw material pretreatment, melting, molding, and post-processing; the solution vector encoding contains multiple key parameters, and the initial solution is generated by fine-tuning the experience values ​​of the company's senior engineers. After calculating the energy consumption, the optimal one is selected, a diversified strategy is adopted in the neighborhood search, and a dynamic update strategy is adopted for the taboo table; after multiple iterations of the taboo search algorithm, an optimization scheme is obtained, and the enterprise's digital management system is used to monitor the equipment operation status and energy consumption data in real time.

[0017] Furthermore, according to the production characteristics of the enterprise, we focus on collecting energy consumption and production data of specific links and establish energy consumption models of corresponding links; the solution vector encoding revolves around specific equipment parameters, the initial solution is randomly generated and combined with the historical production data of the enterprise to screen out the better one, the neighborhood search adopts a combination of parameter fine-tuning and local parameter combination adjustment, and the taboo table setting sets the taboo period differently according to the importance of the parameters and the degree of impact on energy consumption; the algorithm is iterated a certain number of times to obtain the optimization plan, and regular maintenance and calibration of equipment are strengthened during implementation.

[0018] Beneficial Effects

[0019] This energy-saving optimization solution for the waste plastic recycling production process based on the taboo search algorithm has significant effects. In terms of improving energy efficiency, through detailed modeling and energy consumption analysis of each production link, high energy consumption points are accurately located, and with the help of the optimization of the taboo search algorithm, energy consumption is significantly reduced. For example, in the melt extrusion process, by optimizing and adjusting parameters such as heating temperature and screw speed, energy waste is effectively reduced and efficient energy utilization is achieved.

[0020] From the perspective of cost reduction, the reduction of energy consumption directly leads to the reduction of production costs. Taking a waste plastic recycling enterprise as an example, after implementing this plan, the monthly energy cost dropped significantly. At the same time, due to the optimization of the production process, the operation efficiency of the equipment was improved, the equipment maintenance cost and raw material waste were reduced, and the production cost was further compressed.

[0021] In terms of environmental protection, the reduction of energy consumption in the recycling process of waste plastics means the reduction of greenhouse gas and other pollutant emissions. This not only helps the green development of the company itself, but also has positive significance for the sustainable development of the entire environment. In addition, the optimization plan ensures and improves product quality, enhances the competitiveness of the company in the market, and forms a win-win situation of economic and environmental benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 Flowchart of the tabu search algorithm. DETAILED DESCRIPTION

[0023] Example 1

[0024] Production process modeling and energy consumption analysis

[0025] The waste plastic recycling production process is modeled in detail and divided into multiple sub-processes, including raw material pretreatment, melt extrusion, and molding processing. For each sub-process, the source of energy consumption and influencing factors are analyzed. For example, in the raw material pretreatment stage, the energy consumption of the cleaning equipment is related to the equipment power and operating time; in the melt extrusion stage, the power and temperature setting of the heating device, as well as the material characteristics and production efficiency will affect the energy consumption; in the molding processing stage, the energy consumption of different types of molding equipment depends on their respective molding process parameters.

[0026] Energy consumption models are established to quantify the energy consumption of each sub-process and the entire production process. For example, the energy consumption of cleaning equipment is the product of equipment power and operating time; the energy consumption of melt extrusion equipment, in addition to heating power, is also related to the energy required for material heating and energy loss during the production process; the energy consumption of molding equipment (such as injection molding machines) should consider the relationship between injection, pressure holding, cooling and other stages and the relevant parameters of each stage. By establishing an accurate energy consumption model, a quantitative basis is provided for subsequent energy-saving optimization.

[0027] Tabu search algorithm design

[0028] Solution representation: The optimization plan for the waste plastic recycling production process is represented by a solution vector. Each element in the solution vector corresponds to a parameter in the production process, such as the speed of the cleaning equipment, the temperature setting of the melt extruder, the pressure of the molding equipment, etc. Changing the value of the solution vector element can change the operation mode of the production process, thereby affecting energy consumption.

[0029] Initial solution generation: The initial solution can be generated by random generation or experience-based methods. Random generation can ensure a wide search space, but may require more iterations to find a better solution; experience-based generation can use existing production experience to approach a better solution faster. In practical applications, multiple initial solutions can be randomly generated first, and then the solution close to the empirical value can be selected as the initial solution.

[0030] Neighborhood search strategy: Define the neighborhood structure and generate neighborhood solutions by fine-tuning the elements in the adjustment vector. For example, increase or decrease a certain temperature parameter, or fine-tune the running time of a certain device. The neighborhood search strategy should ensure that the neighborhood solutions are diverse, while avoiding too large a search range to reduce computational efficiency. A variety of operations can be used, such as swapping the values ​​of two device parameters, or inserting a device parameter into another position.

[0031] Taboo table setting: Establish a taboo table to record the most recently visited solutions to prevent the algorithm from falling into a local optimum. The taboo table stores key information of the solution vector (such as certain eigenvalues ​​or hash values) and the corresponding taboo period. When generating a new neighborhood solution, check whether it is in the taboo table. If it is and the taboo period has not expired, the solution is a taboo solution and will not be considered unless its objective function value (energy consumption value) is better than the current optimal solution. The taboo period can be dynamically adjusted according to the scale and complexity of the problem, such as gradually shortening it as the number of iterations increases to expand the search range.

[0032] Objective function definition: The objective function is to minimize the energy consumption of the entire waste plastic recycling production process. It comprehensively considers the energy consumption of each sub-process, combines actual production needs, and adds constraints such as product quality requirements and production efficiency requirements.

[0033] Search termination conditions: Set the search termination conditions, such as reaching the maximum number of iterations, or the objective function value has not improved significantly within a certain number of iterations. When these conditions are met, the algorithm stops searching and outputs the current optimal solution.

[0034] Optimization plan implementation and feedback

[0035] According to the optimal solution obtained by the taboo search algorithm, a detailed energy-saving optimization plan for the waste plastic recycling production process is formulated. The plan includes adjusting the production equipment parameters, such as adjusting the heating temperature of the melt extruder according to the optimized temperature; optimizing the production process sequence; and reasonably allocating energy supply.

[0036] Implement the optimization plan in actual production and monitor indicators such as energy consumption and product quality in real time. Compare the actual monitoring data with the data before optimization to evaluate the effect of the plan. If the effect does not meet expectations, analyze the reasons. It may be that the actual production environment is different from the modeling assumptions, or the algorithm is trapped in a local optimum. According to the analysis results, adjust the energy consumption model, taboo search algorithm parameters or neighborhood search strategy, and re-optimize the calculation until a satisfactory energy-saving effect is achieved.

[0037] Production process research: Conduct field research in the production workshop of waste plastic recycling enterprises to understand in detail each link of the production process, including equipment models, specifications, quantities, and the order and method of material flow between various equipment. Draw a detailed schematic diagram of the production process, marking the equipment name, operation process and key parameters of each link.

[0038] Energy consumption data collection: Energy consumption monitoring equipment, such as power monitors and heat meters, is installed on production equipment to collect real-time energy consumption data of each device under different production conditions. At the same time, relevant information such as equipment operating time, production output, raw material type and quality are recorded. The collected data covers at least one production cycle (such as one week or one month) to ensure data representativeness.

[0039] Data collation and analysis: Collect and analyze the collected energy consumption data and production-related information. Statistical energy consumption data by production process, and calculate the average energy consumption and energy consumption fluctuation range of each sub-process. Analyze the correlation between energy consumption and production parameters (such as equipment speed, temperature, pressure, etc.) to provide data support for energy consumption modeling. For example, it is found that the energy consumption of the melt extruder increases with the increase of heating temperature, and decreases with the improvement of production efficiency within a certain range.

[0040] Sub-process energy consumption model construction:

[0041] Raw material pretreatment: This stage mainly includes sorting, cleaning and crushing operations. Taking cleaning equipment as an example, its energy consumption is related to the equipment power and operating time. When cleaning different types and qualities of waste plastics, the cleaning time and power may be different. Based on actual data, a relationship model between cleaning time and raw material characteristics (such as dirt level, plastic material, etc.) can be established to optimize energy consumption calculation.

[0042] Melt extrusion stage: The energy consumption of the melt extruder is used to heat the waste plastic and extrude it. Its energy consumption is related to the heating power, heating time, material properties (such as specific heat capacity, heating requirements) and energy loss in the production process (such as heat dissipation and mechanical friction loss).

[0043] Molding process: The energy consumption of molding equipment is related to the equipment type (such as injection molding machine, blow molding machine, etc.) and molding process parameters. Taking the injection molding machine as an example, the energy consumption in the injection stage is related to the injection pressure and speed, the holding stage is related to the holding pressure and time, and the cooling stage is related to the cooling medium flow and temperature. By analyzing the working principle and energy consumption characteristics, the corresponding energy consumption model is established.

[0044] Integration of total energy consumption model: Integrate the energy consumption models of each sub-process to obtain the total energy consumption model of the entire waste plastic recycling production process. This model covers the energy consumption of raw material pretreatment, melt extrusion, molding processing, etc., and also considers other possible energy consumption links (such as drying links, post-processing links, etc.). Through this model, energy consumption can be quickly calculated according to different production parameter settings.

[0045] Solution encoding and initialization:

[0046] Solution encoding: Encode the key parameters in the production process, such as the speed and temperature of the cleaning equipment, the screw speed, heating temperature, extrusion pressure of the melt extruder, the injection pressure and holding time of the molding equipment, into a solution vector. Each parameter has a range of values ​​determined according to the equipment technical specifications and production process requirements.

[0047] Initial solution generation: The initial solution is generated by combining random generation and empirical value. First, determine the approximate reasonable range of some parameters based on production experience, and then randomly generate parameter values ​​in the initial solution vector within these ranges. For example, the speed of the cleaning equipment is generally between 100-500r / min, and the initial value is randomly generated within this range. Generate multiple initial solutions (such as 10), select the solution with a relatively low energy consumption calculation value as the initial solution, and enter the taboo search algorithm iteration.

[0048] Neighborhood search and tabu table operations:

[0049] Neighborhood search strategy: Use multiple neighborhood search operations:

[0050] Parameter fine-tuning operation: Make a small adjustment to a parameter in the solution vector, such as increasing or decreasing the heating temperature of the melt extruder by a certain degree (such as 5°C), to generate a new neighborhood solution.

[0051] Parameter exchange operation: Exchange the values ​​of two parameters in the solution vector, such as exchanging the speed of the cleaning equipment and the injection pressure of the molding equipment, to obtain a new neighborhood solution.

[0052] Parameter insertion operation: insert a parameter in the solution vector into another position, such as inserting the screw speed of the melt extruder after the holding time, to form a new neighborhood solution.

[0053] Taboo table operation: Establish a taboo table to store the key features of neighborhood solutions (such as the hash value of the solution vector) and the corresponding taboo period. When generating a new neighborhood solution, calculate its hash value and check whether it is in the taboo table. If it is and the taboo period has not expired, the solution is a taboo solution. However, if the energy consumption value of the taboo solution is better than the current optimal solution, the solution is unbanned and allowed to participate in subsequent searches. After each iteration, the taboo table is updated, the newly accessed solution is added, and the taboo period of the existing taboo solution is shortened. When the taboo period is 0, the solution is removed from the taboo table.

[0054] Objective function calculation and optimal solution update:

[0055] Objective function calculation: For each generated neighborhood solution, its energy consumption value, i.e., the objective function value, is calculated according to the established energy consumption model.

[0056] Optimal solution update: In each iteration, compare the objective function value of the current neighborhood solution with the current optimal solution. If the energy consumption of the neighborhood solution is lower, update the current optimal solution to the neighborhood solution. Record the optimal solution and its objective function value of each iteration so that the global optimal solution can be output when the algorithm terminates.

[0057] Search termination condition judgment: Set the search termination condition. When one of the following conditions is met, the algorithm stops searching:

[0058] When the maximum number of iterations is reached, for example, if it is set to 1000, the algorithm stops.

[0059] If the objective function value does not improve significantly within a certain number of iterations, such as if the objective function value of the optimal solution changes by less than a certain threshold (such as 0.1%) in 50 consecutive iterations, the objective function value is considered to have converged and the algorithm stops.

[0060] Optimization plan formulation: After the taboo search algorithm is terminated, a detailed energy-saving optimization plan for the waste plastic recycling production process is formulated based on the parameter values ​​of the optimal solution vector. For example, the equipment operating parameters are adjusted according to the optimized speed and temperature parameters of the cleaning equipment; the corresponding settings are made according to the optimal screw speed, heating temperature and extrusion pressure of the melt extruder; and the operation process is optimized according to the optimal injection pressure, holding time and other parameters of the molding equipment.

[0061] Implementation and monitoring: Implement optimization plans on waste plastic recycling production lines, and monitor production equipment operating status, energy consumption, product quality and other indicators in real time. Collect data through energy consumption monitoring equipment and quality testing instruments installed on the equipment, such as recording equipment power and heat energy consumption data every hour, and testing the size accuracy, strength, and appearance of recycled plastic products every half an hour.

[0062] Effect evaluation and feedback: Compare the production data before and after the implementation of the optimization plan, and calculate indicators such as the energy consumption reduction ratio and product quality change. If the energy consumption reduction does not meet expectations or there are problems with product quality, analyze the reasons, which may be errors in the energy consumption model, changes in the actual production environment (such as fluctuations in raw material quality, equipment aging), or deviations in the execution of the optimization plan. According to the analysis results, correct the energy consumption model, or adjust the taboo search algorithm parameters (such as neighborhood search range, taboo period), re-optimize the calculation and implement it again until a satisfactory energy-saving effect is achieved and the product quality meets the requirements. Through continuous implementation, evaluation and feedback, the energy utilization efficiency and product quality of the waste plastic recycling production process are continuously optimized.

[0063] Example 2

[0064] Comprehensive optimization of large-scale waste plastic recycling enterprises, data collection and model establishment: Detailed energy consumption data of more than 20 types of equipment in each production line within the enterprise was collected over a quarter. Through analysis, it was found that the melting link accounted for the largest proportion of energy consumption, about 45%, followed by the molding link, accounting for 30%. Based on this, a detailed energy consumption model covering more than 10 links such as raw material pretreatment, melting, molding, and post-processing was established.

[0065] Application of taboo search algorithm: The solution vector encoding contains more than 50 key parameters, such as the temperature and screw speed of different types of melt extruders, the injection pressure and holding time of various molding equipment, etc. The initial solution is generated by fine-tuning the experience value of senior engineers of the enterprise, and the optimal one is selected after calculating the energy consumption. In the neighborhood search, a diversified strategy is adopted, such as setting different adjustment steps for different equipment parameters. The taboo table adopts a dynamic update strategy to adjust the taboo period according to the number of iterations and search results.

[0066] Optimization plan formulation and implementation: After 2,000 iterations of the taboo search algorithm, an optimization plan was obtained. For example, the heating temperature of some melt extruders was reduced by 10-20°C, the screw speed was adjusted, and the holding time of the molding equipment was shortened by 10-15%. During the implementation process, the enterprise digital management system was used to monitor the equipment operation status and energy consumption data in real time.

[0067] Example 3

[0068] Data collection and model building: In view of the production characteristics of the enterprise, the energy consumption and production data of the three links of cleaning, extrusion and molding were collected for two months. It was found that the cleaning equipment had high energy consumption due to long-term high-load operation, and the extrusion link had unstable temperature control, which affected product quality and energy consumption. Based on this, an energy consumption model for the three links was established, taking into account factors such as cleaning equipment power, cleaning time, raw material characteristics, extruder temperature, pressure, and extrusion speed.

[0069] Application of taboo search algorithm: The solution vector encoding mainly revolves around more than 20 parameters such as the speed, temperature, cleaning agent concentration of the cleaning equipment, the temperature, pressure, screw speed of the extruder, etc. 50 initial solutions are randomly generated, and the better ones are selected in combination with the company's historical production data. The neighborhood search adopts a combination of parameter fine-tuning and local parameter combination adjustment. The taboo table setting sets the taboo period differently according to the importance of the parameters and the degree of impact on energy consumption.

[0070] Optimization plan formulation and implementation: After 1,500 iterations of the algorithm, an optimization plan was obtained. For example, the speed of the cleaning equipment was reduced by 10%, the concentration of the cleaning agent was adjusted, and the temperature curve and pressure control of the extruder were optimized. During the implementation process, regular maintenance and calibration of the equipment were strengthened to ensure effective execution of parameter adjustments.

[0071] Example 4

[0072] Data collection and model building: Data collection was conducted for one month on the only three links of the company: sorting, cleaning, and melt granulation. Due to the old equipment, it was found that the heating system of the melt granulator was aging, the energy consumption was high, and the temperature control was inaccurate. A simple energy consumption model was established, taking into account factors such as sorting labor energy consumption, cleaning equipment power and time, and melt granulator heating power and operating time.

[0073] Application of taboo search algorithm: The solution vector encoding targets more than 10 key parameters such as cleaning equipment speed, melt granulator temperature, screw speed, etc. The initial solution is generated by referring to the experience of similar enterprises and 20 solutions are selected, and the one with the lowest energy consumption is selected. The neighborhood search adopts a simple parameter fine-tuning method, such as adjusting the temperature of the melt granulator by 5°C each time. The taboo table is simple to set, and the fixed taboo period is 50 iterations.

[0074] Optimization plan formulation and implementation: After 800 iterations of the algorithm, an optimization plan was formulated, such as partial modification of the heating system of the melt granulator and adjustment of the running time and speed of the cleaning equipment. During the implementation process, the company tracked energy consumption and product quality through manual records and simple monitoring equipment.

[0075] Example 5

[0076] Data collection and evaluation index determination: Collect energy consumption data and product quality data of each link in the production process, such as crushing, cleaning, mixing, extrusion molding, etc., including indicators such as the flatness and strength of the board. The collection period is three months. Determine the reduction of energy consumption and the improvement of product quality as comprehensive evaluation indicators, and construct an evaluation function.

[0077] Application of genetic algorithm: Encode key parameters in the production process, such as crushing particle size, cleaning time, mixing ratio, extrusion temperature and speed, into chromosomes. The initial population randomly generates 100 individuals. Define the fitness function, taking into account the energy consumption reduction ratio and product quality improvement. In genetic operation, roulette selection method, single point crossover and uniform mutation are used for selection.

[0078] Optimization plan formulation and implementation: After 500 generations of genetic evolution, the optimized parameter combination was obtained. For example, adjusting the crushing particle size makes subsequent cleaning more efficient and reduces energy consumption; optimizing the mixing ratio combined with adjusting the extrusion temperature and speed improves the quality of the board. During the implementation process, each parameter is gradually adjusted, and energy consumption and product quality are monitored in real time.

[0079] Example 6

[0080] Data collection and model building: Collect energy consumption data, operating hours, and maintenance records of various types of equipment, such as injection molding machines, blow molding machines, and dryers, for a period of two months. Build a model of the relationship between equipment energy consumption and operating time, taking into account factors such as equipment startup energy consumption, operating energy consumption, and standby energy consumption.

[0081] Application of simulated annealing algorithm: Encode the startup time, running time, downtime, etc. of the equipment into a state vector. The initial state is randomly generated, and the initial temperature, cooling rate and termination temperature are set. The objective function is defined as the minimum total energy consumption. During the search process, according to the rules of the simulated annealing algorithm, the inferior solution is accepted with a certain probability and the local optimum is jumped out.

[0082] Optimization plan formulation and implementation: After multiple iterations, the optimized equipment operation time arrangement is obtained. For example, equipment is started at different times to avoid concentrated power consumption peaks; equipment downtime is reasonably arranged to reduce standby energy consumption. During the implementation process, the equipment operation time is adjusted through the automatic control system and energy consumption changes are monitored.

[0083] Example 7

[0084] Data collection and environment definition: collect production order information, equipment status, raw material inventory and other data, and update them in real time. Define the state space as current order demand, equipment operation status, raw material inventory level, etc.; the action space is production task allocation, equipment start and stop decision, etc.; the reward function is designed based on factors such as energy consumption reduction, production efficiency improvement, and order completion rate on time.

[0085] Application of reinforcement learning algorithm: Use the deep Q network (DQN) algorithm to build a neural network model. Learn the optimal production scheduling strategy through continuous interaction with the production environment. During the training process, use the experience replay mechanism to improve data utilization and accelerate convergence.

[0086] Optimization plan formulation and implementation: As training progresses, the model gradually learns to make optimal scheduling decisions based on real-time conditions. For example, according to the urgency of the order and the energy consumption of the equipment, production tasks are reasonably allocated, and equipment with low energy consumption and high efficiency is started first. During the implementation process, the reinforcement learning model is integrated into the enterprise production management system to adjust the production schedule in real time.

[0087] Example 8

[0088] Data collection and preprocessing: Collect production energy consumption data, equipment operating parameters, ambient temperature and other data for the past year. Normalize the data and divide it into training and test sets.

[0089] Particle swarm optimization-support vector machine model construction: Use support vector machine (SVM) to build an energy consumption prediction model, and use particle swarm optimization algorithm (PSO) to optimize SVM parameters, such as penalty factor C and kernel function parameter γ. PSO continuously iterates to find the optimal parameter combination to minimize the prediction error of SVM.

[0090] Energy-saving optimization implementation: Use the optimized SVM model to predict future energy consumption, and adjust production plans and equipment parameters in advance based on the prediction results. For example, when it is predicted that energy consumption will increase, check whether the equipment has hidden dangers of failure in advance, and adjust production process parameters, such as reducing extrusion speed, optimizing heating temperature, etc.

[0091] Implementation effect: The accuracy of energy consumption prediction reached over 90%. Through early optimization, energy consumption was reduced by 14%, effectively avoiding the cost increase caused by excessive energy consumption.

[0092] Example 9

[0093] Data collection and model building: Install various sensors on production equipment to collect temperature, pressure, current, speed and other data in real time. Build an autoencoder model based on deep learning to extract features and detect anomalies on the collected data.

[0094] Energy-saving optimization strategy: The automatic encoder learns the data characteristics under normal production conditions. When the data deviates from the normal characteristics, it is judged as abnormal. According to the abnormal type, combined with the expert experience database, the corresponding energy-saving optimization strategy is formulated. For example, when the temperature of the melting equipment is detected to be abnormally high, it is judged that it may be a heating element failure, and maintenance is arranged in time, and the heating parameters are adjusted to reduce energy consumption.

Claims

1. An energy-saving optimization scheme for waste plastic recycling production process based on taboo search algorithm, characterized in that: The following steps are involved: Model the waste plastic recycling production process in detail, divide it into multiple sub-processes, analyze the sources and influencing factors of energy consumption in each sub-process, and establish an energy consumption model. The energy consumption model is constructed based on factors such as equipment power parameters, operating time, and production output to quantify the energy consumption of each sub-process and the entire production process; Designing a taboo search algorithm, including representing the optimization scheme of the waste plastic recycling production process as a solution vector, where each element in the solution vector corresponds to a production process parameter; generating an initial solution by random generation or experience-based method; The neighborhood structure is defined, and the neighborhood solutions are generated by making slight adjustments to the elements in the solution vector; a taboo table is established to record the most recently visited solutions to prevent the algorithm from falling into the local optimum; the objective function is to minimize the energy consumption of the entire waste plastic recycling production process, and the objective function is expressed as: Where E represents the total energy consumption, ei represents the energy consumption of the i-th sub-process, wi represents the energy consumption weight of the i-th sub-process, n represents the number of sub-processes, and the objective function comprehensively considers the energy consumption of each sub-process, and other constraints can be added according to actual production needs; Set the search termination condition. When the search termination condition is met, the algorithm stops searching and outputs the optimal solution currently found. According to the optimal solution obtained by the taboo search algorithm, a detailed energy-saving optimization plan for the waste plastic recycling production process is formulated, and the plan is implemented in actual production. The energy consumption and product quality indicators in the production process are monitored in real time. The actual monitoring data are compared with the data before optimization to evaluate the effect of the optimization plan. If the expectation is not met, the energy consumption model, the parameters of the taboo search algorithm or the neighborhood search strategy are adjusted and improved, and the optimization calculation is performed again until a satisfactory energy-saving effect is achieved.

2. The energy-saving optimization scheme for waste plastic recycling production process based on taboo search algorithm according to claim 1 is characterized in that: The waste plastic recycling production process is modeled in detail, specifically including modeling of raw material pretreatment, melt extrusion, and molding processing sub-processes, wherein the raw material pretreatment link considers the power of the cleaning equipment, the operating time, and the relationship between the cleaning time and the raw material characteristics; the melt extrusion link considers factors such as the power of the heating device, the temperature setting, the specific heat capacity of the material, the heating requirements, and the production efficiency; the molding processing link considers the type of molding equipment and the molding process parameters.

3. The energy-saving optimization scheme for waste plastic recycling production process based on taboo search algorithm according to claim 1 is characterized in that: The solution vector code includes key parameters in the production process, each parameter has a corresponding value range, and the value range is determined according to the technical specifications of the equipment and the production process requirements.

4. The energy-saving optimization scheme for waste plastic recycling production process based on taboo search algorithm according to claim 1 is characterized in that: The generation of the initial solution adopts a method combining random generation and empirical values. First, the approximate reasonable range of some parameters is determined based on production experience, and the parameter values ​​in the initial solution vector are randomly generated within this range. After generating multiple initial solutions, the solution with a relatively low energy consumption calculation value is selected as the initial solution.

5. The energy-saving optimization scheme for waste plastic recycling production process based on taboo search algorithm according to claim 1 is characterized in that: The neighborhood search strategy adopts multiple neighborhood search operations, including parameter fine-tuning operation, increasing or decreasing a certain value of a parameter in the solution vector; parameter exchange operation, exchanging the values ​​of two parameters in the solution vector; parameter insertion operation, inserting a parameter in the solution vector into another position.

6. The energy-saving optimization scheme for waste plastic recycling production process based on taboo search algorithm according to claim 1 is characterized in that: The taboo table stores the key features of neighborhood solutions and the corresponding taboo periods. When a new neighborhood solution is generated, it is checked whether the solution is in the taboo table. If it is and the taboo period has not expired, the solution is regarded as a taboo solution unless the unbanning condition is met, that is, the objective function value of the solution is better than the current optimal solution. Otherwise, the solution is not considered. The taboo table is updated after each iteration, the newly accessed solution is added to the taboo table, and the taboo period of the existing taboo solution is shortened. When the taboo period is 0, the solution is removed from the taboo table.

7. The energy-saving optimization scheme for waste plastic recycling production process based on taboo search algorithm according to claim 1 is characterized in that: The search termination condition includes that the algorithm stops when the maximum number of iterations is reached; or if the objective function value does not improve significantly within a certain number of iterations, it is considered that the objective function value has converged and the algorithm stops.

8. The energy-saving optimization scheme for waste plastic recycling production process based on taboo search algorithm according to claim 1 is characterized in that: When implementing the optimization plan in actual production, the operating status of the production equipment, energy consumption, product quality and other indicators are monitored in real time. Relevant data are collected through energy consumption monitoring equipment and quality inspection instruments installed on the equipment. The power consumption and heat energy consumption data of each device are recorded once an hour. At regular intervals, the quality of the recycled plastic products produced is inspected, including dimensional accuracy, strength and appearance.

9. An energy-saving optimization method for a large-scale waste plastic recycling enterprise, based on the waste plastic recycling production process energy-saving optimization scheme based on the taboo search algorithm as described in any one of claims 1 to 8, characterized in that: This includes collecting detailed energy consumption data for various equipment on various production lines within the enterprise over a quarter, and establishing a sophisticated energy consumption model covering multiple links such as raw material pretreatment, melting, molding, and post-processing. The solution vector encoding contains multiple key parameters, and the initial solution is generated by fine-tuning the experience values ​​of senior engineers in the enterprise. After calculating the energy consumption, the optimal one is selected, and a diversified strategy is adopted in the neighborhood search, and a dynamic update strategy is adopted for the taboo table. After multiple iterations of the taboo search algorithm, an optimization solution was obtained, and the enterprise digital management system was used to monitor the equipment operating status and energy consumption data in real time.

10. An energy-saving optimization method for a medium-sized waste plastic recycling enterprise, based on the waste plastic recycling production process energy-saving optimization scheme based on the taboo search algorithm described in any one of claims 1 to 8, characterized in that: According to the production characteristics of the enterprise, we focus on collecting energy consumption and production data of specific links and establishing energy consumption models of corresponding links; the solution vector encoding revolves around specific equipment parameters, the initial solution is randomly generated and combined with the company's historical production data to screen out the better one, the neighborhood search adopts a combination of parameter fine-tuning and local parameter combination adjustment, and the taboo table setting sets the taboo period differently according to the importance of the parameters and the degree of impact on energy consumption; the algorithm is iterated a certain number of times to obtain the optimization plan, and regular maintenance and calibration of equipment are strengthened during implementation.

Images