A multi-objective evolutionary energy medium configuration method

Abstract

The present invention relates to a multi-objective evolutionary energy medium configuration method, with the goal of minimizing cost and energy consumption, including: 1: obtaining energy-related data in steel production, determining key nodes, constructing a network topology based on energy flow, and establishing key node constraints; 2: obtaining the production task of steel output and determining the calculation cycle; 3: considering the time-sharing electricity price characteristics and the buffering effect of the gas tank, establishing an energy medium configuration model with minimizing cost and energy consumption as the objective function; 4: improving the decomposition-based multi-objective evolutionary algorithm MOEA / D in the decoding operation and neighborhood update steps respectively, and using the improved decomposition-based multi-objective evolutionary algorithm to optimize and solve the model to obtain the optimal energy medium configuration scheme. The present invention uniformly regulates the energy media such as gas, steam and electricity involved in the system at the generation end, consumption end and storage end, realizes efficient conversion between various energy media, and effectively reduces production costs and energy consumption.

Description

Technical Field

[0001] The invention belongs to the field of energy configuration in steel production, and specifically is an energy medium configuration method with the goal of minimizing cost and energy consumption. Background Art

[0002] As the concept of sustainable development gradually takes root in people's minds, coupled with the current situation of increasing energy shortage, energy conservation and consumption reduction are imperative. However, as a typical energy-intensive industry, my country's steel industry urgently needs to conduct in-depth research on how to reduce steel energy consumption and production costs, so as to maintain economic stability while reducing energy consumption and achieving green and sustainable development.

[0003] The main energy flow path of steel is to recycle the rich gas and waste heat and pressure generated in the production process, and convert them into steam and electricity required for the production process through energy conversion devices to achieve energy reuse. Energy medium configuration is to analyze the on-site energy production and consumption, and reasonably mobilize energy according to the changes in supply and demand and the operating status of equipment, reduce energy dispersion, and improve energy utilization efficiency. Therefore, formulating an energy medium configuration plan that meets the current production status is the basis for energy optimization and allocation, and plays a vital role in energy conservation and consumption reduction. However, due to the characteristics of steel with more variables and complex constraints, multi-objective energy medium configuration has the advantages of difficult optimization and great energy-saving potential. However, there are few related studies, and it is urgent to conduct in-depth discussions on such issues.

[0004] The energy medium configuration method proposed in the present invention fully considers the whole process of steel production, the time-of-use electricity price characteristics and the buffering effect of gas tanks, constructs an energy medium configuration model with minimizing operating costs and total energy consumption as objective functions, and combines a decomposition-based multi-objective evolutionary algorithm with adaptive neighborhood updates to formulate and optimize energy medium configuration plans to reduce costs and energy consumption. Summary of the invention

[0005] The purpose of the present invention is to provide an energy medium configuration method with the goal of minimizing cost and energy consumption, so as to solve the problems in the existing energy medium configuration technology, such as the large number of steel energy medium configuration variables, complex constraints and high difficulty in model solving.

[0006] The energy medium configuration method of the present invention comprises the following steps:

[0007] Step 1: Obtain energy-related data in steel production, identify key nodes, build a network topology based on energy flow, and establish key node constraints;

[0008] Step 2: Obtain the production task of steel output and determine the calculation cycle;

[0009] Step 3: Considering the time-of-use electricity price characteristics and the buffering effect of the gas cabinet, an energy medium configuration model is established with minimizing cost and energy consumption as the objective function. The energy medium includes gas, steam, and electricity.

[0010] Step 4: According to the characteristics of the energy medium configuration model, the decomposition-based multi-objective evolutionary algorithm MOEA / D is improved in the decoding operation and neighborhood update steps respectively, and the improved decomposition-based multi-objective evolutionary algorithm is used to optimize and solve the model to obtain the optimal energy medium configuration plan.

[0011] The energy medium configuration of steel production in step one involves a production process subsystem, a multi-media storage subsystem, a multi-media conversion subsystem and a multi-media transportation subsystem. The key processes included in the production process subsystem are: sintering, coking, steelmaking, ironmaking, hot rolling and cold rolling. The key equipment in the multi-media storage subsystem includes: blast furnace gas holder, converter gas holder, coke oven gas holder and corresponding discharge tower. The key equipment in the multi-media conversion subsystem includes: generator set, steam turbine, combustion boiler and waste heat recovery device. The multi-media transportation subsystem includes: gas pipeline network, steam pipeline network and electric energy medium pipeline network; accordingly, the key nodes in the network topology structure based on energy flow include: key processes in the production process subsystem, key equipment in the multi-media storage subsystem, and key equipment in the multi-media conversion subsystem. The energy medium relationship between these key nodes and the multi-media transportation subsystem constitutes a network topology structure diagram.

[0012] Step 1: Obtain the historical data of energy production and consumption of each node under normal operation in the past multiple cycles from the energy database of the steel enterprise, and perform data cleaning, including: eliminating missing data and data information of abnormal production scenarios.

[0013] Step one also establishes process constraints, energy balance constraints and logistics balance constraints for key equipment in energy conversion and storage, including key equipment in the multimedia storage subsystem and key equipment in the multimedia conversion subsystem.

[0014] Step 2 is to obtain the current production tasks from the production system, including work plans and maintenance plans, clarify the output requirements in the production tasks, make accident assumptions and operating condition assumptions, and determine the production process path and process unit selection.

[0015] The calculation period is determined to be one day, and the 24 consecutive hours of one day are discretized and divided into T time periods to simplify the calculation complexity; considering the energy response characteristics of steel, T is set to an integer of 12 to 36.

[0016] In the objective function of step three, the target cost is set as the sum of the manufacturing cost of purchased coal, purchased natural gas, steam-producing equipment, equipment maintenance cost of power generation equipment, purchased electricity cost and external power transmission income; the target energy consumption is set as the sum of the consumption of purchased energy, secondary energy, waste heat energy, external energy and inventory change energy of the three energy media of gas, steam and electricity.

[0017] The constraints of the energy medium configuration model, in addition to the process constraints, energy balance constraints and logistics balance constraints established in step one, also include: process saturation constraints, energy medium supply and demand balance constraints, mixed gas calorific value constraints, non-negative constraints on all variables, constraints that purchased electricity and transmitted electricity cannot be greater than 0 at the same time, and the constraint that the actual output is greater than or equal to the output demand described in step two.

[0018] The decision variable of the energy medium configuration model is the process saturation, which represents the production and operation status of each process. The process saturation is the ratio of the current process production to the maximum production; the maximum production is the maximum production that can be achieved under comprehensive consideration of product mix, labor, plant, equipment, raw materials, etc.; the process saturation constraint reflects the production equipment limitation and the coordination requirements of the upper and lower processes, as shown in the following formula:

[0019]

[0020]

[0021] in, represents the process saturation of process i at time t, and Respectively represent the minimum and maximum values; p i+1 is the ratio coefficient of the output of process i+1 product to the output of process i product under normal operation, q i For normal operation, the process i The ratio coefficient of the product output and the process saturation of process i, is the inventory of product i at time t.

[0022] Step 4: In the decoding operation and neighborhood update step of the decomposition-based multi-objective evolutionary algorithm MOEA / D, a decoding mechanism based on energy consumption rules and a neighborhood adaptive update based on the population evolution degree are introduced to improve the algorithm and improve the convergence ability of the algorithm. The improved decomposition-based multi-objective evolutionary algorithm includes the following steps:

[0023] Step 4.1, algorithm initialization: determine the decision variables, algorithm parameters, and weight vector set; set multiple groups of random number sequences to generate a solution in the initial population of the corresponding energy medium configuration model; calculate the target cost and target energy consumption of the energy medium configuration model according to the decoding mechanism based on the energy production and consumption rules; take the minimum cost and minimum energy consumption as the initial ideal point; record the non-dominated optimal solution set in the initial population of the energy medium configuration model as an external archive;

[0024] The decoding mechanism based on energy production and consumption rules regards the key equipment in the multi-media storage subsystem and the multi-media conversion subsystem as an energy system, eliminating the energy medium allocation variables of each equipment; the energy production and consumption rules include:

[0025] (1) Set the gas usage priority: blast furnace gas > coke oven gas > converter gas;

[0026] (2) Set the medium generation priority: steam production > electricity production;

[0027] (3) Equipment power generation rules: Compare the self-generation cost of power generation equipment with the current power purchase cost and power sales revenue, and divide it into Class I, Class II, and Class III equipment according to the power generation cost from high to low, that is, represents the power generation cost of Class I power generation equipment, represents the electricity purchase cost at time t, represents the power generation cost of Class II power generation equipment, represents the electricity sales revenue at time t, Indicates the power generation cost of Class III power generation equipment; Class I power generation equipment generates electricity at the minimum output; Class II power generation equipment changes according to the satisfaction of power demand: if the power supply is less than the demand, the minimum value of the power supply and demand difference and the maximum output of Class II power generation equipment is used for power generation; if the power supply is greater than or equal to the demand, Class II power generation equipment generates electricity at the minimum output; Class III power generation equipment generates electricity at the maximum output;

[0028] The decoding mechanism based on the above energy production and consumption rules includes: calculating the supply of surplus gas and the demand for steam and electricity of each process at the current moment according to the saturation of each process at each moment and the historical data of energy production and consumption; using the energy production and consumption rules to calculate the maximum gas consumption, maximum steam production and maximum electricity production of the energy system; determining the amount of purchased energy and delivered energy at each moment, and obtaining the target cost and target energy consumption of the energy medium configuration model;

[0029] Step 4.2, evolutionary operation: randomly select two numbers from the neighborhood, use the DE crossover operator and the polynomial mutation operator to perform crossover and mutation operations respectively, and generate a new offspring energy medium configuration solution;

[0030] Step 4.3, constraint processing: For out-of-bounds values, the difference between the out-of-bounds parameter value of the infeasible solution and the boundary value is scaled to make the out-of-bounds value return to the value near the boundary;

[0031] Step 4.4, decoding operation: calculating the target cost and target energy consumption of the energy medium configuration model according to the decoding mechanism based on the energy production and consumption rules in step 4.1;

[0032] Step 4.5, decomposition operation: decompose the cost and energy consumption in the multi-objective of the energy medium configuration model into quantifiable single objective values ​​according to the normalized Chebyshev aggregation function and weight vector;

[0033] Step 4.6, ideal point update: update the ideal point according to the latest minimum cost and minimum energy consumption;

[0034] Step 4.7, neighborhood update: introduce the population evolution degree information, and adaptively update the neighborhood of each sub-problem; the coefficients that express the population evolution degree include the relative decline rate of the sub-problem and the basic neighborhood size of the population. By introducing these two coefficients, the neighborhood size of the sub-problem is adaptively updated to enhance the guiding role of the individual convergence process of the population;

[0035] Step 4.8, update the external archive: update the external archive according to the non-dominated solutions in the offspring energy medium configuration solution;

[0036] Step 4.9: If the number of iterations reaches the set maximum value, stop the calculation and output the external file; otherwise, jump to step 4.2;

[0037] Step 4.10: Output the optimal energy medium configuration plan: According to the external archives and decision maker preferences, select the optimal energy medium configuration plan from the external archives for output.

[0038] The advantages of the present invention are: the present method can effectively realize the optimization of energy medium configuration. Compared with the current situation of energy allocation and control, the present invention uniformly regulates the energy media such as gas, steam and electricity involved in the system at the generation end, consumption end and storage end to realize efficient conversion between various energy media. In view of the problems of many variables, complex constraints and high difficulty in model solving in the multi-media steel energy planning model, the existing multi-objective evolutionary algorithm based on decomposition is improved. In the decoding operation and neighborhood update steps of the multi-objective evolutionary algorithm based on decomposition, a decoding mechanism based on energy production and consumption rules and a neighborhood adaptive update based on the degree of population evolution are introduced to improve the algorithm, thereby improving the convergence ability of the algorithm. The obtained intra-day energy medium configuration plan can effectively reduce the cost and energy consumption required for production and improve production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is the overall flow chart of the present invention.

[0040] Figure 2 This is the network topology diagram of a steel enterprise. DETAILED DESCRIPTION

[0041] The present invention is further described in detail below with reference to the accompanying drawings and embodiments.

[0042] like Figure 1 As shown, the present invention generally includes 4 steps:

[0043] Step 1: Structure determination: Obtain steel energy-related data from the steel database, determine key nodes, build a network topology based on energy flow, and establish key node constraints.

[0044] Step 2: Task input: Obtain the production task of steel output and determine the calculation cycle.

[0045] Step 3: Model building. Considering the time-of-use electricity price characteristics and the buffering effect of gas cabinets, an energy medium configuration model is established with minimizing cost and energy consumption as the objective function.

[0046] Step 4: Intelligent optimization. According to the characteristics of the energy medium configuration model, the decomposition-based multi-objective evolutionary algorithm MOEA / D is improved in the decoding operation and neighborhood update steps respectively, and the improved decomposition-based multi-objective evolutionary algorithm is used to optimize and solve the model to obtain the optimal energy medium configuration solution.

[0047] In step one, the steel energy medium configuration system mainly includes: production process subsystem, multi-media storage subsystem, multi-media conversion subsystem and multi-media transportation subsystem. The production process subsystem uses natural resources as raw materials, goes through multiple processes, and finally obtains the final product. During this period, a lot of recyclable resources will be generated, and the main recycled energy is rich gas, waste heat and electricity. Rich gas is temporarily stored through the multi-media storage subsystem or converted into energy (steam and electricity) required by the production process subsystem through the multi-media conversion subsystem. Waste heat is directly recovered through the multi-media conversion subsystem. The steel energy medium configuration system will realize the energy transmission between each subsystem and the interaction with the external power grid through the multi-media transportation subsystem. If the supply of electricity is greater than the demand, it will be sent to the power grid to obtain benefits; if the supply of electricity is less than the demand, it will purchase electricity from the power grid to meet production needs.

[0048] In the network topology based on energy flow, key nodes include: key processes in the steel production process subsystem (sintering, coking, steelmaking, ironmaking, hot rolling and cold rolling), key equipment in the multi-media storage subsystem (blast furnace gas tank, converter gas tank, coke oven gas tank and corresponding discharge tower) and key equipment in the multi-media conversion subsystem (generator set, gas turbine, combustion boiler and waste heat recovery device). The energy medium relationship between these key nodes and the multi-media transportation subsystem (gas pipeline network, steam pipeline network and power energy medium pipeline network) constitutes the network topology diagram of the steel enterprise based on energy flow, such as Figure 2 shown.

[0049] The present invention considers the key equipment in the multi-media storage subsystem and the multi-media conversion subsystem as an energy system. Compared with the current situation of energy allocation and control, the energy media such as gas, steam and electricity involved in the system can be uniformly regulated at the generation end, consumption end and storage end to achieve efficient conversion between various energy media.

[0050] Then, the energy consumption history data of each node under normal operation in the past multiple cycles is obtained from the energy database of the steel enterprise, and data cleaning is performed, including: eliminating missing data and data information of abnormal production scenarios. Process constraints, energy balance constraints and logistics balance constraints are established for key equipment in energy conversion and storage (combustion boilers, steam turbines, generator sets and waste heat recovery devices, blast furnace gas holders, converter gas holders, coke oven gas holders and corresponding discharge towers).

[0051] Step 2 obtains the current stage of production tasks (operation plans and maintenance plans) from the production system, clarifies the output requirements in the production tasks (the actual output must be greater than or equal to the output requirements), makes accident assumptions and working condition assumptions, and determines the production process path and process units. Determine the calculation cycle as days, and discretize the 24 consecutive hours of a day, dividing it into T time periods to simplify the calculation complexity. T is preferably set to an integer of 12 to 36. In the embodiment, T=24, and a day is divided into 24 time periods, with each hour as the minimum cycle, for a total of 24 cycles. Considering that the energy allocation response cycle of gas pipelines, steam pipelines, and self-generating units is 20 to 40 minutes, selecting 1 hour as the time granularity can simplify the simulation complexity while being close to actual operation.

[0052] In step 3, considering the characteristics of time-of-use electricity prices and the buffering effect of gas tanks, an energy medium configuration model is established with the objective function of minimizing cost and energy consumption. Energy media include gas, steam, and dielectrics. The time-of-use electricity prices are shown in Table 1. The external power price does not change with different electricity consumption periods. However, the external power price is higher in peak hours than in normal hours, and higher than the external power price in valley hours.

[0053] Table 1 Time-of-use electricity price

[0054]

[0055]

[0056] The gas tank has the function of storing gas. When the gas supply exceeds the demand, it can absorb a certain amount of gas to reduce gas emission. When the gas supply is less than the demand, it can release a certain amount of gas to maintain stable and sustainable production. The gas capacity constraints are shown in Table 2.

[0057] Table 2 Gas tank capacity constraints

[0058]

[0059] In the objective function of the energy medium configuration model of the present invention, the target cost is set as the sum of the manufacturing cost of purchased coal, purchased natural gas, steam-producing equipment, the equipment maintenance cost of power generation equipment, the cost of purchased electricity (positive value) and the income from externally transmitted electricity (negative value); the target energy consumption is set as the sum of the consumption of purchased energy (energy purchased from outside the energy system), secondary energy (usable gas and steam energy recovered from the production process subsystem), waste heat energy (usable waste heat and waste pressure energy recovered from the production process subsystem), external energy (energy transmitted from the energy system to the outside world) and inventory change energy (change in stored energy in the production process subsystem and the multi-media storage subsystem) of the three energy media of gas, steam and electricity.

[0060] In addition to the process constraints, energy balance constraints and logistics balance constraints described in step one, the constraints of the energy medium configuration model also include: process saturation constraints, energy medium supply and demand balance constraints, mixed gas calorific value constraints, non-negative constraints on all variables, constraints that purchased electricity and delivered electricity cannot be greater than 0 at the same time, and the constraint that the actual output is greater than or equal to the output demand described in step two.

[0061] The decision variable of the energy medium configuration model of the present invention is the process saturation, which is defined as the ratio of the current process production to the maximum production, representing the production operation of each process. The maximum production is the maximum production that can be achieved by comprehensively considering the product portfolio, labor, plant, equipment, raw materials, etc. The process saturation constraint reflects the production equipment limitation and the coordination requirements of the upper and lower processes, as shown in the following formula:

[0062]

[0063]

[0064] in, represents the process saturation of process i at time t, and Represent their minimum and maximum values ​​respectively, p i+1 is the ratio coefficient of the output of process i+1 product to the output of process i product under normal operation, q i is the ratio coefficient of the product output of process i under normal operation to the process saturation of process i, is the inventory of product i at time t.

[0065] In step four, in view of the fact that the energy medium configuration model has many variables, complex constraints and high difficulty in solving the model, a multi-objective evolutionary algorithm based on decomposition (MOEA / D) is selected and improved to optimize the solution. In the decoding operation and neighborhood update steps of the multi-objective evolutionary algorithm based on decomposition, a decoding mechanism based on energy production and consumption rules and a neighborhood adaptive update based on the degree of population evolution are introduced to improve the algorithm and improve the convergence ability of the algorithm. In the embodiment, the population size of MOEA / D is set in the range of 80 to 120, the number of iterations is set to 1000 to 1500, the crossover probability is 1, the mutation probability is set in the range of 0.1 to 0.3, the evolution degree threshold is set to 0.01 to 0.1, and the population neighborhood size selection range is 8 to 20. The improved multi-objective evolutionary algorithm based on decomposition includes the following steps:

[0066] Step 4.1: Algorithm initialization. Determine the decision variables, algorithm parameters, and weight vector set; set multiple sets of random number sequences to generate a solution in the initial population of energy medium configuration; calculate the target value cost and energy consumption of the energy medium configuration model according to the decoding mechanism of energy production and consumption rules; take the minimum cost and minimum energy consumption as the initial ideal point; record the non-dominated optimal solution set in the initial population of energy medium configuration as an external archive.

[0067] Step 4.2: Evolution operation. Randomly select two numbers from the neighborhood, use the DE crossover operator and polynomial mutation operator to perform crossover and mutation operations respectively, and generate a new offspring energy medium configuration solution.

[0068] Step 4.3: Constraint processing: For out-of-bounds values, the difference between the out-of-bounds parameter value of the infeasible solution and the boundary value is scaled to make the out-of-bounds value return to the value near the boundary, and the search moves towards the feasible region.

[0069] Step 4.4: Decoding operation. Calculate the target value cost and energy consumption of the energy medium configuration model according to the decoding mechanism of the energy production and consumption rules.

[0070] The present invention introduces a decoding mechanism based on energy production and consumption rules, treats the key equipment in the multi-media storage subsystem and the multi-media conversion subsystem as an energy system, eliminates the energy medium allocation variables of each device, simplifies the target value calculation complexity, solves the problems of long steel process, numerous production and consumption equipment, and difficult energy medium allocation, and improves the search efficiency of the algorithm. The energy production and consumption rules are as follows:

[0071] (1) Medium consumption rules. Based on the principle of minimum gas emission and the abundant gas situation of the steel plant in this paper, the gas use priority is set as follows: blast furnace gas > coke oven gas > converter gas.

[0072] (2) Medium production rules. Considering the non-purchasability of steam in this system, the medium generation priority is designed as follows: steam production > electricity production.

[0073] (3) Equipment power generation rules. The self-generation cost of power generation equipment is compared with the current power purchase cost and power sales revenue, and the equipment is divided into Class I to Class III according to the power generation cost from high to low, that is, represents the power generation cost of Class I power generation equipment, represents the electricity purchase cost at time t, represents the power generation cost of Class II power generation equipment, represents the electricity sales revenue at time t, Indicates the power generation cost of Class III power generation equipment. Class I power generation equipment generates electricity at the minimum output. Class II power generation equipment changes according to the satisfaction of power demand: if the power supply is less than the demand, the minimum value of the power supply and demand difference and the maximum output of Class II equipment is used for power generation; if the power supply is greater than or equal to the demand, Class II equipment generates electricity at the minimum output. Class III power generation equipment generates electricity at the maximum output.

[0074] The decoding mechanism based on energy production and consumption rules includes: calculating the supply of each surplus gas volume and the demand for steam and electricity at the current moment according to the decision variables (saturation of each process at each moment) and the historical data of energy production and consumption; using energy production and consumption rules to calculate the maximum gas consumption, maximum steam production and maximum electricity production of the energy system; determining the amount of purchased energy and delivered energy at each moment, and obtaining the target cost and target energy consumption of the final energy medium configuration model.

[0075] Step 4.5: Decomposition operation: Decompose the cost and energy consumption in the multi-objective energy medium configuration model into quantifiable single objective values ​​according to the normalized Chebyshev aggregation function and weight vector.

[0076] Step 4.6: Ideal point update: Update the ideal point according to the latest minimum cost and minimum energy consumption.

[0077] Step 4.7: Neighborhood update: Introduce the population evolution information and adaptively update the neighborhood of each sub-problem.

[0078] In this invention, by introducing the relative decline rate coefficient of the subproblem and the population basic neighborhood scale coefficient, the neighborhood scale of the subproblem is adaptively updated, the guiding role of the population individual convergence process is enhanced, and the convergence ability of the algorithm is effectively improved. The specific calculation is as follows:

[0079] Define the relative decrease rate of the aggregate value of subproblem n as △ n , as shown below:

[0080]

[0081] in, represents the nth solution in the population with the number of iterations g, λ is the weight vector, z * For the ideal point, Represents the aggregation function value of the g-th generation subproblem.

[0082] Calculate the size of the population basic neighborhood T g , as shown below:

[0083]

[0084] Among them, T max and T min are the maximum and minimum values ​​of the population neighborhood size, g represents the number of iterations, and Maxgen represents the maximum number of iterations set.

[0085] Adaptively adjust the neighborhood size corresponding to the subproblem based on the relative decline rate of the subproblem and the population base neighborhood size If the evolutionary pace of the previous generation of individuals is large, the neighborhood size of the current generation of individuals should be appropriately reduced so that the individuals can conduct breadth search at the new convergence point and the population can conduct search in all directions, as shown in the following formula:

[0086]

[0087] Among them, mod(·) function represents the rounding function, and ε is the evolution degree threshold.

[0088] Step 4.8: Update the external archive. Update the external archive according to the non-dominated solutions in the offspring energy medium configuration solution.

[0089] Step 4.9: Termination judgment. If the number of iterations reaches the maximum value, stop the calculation and output the external file. Otherwise, jump to step 4.2.

[0090] Step 4.10: Output the optimal energy medium configuration plan. According to the external archives and decision maker preferences, select the optimal energy medium configuration plan from the external archives and send it to the steel database platform for result display. The energy medium configuration plan includes: the surplus gas supply, steam and electricity demand generated by the production process subsystem in each cycle, and the amount of purchased energy, secondary energy, waste heat energy, external energy and inventory change energy of each energy medium.

Claims

1. A multi-objective evolutionary energy medium configuration method, characterized in that: The following steps are involved: Step 1: Obtain energy-related data in steel production, identify key nodes, build a network topology based on energy flow, and establish key node constraints; Step 2: Obtain the production task of steel output and determine the calculation cycle; Step 3: Considering the time-of-use electricity price characteristics and the buffering effect of the gas cabinet, an energy medium configuration model is established with minimizing cost and energy consumption as the objective function. The energy medium includes gas, steam, and electricity. Step 4: According to the characteristics of the energy medium configuration model, the decomposition-based multi-objective evolutionary algorithm MOEA / D is improved in the decoding operation and the neighborhood update step respectively, and the improved decomposition-based multi-objective evolutionary algorithm is used to optimize and solve the model to obtain the optimal energy medium configuration plan; Step 4: In the decoding operation and neighborhood update step of the decomposition-based multi-objective evolutionary algorithm MOEA / D, a decoding mechanism based on energy consumption rules and a neighborhood adaptive update based on the population evolution degree are introduced to improve the algorithm and improve the convergence ability of the algorithm. The improved decomposition-based multi-objective evolutionary algorithm includes the following steps: Step 4.1, algorithm initialization: determine the decision variables, algorithm parameters, and weight vector set; set multiple groups of random number sequences to generate a solution in the initial population of the corresponding energy medium configuration model; calculate the target cost and target energy consumption of the energy medium configuration model according to the decoding mechanism based on the energy production and consumption rules; take the minimum cost and minimum energy consumption as the initial ideal point; record the non-dominated optimal solution set in the initial population of the energy medium configuration model as an external archive; The decoding mechanism based on energy production and consumption rules regards the key equipment in the multi-media storage subsystem and the multi-media conversion subsystem as an energy system, eliminating the energy medium allocation variables of each equipment; the energy production and consumption rules include: (1) Set the gas usage priority: blast furnace gas > coke oven gas > converter gas; (2) Set the medium generation priority: steam production > electricity production; (3) Equipment power generation rules: Compare the self-generation cost of power generation equipment with the current power purchase cost and power sales revenue, and divide it into Class I, Class II, and Class III equipment according to the power generation cost from high to low, that is, represents the power generation cost of Class I power generation equipment, represents the electricity purchase cost at time t, represents the power generation cost of Class II power generation equipment, represents the electricity sales revenue at time t, Indicates the power generation cost of Class III power generation equipment; Class I power generation equipment generates electricity at the minimum output; Class II power generation equipment changes according to the satisfaction of power demand: if the power supply is less than the demand, the minimum value of the power supply and demand difference and the maximum output of Class II power generation equipment is used for power generation; if the power supply is greater than or equal to the demand, Class II power generation equipment generates electricity at the minimum output; Class III power generation equipment generates electricity at the maximum output; The decoding mechanism based on the above energy production and consumption rules includes: calculating the supply of surplus gas and the demand for steam and electricity of each process at the current moment according to the saturation of each process at each moment and the historical data of energy production and consumption; using the energy production and consumption rules to calculate the maximum gas consumption, maximum steam production and maximum electricity production of the energy system; determining the amount of purchased energy and delivered energy at each moment, and obtaining the target cost and target energy consumption of the energy medium configuration model; Step 4.2, evolutionary operation: randomly select two numbers from the neighborhood, use the DE crossover operator and the polynomial mutation operator to perform crossover and mutation operations respectively, and generate a new offspring energy medium configuration solution; Step 4.3, constraint processing: For out-of-bounds values, the difference between the out-of-bounds parameter value of the infeasible solution and the boundary value is scaled to make the out-of-bounds value return to the value near the boundary; Step 4.4, decoding operation: calculating the target cost and target energy consumption of the energy medium configuration model according to the decoding mechanism based on the energy production and consumption rules in step 4.1; Step 4.5, decomposition operation: decompose the cost and energy consumption in the multi-objective of the energy medium configuration model into quantifiable single objective values ​​according to the normalized Chebyshev aggregation function and weight vector; Step 4.6, ideal point update: update the ideal point according to the latest minimum cost and minimum energy consumption; Step 4.7, neighborhood update: introduce the population evolution degree information, and adaptively update the neighborhood of each sub-problem; the coefficients that express the population evolution degree include the relative decline rate of the sub-problem and the basic neighborhood size of the population. By introducing these two coefficients, the neighborhood size of the sub-problem is adaptively updated to enhance the guiding role of the individual convergence process of the population; By introducing the relative decline rate coefficient of the sub-problem and the population basic neighborhood size coefficient, the neighborhood size of the sub-problem is adaptively updated. The specific calculation is as follows: Define the relative decrease rate of the aggregate value of subproblem n as Δ n , as shown below: in, represents the nth solution in the population with the number of iterations g, λ is the weight vector, z * For the ideal point, Represents the aggregate function value of the g-th generation subproblem; Calculate the size of the population basic neighborhood T g , as shown below: Among them, T max and T min are the maximum and minimum values ​​of the population neighborhood size, g represents the number of iterations, and Maxgen represents the maximum number of iterations set; Adaptively adjust the neighborhood size corresponding to the subproblem based on the relative decline rate of the subproblem and the population base neighborhood size If the evolutionary pace of the previous generation of individuals is large, the neighborhood size of the current generation of individuals should be appropriately reduced so that the individuals can conduct breadth search at the new convergence point and the population can conduct search in all directions, as shown in the following formula: Among them, mod(·) function represents the rounding function, ε is the evolution degree threshold; Step 4.8, update the external archive: update the external archive according to the non-dominated solutions in the offspring energy medium configuration solution; Step 4.9: If the number of iterations reaches the set maximum value, stop the calculation and output the external file; otherwise, jump to step 4.2; Step 4.10: Output the optimal energy medium configuration plan: According to the external archives and decision maker preferences, select the optimal energy medium configuration plan from the external archives for output.

2. A multi-objective evolutionary energy medium configuration method as claimed in claim 1, characterized in that: The energy medium configuration of steel production in step one involves a production process subsystem, a multi-media storage subsystem, a multi-media conversion subsystem and a multi-media transportation subsystem. The key processes included in the production process subsystem are: sintering, coking, steelmaking, ironmaking, hot rolling and cold rolling. The key equipment in the multi-media storage subsystem includes: blast furnace gas holder, converter gas holder, coke oven gas holder and corresponding discharge tower. The key equipment in the multi-media conversion subsystem includes: generator set, steam turbine, combustion boiler and waste heat recovery device. The multi-media transportation subsystem includes: gas pipeline network, steam pipeline network and electric energy medium pipeline network; accordingly, the key nodes in the network topology structure based on energy flow include: key processes in the production process subsystem, key equipment in the multi-media storage subsystem, and key equipment in the multi-media conversion subsystem. The energy medium relationship between these key nodes and the multi-media transportation subsystem constitutes a network topology structure diagram.

3. A multi-objective evolutionary energy medium configuration method as claimed in claim 1, characterized in that: Step 1: Obtain the historical data of energy production and consumption of each node under normal operation in the past multiple cycles from the energy database of the steel enterprise, and perform data cleaning, including: eliminating missing data and data information of abnormal production scenarios.

4. A multi-objective evolutionary energy medium configuration method as claimed in claim 2, characterized in that: Step 1 establishes process constraints, energy balance constraints and logistics balance constraints for key equipment in energy conversion and storage, wherein the key equipment includes key equipment in a multi-media storage subsystem and key equipment in a multi-media conversion subsystem.

5. A multi-objective evolutionary energy medium configuration method as claimed in claim 4, characterized in that: Step 2 is to obtain the current production tasks from the production system, including work plans and maintenance plans, clarify the output requirements in the production tasks, make accident assumptions and operating condition assumptions, and determine the production process path and process unit selection.

6. A multi-objective evolutionary energy medium configuration method as claimed in claim 5, characterized in that: Step 2: Determine the calculation period as day, and discretize the 24 consecutive hours of a day into T time periods to simplify the calculation complexity; considering the energy response characteristics of steel, T is set to an integer of 12 to 36.

7. A multi-objective evolutionary energy medium configuration method as claimed in claim 6, characterized in that: In the objective function of step three, the target cost is set as the sum of the manufacturing cost of purchased coal, purchased natural gas, steam-producing equipment, equipment maintenance cost of power generation equipment, purchased electricity cost and external power transmission income; the target energy consumption is set as the sum of the consumption of purchased energy, secondary energy, waste heat energy, external energy and inventory change energy of the three energy media of gas, steam and electricity.

8. A multi-objective evolutionary energy medium configuration method as claimed in claim 7, characterized in that: In addition to the process constraints, energy balance constraints and logistics balance constraints established in step one, the constraints of the energy medium configuration model in step three also include: process saturation constraints, energy medium supply and demand balance constraints, mixed gas calorific value constraints, non-negative constraints on all variables, constraints that purchased electricity and delivered electricity cannot be greater than 0 at the same time, and the constraint that the actual output is greater than or equal to the output demand described in step two.

9. A multi-objective evolutionary energy medium configuration method as claimed in claim 8, characterized in that: The decision variable of the energy medium configuration model in step three is the process saturation, which represents the production and operation status of each process. The process saturation is the ratio of the current process production to the maximum production; the maximum production is the maximum production that can be achieved by comprehensively considering the product portfolio, labor, plant, equipment, and raw materials; the process saturation constraint reflects the production equipment limitation and the coordination requirements of the upper and lower processes, as shown in the following formula: in, Indicates at time t Process i The process saturation, and Respectively represent the minimum and maximum values; p i+1 For normal operation, the process i+1 Product output and process i The ratio coefficient of product output, q i For normal operation, the process i Product output and process i The ratio coefficient of process saturation, For the moment t Next process i The inventory level of the product.

Images