Multi-objective optimization control method for grate cooler considering efficiency and energy consumption

Abstract

The application discloses a kind of considering efficiency and energy consumption's grate cooler multi-objective optimization control method, comprising:1, on the basis of considering efficiency and energy consumption, establish grate cooler energy efficiency evaluation model;2, based on the establishment of grate cooler energy efficiency evaluation model grate cooler efficiency target and energy consumption target, construct grate cooler multi-objective optimization model;3, utilize MOEA / D algorithm to solve grate cooler multi-objective optimization model, obtain the optimal solution of grate pressure setting value and fan air volume setting value;4, using generalized predictive control algorithm and incremental PID algorithm realizes the rolling optimization of grate pressure and fan air volume.The application is under the premise of considering the efficiency and energy consumption of grate cooler, to grate cooler grate pressure setting value and fan air volume setting value are optimized simultaneously, it is favorable to improve the parameter control level of grate cooler, reduce production energy consumption.

Description

Technical Field

[0001] This invention belongs to the field of grate cooler control technology, specifically a multi-objective optimization control method for grate coolers that considers efficiency and energy consumption. Background Technology

[0002] As one of the main pieces of equipment in cement clinker production, the grate cooler is responsible for cooling the cement clinker and recovering high-temperature heat. Its operation directly affects clinker quality and production energy consumption. To improve grate cooler efficiency and reduce production energy consumption, operators control the grate cooler's material layer thickness and adjust the cooling air volume to maximize the secondary and tertiary air temperatures and lower the outlet clinker temperature, ensuring sufficient cooling of the clinker while maximizing heat recovery. However, the parameter control of the grate cooler system is relatively complex, relying mainly on manual experience, resulting in a generally low level of overall control. Furthermore, due to varying operator skill levels, the grate cooler's operation cannot be maintained stably and continuously, the clinker cooling process cannot be effectively controlled, and the system's energy efficiency cannot reach its optimal level.

[0003] The selection of objectives is one of the key issues in the multi-objective optimization control of grate coolers. Currently, most multi-objective optimization studies on grate coolers focus on efficiency, rarely considering energy consumption. In actual production, both efficiency and energy consumption need to be considered simultaneously to ensure the rationality of parameter adjustments. Furthermore, in the process of optimizing grate cooler parameters, the directly controllable quantities are often used as optimization variables. When the operating conditions of the grate cooler change, these control quantities may fluctuate drastically, failing to maintain the stable operation of the grate cooler system. Summary of the Invention

[0004] To address the aforementioned problems, this invention proposes a multi-objective optimization control method for grate coolers that considers both efficiency and energy consumption. This method aims to simultaneously optimize the setpoints for the grate pressure and fan airflow of the grate cooler while meeting both efficiency and energy consumption targets. This will improve the parameter control level of the grate cooler, reduce production energy consumption, and ensure the efficient operation of the grate cooler.

[0005] To achieve the above-mentioned objectives, the present invention adopts the following technical solution:

[0006] This invention discloses a multi-objective optimization control method for a grate cooler that considers efficiency and energy consumption. The grate cooler has n cooling fans, of which m fans have freely adjustable airflow rates, and are denoted as FA = {FA1, FA2, ..., FA...}. l ,…,FA m}, FA l This indicates that the l-th fan, 1≤l≤m, and the airflow of the remaining fans are constant or controlled by the operator; its characteristic is that the multi-objective optimization control method of the grate cooler is carried out according to the following steps:

[0007] Step 1: Establish an energy efficiency evaluation model for the grate cooler, taking into account efficiency and energy consumption;

[0008] Step 1.1, Data Acquisition:

[0009] The operating data of the grate cooler is collected in N time periods of length t, and the average operating data of the grate cooler in the k-th time period is calculated and denoted as {y1(k),y2(k),y3(k),u1(k),u2(k),…,u l (k),…,u m (k),u m+1 (k),u m+2 (k)}, k=1,2,…,N, where y1(k) represents the average secondary air temperature of the grate cooler in the k-th time period, y2(k) represents the average tertiary air temperature of the grate cooler in the k-th time period, y3(k) represents the average outlet clinker temperature of the grate cooler in the k-th time period, u l (k) represents the l-th fan FA of the grate cooler in the k-th time period. l Average fan air volume, u m+1 (k) represents the average grate pressure of the grate cooler in the k-th time period, u m+2 (k) represents the average raw material feed rate of the grate cooler in the kth time period;

[0010] Calculate the power consumption of the grate cooler in the kth time period, denoted as y4(k);

[0011] Step 1.2: Process the grate cooler operating data for the k-th time period using the moving average filtering method to obtain the filtered grate cooler operating data. in, This represents the secondary wind temperature filter value for the k-th time period. This represents the filtered values ​​of the third wind temperature in the k-th time period. This represents the filtered value of the clinker outlet temperature in the k-th time period. This represents the power consumption filter value for the k-th time period. This indicates the l-th fan FA of the grate cooler during the k-th time period. l Fan airflow filter value, This represents the filter value of the grate pressure in the k-th time period. This represents the filtered value of raw material feed rate for the k-th time period;

[0012] Step 1.3: Based on the filtered grate cooler operating data, establish an energy efficiency evaluation model for the grate cooler:

[0013] by and The i-th output y of the energy efficiency evaluation model is respectively i (k), i = 1, 2, 3, 4, with The j-th input to the energy efficiency evaluation model is respectively Therefore, the energy efficiency evaluation model of the grate cooler in the k-th time period is constructed using equation (1):

[0014]

[0015] In equation (1), z -1 For delay operators, it means lag by 1 step; Let A represent the input parameter matrix for the k-th time interval, obtained from equation (2); i (z -1 ) represents the i-th output The output coefficient polynomial is obtained from equation (3); T i (z -1 ) represents the i-th output The input hysteresis matrix is ​​obtained from equation (4); B i (z -1 ) represents the i-th output The input coefficient polynomial matrix is ​​obtained from equation (5); ε i (k) represents the i-th output. The noise term in the k-th time period;

[0016]

[0017]

[0018] In equation (3), The output coefficient polynomial A is respectively i (z -1 The coefficients of each term of ) Indicates the i-th output The output order;

[0019]

[0020] In equation (4), Indicates the i-th output The j-th input The corresponding lag order, τ j For the j-th input The number of lag steps;

[0021] B i (z -1 )=[b i,1 ,b i,2 ,…,b i,j ,…,b i,m+2(5)

[0022] In equation (5), b i,j Indicates the i-th output The j-th input Given the input coefficient polynomial, we have:

[0023]

[0024] In equation (6), The input coefficient polynomial b are respectively i,j The coefficients of each term, Indicates the i-th output The j-th input The input order;

[0025] Step 1.4: Parameter identification of the grate cooler energy efficiency evaluation model;

[0026] Initialize the structural parameters of the grate cooler energy efficiency evaluation model, including: the i-th output. Output order The i-th output The j-th input Input order Input hysteresis matrix T i (z -1 );

[0027] The filtered grate cooler operating data is input into the grate cooler's energy efficiency evaluation model, and the model parameters, including the output coefficient polynomial A, are identified using a recursive least squares algorithm. i (z -1 The coefficients of each term of ) Input coefficient polynomial b i,j coefficients of each term

[0028] Step 2: Construct a multi-objective optimization model for the grate cooler;

[0029] Step 2.1: Use formula (7) to establish the efficiency target and energy consumption target of the grate cooler in the kth time period. The efficiency target includes the secondary air temperature evaluation index, the tertiary air temperature evaluation index and the outlet clinker temperature evaluation index. The energy consumption target is the power consumption evaluation index.

[0030] minF(X(k))=(f1(X(k)),f2(X(k)),f3(X(k)),f4(X(k))) T (7)

[0031] In equation (7), f1(X(k)) represents the secondary wind temperature evaluation index for the k-th time period, and f2(X(k)) represents the three wind temperature evaluation indicators for the k-th time period, and f3(X(k)) represents the clinker outlet temperature evaluation index for the k-th time period, and f4(X(k)) represents the power consumption evaluation index for the k-th time period, and X(k) represents the decision variable for the k-th time period, and Let j′ represent the j′-th decision variable in the k-th time interval, where 1≤j′≤m+1;

[0032] Step 2.2: Establish the constraints for the multi-objective optimization model of the grate cooler;

[0033] Step 3: Solve the multi-objective optimization model of the grate cooler for the k-th time period using the MOEA / D algorithm to obtain the optimal solution of the multi-objective optimization model of the grate cooler for the (k+1)-th time period. Then, identify the m fans FA1 to FA1 corresponding to the optimal solution. m The fan air volume and grate pressure are respectively used as the values ​​of m fans FA1 to FA1 in the (k+1)th time period. m The setpoints for the fan airflow and the grate pressure;

[0034] Step 4: Based on the (k+1)th time period, determine the m wind turbines FA1 to FA2. m The setpoints for fan airflow and grate pressure are determined using an incremental PID algorithm for m fans FA1 to FA2. m The fan airflow controller, based on the generalized predictive control algorithm, and the grate pressure controller respectively control the airflow of m fans FA1 to FA2 in the (k+1)th time period. m The fan air volume and grate pressure are adaptively adjusted to achieve multi-objective optimization control of the grate cooler.

[0035] The multi-objective optimization control method for a grate cooler considering efficiency and energy consumption, as described in this invention, is also characterized in that step 2.2 includes:

[0036] Step 2.2.1: Construct the operating constraints of the grate cooler using equation (8):

[0037]

[0038] In equation (8), u j′min Let u represent the j′-th decision variable. j′ The minimum value of (k), u j′max Let u represent the j′-th decision variable. j′ The maximum value of (k);

[0039] Step 2.2.2: Construct the stability constraints of the grate cooler using equation (9):

[0040]

[0041] In equation (9), SP j′ (k) represents the j′-th decision variable in the k-th time period. The set value, Δu j′ Let j′ represent the j′-th decision variable. and its setting value SP j′ The maximum deviation between (k).

[0042] Step 3 includes:

[0043] Step 3.1: Initialize algorithm parameters:

[0044] Step 3.1.1: Initialize the population size to P, the neighborhood size to T, the current iteration count to G = 0, and the maximum iteration count to G. max The non-dominated solution set EP for the k-th time interval is an empty set;

[0045] Step 3.1.2: Randomly generate an initial population as the Gth generation population, construct P weight vectors and assign them to each individual in the Gth generation population in turn;

[0046] Step 3.1.3: For each individual in the G-th generation population, calculate the Euclidean distance between the weight vector of each individual and the weight vector of other individuals, and select the individuals corresponding to the top T weight vectors with the smallest Euclidean distance as the neighbor set of the corresponding individual.

[0047] Step 3.1.4: Calculate the secondary wind temperature evaluation index, tertiary wind temperature evaluation index, clinker outlet temperature evaluation index and power consumption evaluation index for all individuals in the G generation population using formula (7). Select the minimum value of the secondary wind temperature evaluation index, the minimum value of the tertiary wind temperature evaluation index, the minimum value of the clinker outlet temperature evaluation index and the minimum value of the power consumption evaluation index for all individuals as the ideal point set for the G generation.

[0048] Step 3.2: Update the MOEA / D solution set:

[0049] Step 3.2.1: For each individual in the G-generation population, randomly select two neighbor individuals from the set of neighbors of each individual, and use the selected two neighbor individuals to perform differential evolution on the corresponding individual to generate a new individual in the G-generation population.

[0050] Step 3.2.2: Update the ideal point set based on the secondary wind temperature evaluation index, tertiary wind temperature evaluation index, outlet clinker temperature evaluation index, and power consumption evaluation index of each new individual in the G-generation population to obtain the ideal point set of the G+1 generation.

[0051] Step 3.2.3: Update the neighbor set:

[0052] The neighbor set of each individual in the G-th generation population is updated using the Chebyshev aggregation method to obtain the G+1-th generation population and the neighbor set of each individual in the G+1-th generation population.

[0053] Step 3.2.5: After generating new non-dominated solutions based on the secondary wind temperature evaluation index, tertiary wind temperature evaluation index, outlet clinker temperature evaluation index and power consumption evaluation index of each new individual in the G-generation population, add them to the non-dominated solution set EP of the k-th time period, and remove the dominant solutions dominated by the new non-dominated solutions from EP.

[0054] Step 3.3: Assign G+1 to G, and determine if G < G. max Check if the condition is met. If it is met, proceed to step 3.2; otherwise, it indicates that G has been completed. max In the next iteration, output the non-dominated solution set EP of the multi-objective optimization model for the grate cooler in the k-th time period;

[0055] Step 3.4: Select a non-dominated solution from the non-dominated solution set EP of the k-th time period as the optimal solution of the multi-objective optimization model of the grate cooler in the (k+1)-th time period.

[0056] The present invention provides an electronic device, comprising a memory and a processor, wherein the memory is used to store a program that supports the processor in executing the multi-objective optimization control method for the grate cooler, and the processor is configured to execute the program stored in the memory.

[0057] The present invention discloses a computer-readable storage medium on which a computer program is stored, wherein the computer program, when executed by a processor, performs the steps of the multi-objective optimization control method for the grate cooler.

[0058] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0059] 1. This invention, considering the efficiency and energy consumption of the grate cooler, synchronously optimizes the setpoints for the grate pressure and fan airflow, and combines predictive control methods to achieve adaptive adjustment of the grate cooler parameters, thus overcoming the shortcomings of manual operation in parameter control. In practical applications, users can flexibly select the setpoints for the grate pressure and fan airflow according to different objectives based on the actual control needs of the grate cooler, meeting the control requirements under various operating conditions.

[0060] 2. This invention uses the MOEA / D algorithm to solve the multi-objective optimization model of the grate cooler, which has low computational complexity and fast convergence speed. It can not only optimize the setpoint parameters of the grate cooler offline, but also better meet the online control requirements of the grate cooler. Attached Figure Description

[0061] Figure 1 This is a schematic diagram of the working principle of the grate cooler of the present invention;

[0062] Figure 2 This is a block diagram of the multi-objective optimization control system for the grate cooler of the present invention. Detailed Implementation

[0063] In this embodiment, a multi-objective optimization control method for a grate cooler considering efficiency and energy consumption includes a grate cooler with n cooling fans, of which m fans have freely adjustable airflow rates. The m fans are denoted as FA = {FA1, FA2, ..., FA...}. l ,…,FA m}, FA l This indicates the l-th fan, where 1 ≤ l ≤ m, and the airflow of the remaining fans is constant or controlled by the operator; Figure 1 The diagram shown illustrates the working principle of a grate cooler in a 5000t / d cement plant production line. Figure 1 The grate cooler shown has 11 cooling fans: F1A, F1B, F2, F2L, F2R, F3, F4, F5, F6, F7, and F8. Under normal operating conditions, the airflow of fans F1A, F2, F2L, and F2R can be freely adjusted according to the on-site production situation. Therefore, fans F1A, F2, F2L, and F2R can be designated as FA1, FA2, FA3, and FA4, respectively, while the airflow of the remaining fans remains constant. The multi-objective optimization control method for the grate cooler, considering efficiency and energy consumption, is carried out according to the following steps:

[0064] Step 1: Establish an energy efficiency evaluation model for the grate cooler, taking into account efficiency and energy consumption;

[0065] Step 1.1, Data Acquisition:

[0066] The operating data of the grate cooler is collected in N time periods of length t, and the average operating data of the grate cooler in the k-th time period is calculated and denoted as {y1(k),y2(k),y3(k),u1(k),u2(k),…,u l (k),…,u m (k),u m+1 (k),u m+2 (k)}, k=1,2,…,N, where y1(k) represents the average secondary air temperature of the grate cooler in the k-th time period, y2(k) represents the average tertiary air temperature of the grate cooler in the k-th time period, y3(k) represents the average outlet clinker temperature of the grate cooler in the k-th time period, u l (k) represents the l-th fan FA of the grate cooler in the k-th time period. l Average fan air volume, u m+1(k) represents the average grate pressure of the grate cooler in the k-th time period, u m+2 (k) represents the average raw material feed rate of the grate cooler in the kth time period;

[0067] Calculate the power consumption of the grate cooler in the kth time period, denoted as y4(k);

[0068] Due to the production delay of the grate cooler, the optimization of the grate cooler's setpoint requires a certain time interval. That is, within the time interval t between two adjacent optimizations, the grate cooler will maintain the same setpoint. Therefore, the average operating data of the grate cooler for multiple time periods with t as the time length is collected to reflect the control effect of different setpoints and to provide data support for subsequent setpoint optimization.

[0069] Power consumption refers to the total power consumption of the grate cooler in each time period;

[0070] Step 1.2: Process the grate cooler operating data for the k-th time period using the moving average filtering method to obtain the filtered grate cooler operating data. in, This represents the secondary wind temperature filter value for the k-th time period. This represents the filtered values ​​of the third wind temperature in the k-th time period. This represents the filtered value of the clinker outlet temperature in the k-th time period. This represents the power consumption filter value for the k-th time period. This indicates the l-th fan FA of the grate cooler during the k-th time period. l Fan airflow filter value, This represents the filter value of the grate pressure in the k-th time period. This represents the filtered value of raw material feed rate for the k-th time period;

[0071] Step 1.3: Based on the filtered grate cooler operating data, establish an energy efficiency evaluation model for the grate cooler:

[0072] Step 1.3.1, with and The i-th output of the energy efficiency evaluation model is respectively i=1,2,3,4, with The j-th input to the energy efficiency evaluation model is respectively 1≤j≤m+2, thus the energy efficiency evaluation model of the grate cooler in the k-th time period is constructed using equation (1):

[0073]

[0074] In equation (1), z -1 For delay operators, it means lag by 1 step; Let A represent the input parameter matrix for the k-th time interval, obtained from equation (2); i (z -1 ) represents the i-th output The output coefficient polynomial is obtained from equation (3); T i (z -1 ) represents the i-th output The input hysteresis matrix is ​​obtained from equation (4); B i (z -1 ) represents the i-th output The input coefficient polynomial matrix is ​​obtained from equation (5); ε i (k) represents the i-th output. The noise term in the k-th time period;

[0075]

[0076]

[0077] In equation (3), The output coefficient polynomial A is respectively i (z -1 The coefficients of each term of ) Indicates the i-th output The output order;

[0078]

[0079] In equation (4), Indicates the i-th output The j-th input The corresponding lag order, τ j For the j-th input The number of lag steps;

[0080] B i (z -1 )=[b i,1 ,b i,2 ,…,b i,j ,…,b i,m+2 (5)

[0081] In equation (5), b i,j Indicates the i-th output The j-th input Given the input coefficient polynomial, we have:

[0082]

[0083] In equation (6), The input coefficient polynomial b are respectively i,j The coefficients of each term, Indicates the i-th output The j-th input The input order;

[0084] Step 1.4: Parameter identification of the grate cooler energy efficiency evaluation model;

[0085] Initialize the structural parameters of the grate cooler energy efficiency evaluation model, including: the i-th output. Output order The i-th output The j-th input Input order Input hysteresis matrix T i (z -1 );

[0086] The filtered grate cooler operating data is input into the grate cooler's energy efficiency evaluation model, and the model parameters, including the output coefficient polynomial A, are identified using a recursive least squares algorithm. i (z -1 The coefficients of each term of ) Input coefficient polynomial b i,j coefficients of each term

[0087] Step 2: Construct a multi-objective optimization model for the grate cooler;

[0088] Step 2.1: Use formula (7) to establish the efficiency target and energy consumption target of the grate cooler in the kth time period. The efficiency target includes the secondary air temperature evaluation index, the tertiary air temperature evaluation index and the outlet clinker temperature evaluation index. The energy consumption target is the power consumption evaluation index. The secondary air temperature evaluation index and the tertiary air temperature evaluation index are used to measure the thermal efficiency of the grate cooler, and the outlet clinker temperature is used to measure the cooling efficiency of the grate cooler.

[0089] minF(X(k))=(f1(X(k)),f2(X(k)),f3(X(k)),f4(X(k))) T (7)

[0090] In equation (7), f1(X(k)) represents the secondary wind temperature evaluation index for the k-th time period, and f2(X(k)) represents the three wind temperature evaluation indicators for the k-th time period, and f3(X(k)) represents the clinker outlet temperature evaluation index for the k-th time period, and f4(X(k)) represents the power consumption evaluation index for the k-th time period, and X(k) represents the decision variable for the k-th time period, and Let j′ represent the j′-th decision variable in the k-th time interval, where 1≤j′≤m+1;

[0091] Step 2.2: Establish the constraints for the multi-objective optimization model of the grate cooler:

[0092] Step 2.2.1: Construct the operating constraints of the grate cooler using equation (8):

[0093] To ensure the safe operation of the grate cooler, the parameter values ​​of each decision variable must be within the allowable range of the grate cooler.

[0094]

[0095] In equation (8), u j′min Let j′ represent the j′-th decision variable. The minimum value, u j′max Let j′ represent the j′-th decision variable. The maximum value;

[0096] Step 2.2.2: Construct the stability constraints of the grate cooler using equation (9):

[0097] In this embodiment, in order to keep the set value of the grate cooler relatively stable, the deviation between the decision variables and their set values ​​in each time period should be kept within a certain range during the optimization process.

[0098]

[0099] In equation (9), SP j′ (k) represents the j′-th decision variable in the k-th time period. The set value, Δu j′ Let j′ represent the j′-th decision variable. and its setting value SP j′ The maximum deviation between (k);

[0100] Step 3: Solve the multi-objective optimization model of the grate cooler for the k-th time period based on the MOEA / D algorithm;

[0101] Step 3.1: Initialize algorithm parameters:

[0102] Step 3.1.1: Initialize the population size to P, the neighborhood size to T, the current iteration count to G = 0, and the maximum iteration count to G. max The non-dominated solution set EP for the k-th time interval is an empty set;

[0103] Step 3.1.2: Randomly generate an initial population as the Gth generation population, construct P weight vectors and assign them to each individual in the Gth generation population in turn;

[0104] Step 3.1.3: For each individual in the G-th generation population, calculate the Euclidean distance between the weight vector of each individual and the weight vector of other individuals, and select the individuals corresponding to the top T weight vectors with the smallest Euclidean distance as the neighbor set of the corresponding individual.

[0105] Step 3.1.4: Calculate the secondary wind temperature evaluation index, tertiary wind temperature evaluation index, clinker outlet temperature evaluation index and power consumption evaluation index for all individuals in the G generation population using formula (7). Select the minimum value of the secondary wind temperature evaluation index, the minimum value of the tertiary wind temperature evaluation index, the minimum value of the clinker outlet temperature evaluation index and the minimum value of the power consumption evaluation index for all individuals as the ideal point set for the G generation.

[0106] Step 3.2: Update the MOEA / D solution set:

[0107] Step 3.2.1: For each individual in the G-generation population, randomly select two neighbor individuals from the set of neighbors of each individual, and use the selected two neighbor individuals to perform differential evolution on the corresponding individual to generate a new individual in the G-generation population.

[0108] Step 3.2.2: Update the ideal point set based on the secondary wind temperature evaluation index, tertiary wind temperature evaluation index, outlet clinker temperature evaluation index, and power consumption evaluation index of each new individual in the Gth generation population to obtain the ideal point set of the G+1th generation.

[0109] Specifically, the secondary wind temperature evaluation index, tertiary wind temperature evaluation index, clinker outlet temperature evaluation index, and power consumption evaluation index of each new individual in the Gth generation population are compared with the secondary wind temperature evaluation index, tertiary wind temperature evaluation index, clinker outlet temperature evaluation index, and power consumption evaluation index in the ideal point set. The minimum value of the secondary wind temperature evaluation index, the minimum value of the tertiary wind temperature evaluation index, the minimum value of the clinker outlet temperature evaluation index, and the minimum value of the power consumption evaluation index are taken as the ideal point set of the G+1th generation.

[0110] Step 3.2.3: Update the neighbor set:

[0111] The neighbor set of each individual in the G-th generation population is updated using the Chebyshev aggregation method to obtain the G+1-th generation population and the neighbor set of each individual in the G+1-th generation population.

[0112] For each individual in the G-th generation population, calculate the Chebyshev aggregation function value of all neighboring individuals of each individual, and the Chebyshev aggregation function value of the new individual generated by each individual. If the Chebyshev aggregation function value of a certain neighboring individual of the corresponding individual is greater than the Chebyshev aggregation function value of the new individual generated by the corresponding individual, replace the corresponding neighboring individual with the corresponding new individual. Finally, we obtain the G+1 generation population and the neighbor set of all individuals in the G+1 generation population.

[0113] Step 3.2.5: After generating new non-dominated solutions based on the secondary wind temperature evaluation index, tertiary wind temperature evaluation index, outlet clinker temperature evaluation index and power consumption evaluation index of each new individual in the G-generation population, add them to the non-dominated solution set EP of the k-th time period, and remove the dominant solutions dominated by the new non-dominated solutions from EP.

[0114] Step 3.3: Assign G+1 to G, and determine if G < G. max Check if the condition is met. If it is met, proceed to step 3.2; otherwise, it indicates that G has been completed. max In the next iteration, output the non-dominated solution set EP of the multi-objective optimization model for the grate cooler in the k-th time period;

[0115] Step 3.4: Select a non-dominated solution from the non-dominated solution set EP of the k-th time period as the optimal solution of the multi-objective optimization model of the grate cooler in the (k+1)-th time period, and take the m fans FA1 to FA1 corresponding to the optimal solution. m The fan air volume and grate pressure are respectively used as the values ​​of m fans FA1 to FA1 in the (k+1)th time period. m The setpoints for the fan airflow and the grate pressure;

[0116] The selection of the optimal solution needs to consider actual production needs and the emphasis on different objectives. In this embodiment, in order to ensure the smooth change of the grate cooler's operating state before and after the setpoint optimization, the m fans FA1 to FA1 corresponding to each non-dominated solution are calculated. m The fan air volume and grate pressure are compared with the number of fans FA1 to FA1 in the current time period. m The mean square error between the fan air volume setpoint and the grate pressure setpoint is used to select the non-dominated solution corresponding to the minimum mean square error as the optimal solution.

[0117] Step 4: Based on the (k+1)th time period, determine the m wind turbines FA1 to FA2. m The setpoints for fan airflow and grate pressure are determined using an incremental PID algorithm for m fans FA1 to FA2. m The fan airflow controller, based on the generalized predictive control algorithm, and the grate pressure controller respectively control the airflow of m fans FA1 to FA2 in the (k+1)th time period. mThe fan airflow and grate pressure are adaptively adjusted to achieve multi-objective optimization control of the grate cooler, such as... Figure 2 As shown.

[0118] In this embodiment, an electronic device includes a memory and a processor. The memory stores a program that supports the processor in executing the above-described method, and the processor is configured to execute the program stored in the memory.

[0119] In this embodiment, a computer-readable storage medium stores a computer program, which is executed by a processor to perform the steps of the above method.

Claims

1. A multi-objective optimization control method for a grate cooler considering efficiency and energy consumption, wherein the grate cooler has n cooling fans, wherein, The air volume of m fans can be freely adjusted, and is denoted as FA={FA1,FA2,…,FA... l ,…,FA m }, FA l The l-th fan is defined as follows: 1 ≤ l ≤ m, and the airflow of the remaining fans is constant or controlled by the operator; the multi-objective optimization control method for the grate cooler is characterized by the following steps: Step 1: Establish an energy efficiency evaluation model for the grate cooler, taking into account efficiency and energy consumption; Step 1.1, Data Acquisition: The operating data of the grate cooler is collected in N time periods of length t, and the average operating data of the grate cooler in the k-th time period is calculated and denoted as {y1(k),y2(k),y3(k),u1(k),u2(k),…,u l (k),…,u m (k),u m+1 (k),u m+2 (k)}, k=1,2,…,N, where y1(k) represents the average secondary air temperature of the grate cooler in the k-th time period, y2(k) represents the average tertiary air temperature of the grate cooler in the k-th time period, y3(k) represents the average outlet clinker temperature of the grate cooler in the k-th time period, u l (k) represents the l-th fan FA of the grate cooler in the k-th time period. l Average fan air volume, u m+1 (k) represents the average grate pressure of the grate cooler in the k-th time period, u m+2 (k) represents the average raw material feed rate of the grate cooler in the kth time period; Calculate the power consumption of the grate cooler in the kth time period, denoted as y4(k); Step 1.2: Process the grate cooler operating data for the k-th time period using the moving average filtering method to obtain the filtered grate cooler operating data. in, This represents the secondary wind temperature filter value for the k-th time period. This represents the filtered values ​​of the third wind temperature in the k-th time period. This represents the filtered value of the clinker outlet temperature in the k-th time period. This represents the power consumption filter value for the k-th time period. This indicates the l-th fan FA of the grate cooler during the k-th time period. l Fan airflow filter value, This represents the filter value of the grate pressure in the k-th time period. This represents the filtered value of raw material feed rate for the k-th time period; Step 1.3: Based on the filtered grate cooler operating data, establish an energy efficiency evaluation model for the grate cooler: by and The i-th output of the energy efficiency evaluation model is respectively i=1,2,3,4, with The j-th input to the energy efficiency evaluation model is respectively 1≤j≤m+2, thus the energy efficiency evaluation model of the grate cooler in the k-th time period is constructed using equation (1): In equation (1), z -1 For delay operators, it means lag by 1 step; Let A represent the input parameter matrix for the k-th time interval, obtained from equation (2); i (z -1 ) represents the i-th output The output coefficient polynomial is obtained from equation (3); T i (z -1 ) represents the i-th output The input hysteresis matrix is ​​obtained from equation (4); B i (z -1 ) represents the i-th output The input coefficient polynomial matrix is ​​obtained from equation (5); ε i (k) represents the i-th output. The noise term in the k-th time period; In equation (3), The output coefficient polynomial A is respectively i (z -1 The coefficients of each term of ) Indicates the i-th output The output order; In equation (4), Indicates the i-th output The j-th input The corresponding lag order, τ j For the j-th input The number of lag steps; B i (z -1 )=[b i,1 ,b i,2 ,…,b i,j ,…,b i,m+2 ] (5) In equation (5), b i,j Indicates the i-th output The j-th input Given the input coefficient polynomial, we have: In equation (6), The input coefficient polynomial b are respectively i,j The coefficients of each term, Indicates the i-th output The j-th input The input order; Step 1.4: Parameter identification of the grate cooler energy efficiency evaluation model; Initialize the structural parameters of the grate cooler energy efficiency evaluation model, including: the i-th output. Output order The i-th output The j-th input Input order Input hysteresis matrix T i (z -1 ); The filtered grate cooler operating data is input into the grate cooler's energy efficiency evaluation model, and the model parameters, including the output coefficient polynomial A, are identified using a recursive least squares algorithm. i (z -1 The coefficients of each term of ) Input coefficient polynomial b i,j coefficients of each term Step 2: Construct a multi-objective optimization model for the grate cooler; Step 2.1: Use formula (7) to establish the efficiency target and energy consumption target of the grate cooler in the kth time period. The efficiency target includes the secondary air temperature evaluation index, the tertiary air temperature evaluation index and the outlet clinker temperature evaluation index. The energy consumption target is the power consumption evaluation index. min F(X(k))=(f1(X(k)),f2(X(k)),f3(X(k)),f4(X(k))) T (7) In equation (7), f1(X(k)) represents the secondary wind temperature evaluation index for the k-th time period, and f2(X(k)) represents the three wind temperature evaluation indicators for the k-th time period, and f3(X(k)) represents the clinker outlet temperature evaluation index for the k-th time period, and f4(X(k)) represents the power consumption evaluation index for the k-th time period, and X(k) represents the decision variable for the k-th time period, and Let j′ represent the j′-th decision variable in the k-th time interval, where 1≤j′≤m+1; Step 2.2: Establish the constraints for the multi-objective optimization model of the grate cooler; Step 3: Solve the multi-objective optimization model of the grate cooler for the k-th time period using the MOEA / D algorithm to obtain the optimal solution of the multi-objective optimization model of the grate cooler for the (k+1)-th time period. Then, identify the m fans FA1 to FA1 corresponding to the optimal solution. m The fan air volume and grate pressure are respectively used as the values ​​of m fans FA1 to FA1 in the (k+1)th time period. m The setpoints for the fan airflow and the grate pressure; Step 4: Based on the (k+1)th time period, determine the m wind turbines FA1 to FA2. m The setpoints for fan airflow and grate pressure are determined using an incremental PID algorithm for m fans FA1 to FA2. m The fan airflow controller, based on the generalized predictive control algorithm, and the grate pressure controller respectively control the airflow of m fans FA1 to FA2 in the (k+1)th time period. m The fan air volume and grate pressure are adaptively adjusted to achieve multi-objective optimization control of the grate cooler.

2. The multi-objective optimization control method for a grate cooler considering efficiency and energy consumption according to claim 1, characterized in that, Step 2.2 includes: Step 2.2.1: Construct the operating constraints of the grate cooler using equation (8): In equation (8), u j′min Let j′ represent the j′-th decision variable. The minimum value, u j′max Let u represent the j′-th decision variable. j′ The maximum value of (k); Step 2.2.2: Construct the stability constraints of the grate cooler using equation (9): In equation (9), SP j′ (k) represents the j′-th decision variable in the k-th time period. The set value, Δu j′ Let j′ represent the j′-th decision variable. and its setting value SP j′ The maximum deviation between (k).

3. The multi-objective optimization control method for a grate cooler considering efficiency and energy consumption according to claim 2, characterized in that, Step 3 includes: Step 3.1: Initialize algorithm parameters: Step 3.1.1: Initialize the population size to P, the neighborhood size to T, the current iteration count to G = 0, and the maximum iteration count to G. max The non-dominated solution set EP for the k-th time interval is an empty set; Step 3.1.2: Randomly generate an initial population as the Gth generation population, construct P weight vectors and assign them to each individual in the Gth generation population in turn; Step 3.1.3: For each individual in the G-th generation population, calculate the Euclidean distance between the weight vector of each individual and the weight vector of other individuals, and select the individuals corresponding to the top T weight vectors with the smallest Euclidean distance as the neighbor set of the corresponding individual. Step 3.1.4: Calculate the secondary wind temperature evaluation index, tertiary wind temperature evaluation index, clinker outlet temperature evaluation index and power consumption evaluation index for all individuals in the G generation population using formula (7). Select the minimum value of the secondary wind temperature evaluation index, the minimum value of the tertiary wind temperature evaluation index, the minimum value of the clinker outlet temperature evaluation index and the minimum value of the power consumption evaluation index for all individuals as the ideal point set for the G generation. Step 3.2: Update the MOEA / D solution set: Step 3.2.1: For each individual in the G-generation population, randomly select two neighbor individuals from the set of neighbors of each individual, and use the selected two neighbor individuals to perform differential evolution on the corresponding individual to generate a new individual in the G-generation population. Step 3.2.2: Update the ideal point set based on the secondary wind temperature evaluation index, tertiary wind temperature evaluation index, outlet clinker temperature evaluation index, and power consumption evaluation index of each new individual in the G-generation population to obtain the ideal point set of the G+1 generation. Step 3.2.3: Update the neighbor set: The neighbor set of each individual in the G-th generation population is updated using the Chebyshev aggregation method to obtain the G+1-th generation population and the neighbor set of each individual in the G+1-th generation population. Step 3.2.5: After generating new non-dominated solutions based on the secondary wind temperature evaluation index, tertiary wind temperature evaluation index, outlet clinker temperature evaluation index and power consumption evaluation index of each new individual in the G-generation population, add them to the non-dominated solution set EP of the k-th time period, and remove the dominant solutions dominated by the new non-dominated solutions from EP. Step 3.3: Assign G+1 to G, and determine if G < G. max Check if the condition is met. If it is met, proceed to step 3.2; otherwise, it indicates that G has been completed. max In the next iteration, output the non-dominated solution set EP of the multi-objective optimization model for the grate cooler in the k-th time period; Step 3.4: Select a non-dominated solution from the non-dominated solution set EP of the k-th time period as the optimal solution of the multi-objective optimization model of the grate cooler in the (k+1)-th time period.

4. An electronic device, comprising a memory and a processor, characterized in that, The memory is used to store a program that supports the processor in executing the multi-objective optimization control method for the grate cooler according to claim 1, 2 or 3, and the processor is configured to execute the program stored in the memory.

5. A computer-readable storage medium storing a computer program thereon, characterized in that, The computer program, when run by the processor, executes the steps of the multi-objective optimization control method for the grate cooler as described in claim 1, 2, or 3.

Images