Polynomial modeling method and system for production process indexes

Abstract

The invention relates to the field of industrial production, in particular to a polynomial modeling method and system for production process indexes, and the method comprises the steps: constructing an initial population; iteratively updating the initial population, and selecting a target chromosome; based on the matrix and the least square solution corresponding to the target chromosome, constructing a polynomial calculation model of the target production process index; the method for acquiring the matrix corresponding to the target chromosome and the least square solution comprises the following steps: constructing the matrix based on a gene sequence of the target chromosome; based on the matrix and the constructed matrix, linear correlation columns in the matrix are removed to obtain the matrix; the method is obtained based on equation construction and solution. And according to the removed linear correlation columns, removing corresponding rows in the linear correlation columns to obtain a matrix. The method is used for increasing the interpretability of the model, and the model can be conveniently and directly used for subsequent planning in production practice.

Description

Technical Field

[0001] The present invention relates to the field of industrial production, and in particular to a polynomial modeling method and system for production process indicators. Background Art

[0002] Production process indicators are key performance indicators (KPIs) that measure core elements of industrial production processes. They are crucial for monitoring and optimizing production processes. These indicators have diverse definitions based on different dimensions, encompassing basic raw material properties such as thickness and width, as well as processing variables such as finished dimensions (thickness, width), processing time, and energy consumption (electricity, steam). Furthermore, production process indicators can refer to the operating parameters of specific equipment, such as the power consumption, flow rate, and pressure of a pump or group of pumps.

[0003] In current production practice, the calculation of many production process indicators tends to rely on the construction of neural network models. For example, the power consumption of a parallel pump group, particularly when multiple variable-frequency pumps are involved, is a key indicator of the industry's preference for neural network modeling to improve the accuracy of calculating the total power consumption of the pump group. This technique begins by designing and building a neural network architecture, then collecting actual production data to construct a training dataset. This dataset is then used to train the neural network, ultimately resulting in a model that effectively predicts the power consumption of the pump group. However, while neural network models offer advantages in predictive accuracy, they lack interpretability and are difficult to intuitively explain. Furthermore, neural network modeling is difficult to directly apply to subsequent planning in production practice. For example, optimizing the parallel operation of a parallel pump group typically involves modeling the total power consumption of the pump group. Using neural network modeling results in a black-box model, making it difficult to directly utilize for optimization. Summary of the Invention

[0004] To solve the above problems, the present invention provides a polynomial modeling method and system for production process indicators, which are used to solve at least one aspect of the above problems.

[0005] In a first aspect, the present invention provides a polynomial modeling method for production process indicators, the method comprising: S1. Construct an initial population. In this population, the total number of chromosomes is Each chromosome contains genes, each gene was used Initialize with random integers on 、 The highest power and total number of pre-set modeling independent variables, Model the number of terms for a pre-set polynomial; S2, using genetic algorithm to iteratively update the initial population, and selecting the optimal chromosome in the population after the last iterative update as the target chromosome; S3, based on the matrix corresponding to the target chromosome and the least squares solution , construct a polynomial calculation model for the target production process indicators: , in, is the target production process indicator, represents the independent variable vector composed of each modeling independent variable, is a vector The jth independent variable in , express of Power, is a matrix Middle OK Elements of the column, for Middle elements, ; The matrix corresponding to the target chromosome and the least squares solution The method for obtaining comprises steps S4 to S5: S4. Constructing a matrix based on the gene sequence of the target chromosome , ; Matrix-based and matrix Constructing the Matrix , where the matrix For the pre-collected The independent variable matrix constructed by the historical data of , for The total sample size, , , express of power; Remove Matrix The linearly dependent columns in , we get the matrix ; is a matrix The total number of columns; Matrix-based Constructing the equation ,in , A dependent variable vector constructed using historical production process indicator values ​​corresponding to the historical data collected in advance; Solving the equation for the least squares solution yields The least squares solution of ; S5. Based on the matrix Remove the column number of the linearly correlated column removed from The rows with corresponding serial numbers in the matrix are obtained. .

[0006] Furthermore, , is the gene sequence, express Middle A gene.

[0007] Furthermore, the iterative update of the initial population using a genetic algorithm comprises the following steps: T1. Perform genetic iteration on the chromosomes of the current latest generation population. In each iteration, each chromosome in the current latest generation population is used as a parent chromosome, and the next generation of child chromosomes of each parent chromosome is obtained. For each parent chromosome, the size relationship between its evaluation index and that of its next generation child chromosome is compared, and the chromosome with the smallest evaluation index is selected as the child chromosome of the corresponding parent chromosome; the child chromosomes are collected to construct the next generation population, and then step T1 is continued until the iteration ends; In step T1, the calculation method of the chromosome evaluation index includes: Use the same calculation method as step S4 to calculate the matrix corresponding to the chromosome and the least squares solution , and get the target matrix and the first least squares solution ; calculate and The Euclidean distance between them is used to obtain the evaluation index of the chromosome.

[0008] Furthermore, the method for selecting the optimal chromosome in the population after the last iterative update includes: Aggregating the evaluation indicators of each chromosome in the target population to obtain a set of target evaluation indicators; the target population is the population after the last iterative update; Select an evaluation indicator with the smallest value from each target evaluation indicator as the first evaluation indicator; The relevant chromosomes in the target population corresponding to the first evaluation index are selected to obtain the optimal chromosome in the target population.

[0009] Furthermore, the removal matrix The linearly dependent columns in , we get the matrix , including the following steps: Step 100, define a matrix U, a column vector d and a matrix for storing Initialize the matrix U to be an empty matrix, initialize d to be an empty vector, and initialize r to be an empty vector; Step 101, traverse the matrix Each column , for the currently traversed Each column , respectively perform the following steps 1011 to 1024, wherein the column represent The hth column of Step 1011, cloning The hth column of ; Step 1012, judgment matrix Is there a non-zero element in : If so, clone the vector Get vector , the matrix With the matrix The h-th column vector of Multiply to get vector , and the vector Divide each element in by the vector The elements at the corresponding positions in the vector are obtained , then use renew The value of , then go to step 1013 to continue execution; represent The transposed matrix of If not, go to step 1013 to continue execution; Step 1013, calculate Dot product with itself ; Step 1014, determine Is it true: If yes, then the vector After transposition, add as new row to At the end of Add as a new element to the column vector At the end of Add it to the end of vector r as a new element; otherwise, do nothing; is the pre-set allowable error threshold; Step 102, after step 101 is executed, according to vector The serial number stored in The corresponding column number in the generator matrix , where the matrix Representative Matrix No. List, is a vector Middle elements.

[0010] Furthermore, the target production process indicator is the total flow rate of the parallel water pump group, the total pressure of the parallel water pump group or the power consumption of a single variable frequency pump in the parallel water pump group; when the target production process indicator is the total flow rate or total pressure of the parallel water pump group, the pre-set modeling independent variable is the operating frequency of each water pump in the parallel water pump group; when the target production process indicator is the power consumption of a single variable frequency pump in the parallel water pump group, the pre-set modeling independent variable is the operating frequency of the corresponding variable frequency pump in the parallel water pump group.

[0011] Furthermore, the target production process indicator is the electricity consumption per ton of steel, steam consumption per ton of steel or processing time per ton of steel of the steel production line, and the pre-set modeling independent variables are the raw material thickness, raw material width, steel processing thickness and steel processing width of the steel produced by the steel production line.

[0012] In a second aspect, the present invention provides a polynomial modeling system for production process indicators, the system comprising: The population construction module is used to construct the initial population, in which the total number of chromosomes is Each chromosome contains genes, each gene was used Initialize with random integers on 、 The highest power and total number of pre-set modeling independent variables, Model the number of terms for a pre-set polynomial; A chromosome selection module is connected to the population construction module and is used to iteratively update the initial population using a genetic algorithm and select the optimal chromosome in the population after the last iterative update as the target chromosome; Model building module, connected to the chromosome selection module, for the matrix corresponding to the target chromosome and the least squares solution , construct a polynomial calculation model for the target production process indicators: , in, is the target production process indicator, represents the independent variable vector composed of each modeling independent variable, is a vector The jth independent variable in , express of Power, is a matrix Middle OK Elements of the column, for Middle elements, ; Among them, the matrix corresponding to the target chromosome and the least squares solution Methods for obtaining include: Step 1: Construct a matrix based on the gene sequence of the target chromosome , ; Matrix-based and matrix Constructing the Matrix , where the matrix For the pre-collected The independent variable matrix constructed by the historical data of , for The total sample size, , , express of power; Remove Matrix The linearly dependent columns in , we get the matrix ; is a matrix The total number of columns; Matrix-based Constructing the equation ,in , A dependent variable vector constructed using historical production process indicator values ​​corresponding to the historical data collected in advance; Solving the equation for the least squares solution yields The least squares solution of ; Step 2: Based on the matrix Remove the column number of the linearly correlated column removed from The rows with corresponding serial numbers in the matrix are obtained. .

[0013] In a third aspect, the present invention provides a water pump control method, which is applied to a parallel water pump group. The parallel water pump group has a total of water pumps, water pumps are connected in parallel, of which Power frequency pumps and Variable frequency pumps, , 、 、 , 、 、 All are integers; methods include: Step 100: Calculate the power consumption of each power frequency pump in the parallel water pump group using the historical average method; if there is no power frequency pump in the parallel water pump group, skip this step; Step 110: If there is a variable frequency pump in the parallel water pump group, use the method described in any one of claims 1 to 7 to construct a polynomial calculation model for the total flow of the parallel water pump group. , total pressure polynomial calculation model , and the power consumption polynomial calculation model of each variable frequency pump in the parallel pump group; among them, is the independent variable vector composed of the operating frequency of each water pump in the parallel water pump group as the modeling independent variable. If there is no variable frequency pump in the parallel water pump group, skip this step; Step 120 is the Generate all non-idle-start pump plans containing at least one open power frequency pump for each water pump; Step 130: For each of all generated pump start-up plans, execute: The power consumption of each turned-on power frequency pump under the current pump-on scheme is obtained from the power consumptions calculated in step 100, and the calculation expression of the power consumption polynomial calculation model of each turned-on variable frequency pump under the current pump-on scheme is obtained from the power consumption polynomial calculation models constructed in step 110. The power consumption of each turned-on power frequency pump under the current pump-on scheme obtained is added to the calculation expression of the power consumption polynomial calculation model of each turned-on variable frequency pump under the current pump-on scheme, and the total power consumption calculation formula of the parallel water pump group with respect to the working frequency of each turned-on variable frequency pump under the current pump-on scheme is obtained. ; and in the total flow polynomial calculation model And the total pressure polynomial calculation model In the current pump-on scheme, the modeling independent variables corresponding to each closed pump are assigned to 0, and the modeling independent variables corresponding to each open power frequency pump are assigned to its rated operating frequency or the average of its historical operating frequencies. The total flow target calculation formula of the parallel water pump group under the current pump-on scheme is obtained. And the total pressure target calculation formula ,in is the independent variable vector composed of the operating frequencies of all the variable frequency pumps under the current pump start scheme; if there is no closed pump and open power frequency pump under the current pump start scheme, the total flow target calculation formula of the parallel water pump group under the current pump start scheme is for , the total pressure target calculation formula of the parallel water pump group under the current pump start plan is for , for ; Then construct the nonlinear programming problem: ;

[0014] in, is the preset water outlet flow error threshold of the parallel water pump group. is the preset target water flow rate of the parallel pump group. The minimum pressure that the parallel pump group needs to meet is preset. Indicates the first of all variable frequency pumps that are turned on under the current pump start plan. The modeling independent variables corresponding to the variable frequency pumps are: The pre-set The minimum operating frequency of a variable frequency pump, The pre-set The maximum operating frequency of a variable frequency pump, Modeling independent variables In the independent variable vector The serial number in The total number of all variable frequency pumps turned on under the current pump start plan; Then solve the nonlinear programming problem and get A solution to ; Then, substitute the solution back into 、 and , calculate the total power consumption, total flow and total pressure of the parallel pump group under the current pump start scheme; and, based on the above solution and arrangement, get and When is the value of the modeling independent variable corresponding to each closed pump and each open power frequency pump under the current pump start scheme, the operating frequency of each water pump in the parallel water pump group under the current pump start scheme is obtained by summarizing, that is, the target operating frequency of each water pump in the parallel water pump group under the current pump start scheme is obtained; Step 140: read two pump start plans from all generated pump start plans, and then execute step L1; Step L1, obtaining the calculated total power consumption, total flow rate, and total pressure of the parallel water pump group under the two current pump start-up plans, and based on the obtained total power consumption, total flow rate, and total pressure, selecting one of the two current pump start-up plans as a temporary pump start-up plan, and then executing step L2; Methods for selecting one of the two current pump start-up plans as a temporary pump start-up plan include: If both current pump start-up plans satisfy formula (1) and formula (2), the one with the smallest total power consumption of the parallel water pump group under the two current pump start-up plans is used as the temporary pump start-up plan, and then step L2 is executed; If only one of the two current pump start-up plans satisfies both formula (1) and formula (2), the pump start-up plan that satisfies both formula (1) and formula (2) is used as a temporary pump start-up plan, and then step L2 is executed; If the two current pump start-up plans do not satisfy formula (1) and formula (2) at the same time, then the two current pump start-up plans The one with the smallest value is used as the temporary pump start solution, and then step L2 is executed. represents absolute value calculation; Formula (1) is: ; Formula (2) is: ; in, Represents the total flow of the parallel pump group, Represents the total pressure of the parallel pump group; Step L2: Continue to read the next pump start plan from all the generated non-empty pump start plans, and combine it with the most recent temporary pump start plan obtained from the current moment to form two pump start plans. Then continue to execute step L1 for these two pump start plans until all the generated non-empty pump start plans are read, and use the temporary pump start plan obtained last in step L1 as the target pump start plan. Step 150: Obtain the target operating frequency of each water pump in the parallel water pump group under the target pump start-up scheme, and obtain the operating frequency of each water pump in the parallel water pump group under the target pump start-up scheme; Step 160 : Control the operation of each water pump in the parallel water pump group according to the operating frequency of each water pump in the parallel water pump group under the target pump start-up plan.

[0015] In a fourth aspect, the present invention provides a water pump control system, which is applied to a parallel water pump group. The parallel water pump group has a total of water pumps, water pumps are connected in parallel, of which Power frequency pumps and Variable frequency pumps, , 、 、 , 、 、 All are integers; the system includes: The first module is used to calculate the power consumption of each power frequency pump in the parallel water pump group using the historical average method; The second module is used to construct a polynomial calculation model for the total flow of the parallel water pump group using the polynomial modeling method described in the above aspects. , total pressure polynomial calculation model of parallel pump group , and the power consumption polynomial calculation model of each variable frequency pump in the parallel pump group; among them, is an independent variable vector used to construct the total flow polynomial calculation model and the total pressure polynomial calculation model; The third module is used for Generate all non-idle-start pump plans containing at least one open power frequency pump for each water pump; The fourth module is used to execute the following steps for each of the generated pump start-up plans: The power consumption of each turned-on power frequency pump under the current pump-on scheme is obtained from the calculated power consumptions, and the calculation expression of the power consumption polynomial calculation model of each turned-on variable frequency pump under the current pump-on scheme is obtained from the constructed power consumption polynomial calculation models. The power consumption of each turned-on power frequency pump under the current pump-on scheme obtained is added to the calculation expression of the power consumption polynomial calculation model of each turned-on variable frequency pump under the current pump-on scheme, and the total power consumption calculation formula of the parallel water pump group for the working frequency of each turned-on variable frequency pump under the current pump-on scheme is obtained. ; and in the total flow polynomial calculation model And the total pressure polynomial calculation model In the current pump-on scheme, the modeling independent variables corresponding to each closed pump are assigned to 0, and the modeling independent variables corresponding to each open power frequency pump are assigned to its rated operating frequency or the average of its historical operating frequencies. The total flow target calculation formula of the parallel water pump group under the current pump-on scheme is obtained. And the total pressure target calculation formula ,in is the independent variable vector composed of the operating frequencies of all the variable frequency pumps under the current pump start scheme; if there is no closed pump and open power frequency pump under the current pump start scheme, the total flow target calculation formula of the parallel water pump group under the current pump start scheme is for , the total pressure target calculation formula of the parallel water pump group under the current pump start plan is for , for ; Then construct the nonlinear programming problem: ;

[0016] in, is the preset water outlet flow error threshold of the parallel water pump group. is the preset target water flow rate of the parallel pump group. The minimum pressure that the parallel pump group needs to meet is preset. Indicates the first of all variable frequency pumps that are turned on under the current pump start plan. The modeling independent variables corresponding to the variable frequency pumps are: The pre-set The minimum operating frequency of a variable frequency pump, The pre-set The maximum operating frequency of a variable frequency pump, Modeling independent variables In the independent variable vector The serial number in The total number of all variable frequency pumps turned on under the current pump start plan; Then solve the nonlinear programming problem and get A solution to ; Then, substitute the solution back into 、 and , calculate the total power consumption, total flow and total pressure of the parallel pump group under the current pump start scheme; and, based on the above solution and arrangement, get and When is the value of the modeling independent variable corresponding to each closed pump and each open power frequency pump under the current pump start scheme, the operating frequency of each water pump in the parallel water pump group under the current pump start scheme is obtained by summarizing, that is, the target operating frequency of each water pump in the parallel water pump group under the current pump start scheme is obtained; A fifth module is configured to read two pump start-up plans from all generated pump start-up plans, and then execute step L1; Step L1, obtaining the calculated total power consumption, total flow rate, and total pressure of the parallel water pump group under the two current pump start-up plans, and based on the obtained total power consumption, total flow rate, and total pressure, selecting one of the two current pump start-up plans as a temporary pump start-up plan, and then executing step L2; Methods for selecting one of the two current pump start-up plans as a temporary pump start-up plan include: If both current pump start-up plans satisfy formula (1) and formula (2), the one with the smallest total power consumption of the parallel water pump group under the two current pump start-up plans is used as the temporary pump start-up plan, and then step L2 is executed; If only one of the two current pump start-up plans satisfies both formula (1) and formula (2), the pump start-up plan that satisfies both formula (1) and formula (2) is used as a temporary pump start-up plan, and then step L2 is executed; If the two current pump start-up plans do not satisfy formula (1) and formula (2) at the same time, then the two current pump start-up plans The one with the smallest value is used as the temporary pump start solution, and then step L2 is executed. represents absolute value calculation; Formula (1) is: ; Formula (2) is: ; in, Represents the total flow of the parallel pump group, Represents the total pressure of the parallel pump group; Step L2: Continue to read the next pump start plan from all the generated non-empty pump start plans, and combine it with the most recent temporary pump start plan obtained from the current moment to form two pump start plans. Then continue to execute step L1 for these two pump start plans until all the generated non-empty pump start plans are read, and use the temporary pump start plan obtained last in step L1 as the target pump start plan. The sixth module is used to obtain the target operating frequency of each water pump in the parallel water pump group under the target pump start-up plan, and obtain the operating frequency of each water pump in the parallel water pump group under the target pump start-up plan; The seventh module is used to control the operation of each pump in the parallel pump group according to the operating frequency of each pump in the parallel pump group under the target pump start plan.

[0017] It can be seen from the above technical solutions that the present invention has the following advantages: The present invention constructs an initial population based on the highest power and total number of pre-set modeling independent variables and the pre-set number of polynomial modeling terms, and uses a genetic algorithm to iteratively update the initial population, and selects the optimal chromosome in the population after the last iterative update as the target chromosome. Then, based on the matrix corresponding to the target chromosome, the optimal chromosome is selected. and the least squares solution , constructing a polynomial calculation model for production process indicators, avoiding the use of neural network modeling, increasing the interpretability of the model, facilitating intuitive explanation, and to a certain extent facilitating direct use in subsequent planning in production practice.

[0018] The present invention helps to realize multivariate polynomial approximation of continuous functions, helps to construct a calculation model for production process indicators whose calculation expressions are continuous functions, and is easy to promote and use.

[0019] The water pump control method and system provided by the present invention adopt the polynomial modeling method provided by the present invention to construct the total power consumption calculation model, total flow calculation model and total pressure calculation model of the parallel water pump group, avoiding the calculation model of the total power consumption, total flow and total pressure of the parallel water pump group obtained by using neural network modeling. The calculation model itself is interpretable, which is convenient for direct use to optimize the parallel connection of water pumps.

[0020] The water pump control method and system provided by the present invention help to convert the parallel optimization problem of multiple variable frequency pumps into nonlinear programming, and help to minimize the total energy consumption while satisfying the flow, pressure and frequency constraints. It can be seen that the present invention can increase the interpretability of the model and facilitate the direct use of the model for subsequent optimization, thereby helping to improve the energy efficiency of water pump centralized control and reduce energy consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solution of the present invention, the following is a brief introduction to the drawings required for the description. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1 It is a schematic flow chart of a polynomial modeling method according to an embodiment of the present invention.

[0023] Figure 2 It is a schematic block diagram of an embodiment of the polynomial modeling system of the present invention.

[0024] Figure 3 It is a schematic block diagram of an embodiment of the water pump control system of the present invention. DETAILED DESCRIPTION

[0025] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0026] SLSQP algorithm: Sequential Least Squares Quadratic Programming (SLSQP) is a numerical optimization method used to solve constrained nonlinear optimization problems. It decomposes the original problem into a series of quadratic programming (QP) subproblems and solves them iteratively, ultimately approaching the optimal solution.

[0027] Figure 1 A schematic flow chart of a polynomial modeling method for production process indicators provided by an embodiment of the present invention. The polynomial modeling method for production process indicators provided by an embodiment of the present invention is executed by a computer device.

[0028] Please refer to Figure 1 The polynomial modeling method includes the following steps S1 to S3.

[0029] Step S1: Construct the initial population.

[0030] In this initial population, the total number of chromosomes is Each chromosome contains genes, each gene was used Initialize with random integers on is the highest power of the pre-set modeling independent variable, is the total number of pre-set modeling independent variables, Models the number of terms for a pre-set polynomial.

[0031] It is understandable that each gene of each chromosome in the initial population is randomly selected Integer initialization on .

[0032] The value of can be set by those skilled in the art according to actual conditions, for example, 200 can be taken.

[0033] It is an empirical value, usually any positive integer between 1 and 40. It can be set by technicians in this field according to the number of modeling independent variables to be set. The specific setting rule is usually to set it larger when there are more modeling independent variables and smaller when there are fewer modeling independent variables.

[0034] The specific value of can be set by those skilled in the art according to actual conditions.

[0035] As a preference, The value range is 1 5. In this embodiment Take 3.

[0036] Step S2: Use a genetic algorithm to iteratively update the initial population, and select the optimal chromosome in the population after the last iterative update as the target chromosome.

[0037] Step S3: Based on the matrix corresponding to the target chromosome and the least squares solution , construct a polynomial calculation model for the target production process indicators.

[0038] The polynomial calculation model for constructing the target production process index is: , in, is the target production process indicator (i.e. the production process indicator of the polynomial calculation model to be constructed), represents the independent variable vector composed of each modeling independent variable, is a vector Middle independent variables, , express of Power, is a matrix Middle OK Elements of the column, for Middle elements, .

[0039] In this embodiment, the matrix corresponding to the target chromosome is and the least squares solution The method for obtaining includes the following steps 1 and 2.

[0040] Step 1: Construct a matrix based on the gene sequence of the target chromosome ; Based on the matrix and matrix Constructing the Matrix ; Remove the matrix The linearly dependent columns in , we get the matrix ; Based on the matrix Constructing the equation ; Solve the least squares solution of the equation and get The least squares solution of the target chromosome is obtained .

[0041] The above matrix , is a matrix No. OK Elements of a column.

[0042] The above matrix For the pre-collected The independent variable matrix constructed by the historical data of , is a matrix No. OK Column elements, for The total sample size, , , express of Power, =1,2,…,u, .

[0043] above is a matrix The total number of columns.

[0044] above , the above It is a dependent variable vector constructed using the values ​​of historical production process indicators corresponding to the historical data collected in advance.

[0045] It is understandable that in the specific implementation, the independent variable vector can be collected first The historical data and the corresponding historical production process index values ​​are then used to collect the independent variable vector The historical data constructs the matrix , using the values ​​of the collected historical production process indicators to construct the dependent variable vector .

[0046] In some embodiments, , is the gene sequence of the target chromosome, Indicates gene sequence Middle A gene.

[0047] Step 2: Based on the matrix Remove the column number of the linearly correlated column removed from The row with the corresponding serial number in the matrix corresponding to the target chromosome is obtained. .

[0048] In this embodiment, the n modeling independent variables are: . The independent variable vector . Representative Matrix No. The column elements correspond to the modeling independent variables.

[0049] In some embodiments of the present invention, the iterative updating of the initial population using a genetic algorithm in step S2 includes the following steps: T1. Perform genetic iteration on the chromosomes of the current latest generation population. In each iteration: Take each chromosome in the current latest generation population as a parent chromosome, and obtain the next-generation child chromosomes of each parent chromosome. For each parent chromosome, compare the magnitude relationship between its evaluation index and that of its next-generation child chromosome, and select the chromosome with the smallest evaluation index as the offspring chromosome corresponding to the parent chromosome. Assemble all the offspring chromosomes to construct the next-generation population, and then turn back to continue executing step T1 until the iteration ends.

[0050] Optionally, the calculation methods of the evaluation indexes of the chromosomes involved in step T1 all include: Using the same calculation method as in step one, calculate the matrix corresponding to the chromosome and the least squares solution to obtain the target matrix and the first least squares solution ; Calculate and [[ID=佃]]The Euclidean distance between them to obtain the evaluation index of the chromosome.

[0051] Specifically, [[ID=却]]The Euclidean distance calculation formula between and is: , In the formula, e represents the evaluation index of the calculated chromosome, represents calculating the Euclidean distance.

[0052] Taking chromosome b as an example, performing step one on chromosome b, that is, obtaining the matrix corresponding to chromosome b and the least squares solution , calculating the product between the matrix corresponding to chromosome b and the least squares solution <00005gl>, and then calculating the Euclidean distance between the vector y and the product, that is, obtaining the evaluation index of this chromosome b. [[ID=4也]]

[0053] It can be understood that after each iteration of the present invention, the calculation of the evaluation indexes of all relevant chromosomes in the corresponding iteration process is completed respectively. "

[0054] In some embodiments of the present invention, the genetic iteration of the chromosomes of the current latest generation population in step T1 includes: Traverse the chromosomes of the current latest generation population, and respectively perform steps T11 to T13 on each traversed chromosome: T11. Copy the currently traversed chromosome and denote it as a parent chromosome; randomly generate a random number rand1 in the interval [0, 1), and determine whether rand1 < rl holds:​ If so, randomly select a chromosome from other chromosomes except the paternal chromosome in the current latest generation population as the maternal chromosome, and perform gene crossover on the gene sequences of the paternal chromosome and the maternal chromosome to obtain the sub-chromosome corresponding to the paternal chromosome, denoted as the transitional chromosome. Then, take r2 = 0.7 and execute step T12; If not, denote the paternal chromosome as the transitional chromosome, take r2 = 1, and then execute step T12; T12: Randomly generate a random number rand2 in the interval [0, 1). If rand2 < r2, execute step T13. If rand2 ≥ r2, the transitional chromosome is the offspring chromosome of the paternal chromosome; T13: Randomly generate an integer z1 as the number of mutations. Then, randomly generate z1 mutation position indexes in the transitional chromosome. Then, randomly take z1 integers z1 times from the interval [0, and, according to the z1 mutation position indexes, randomly replace the genes at the corresponding positions on the transitional chromosome with the z1 integers to obtain the offspring chromosome of the paternal chromosome; where r1 is a preset probability of crossover and r2 is a preset probability of mutation; .

[0055] In some other embodiments of the present invention, the implementation method of selecting the optimal chromosome in the population after the last iterative update in step S2 includes: Collect the evaluation indexes of each chromosome in the target population to obtain a set of target evaluation indexes; the population is the population after the last iterative update; Select the evaluation index with the smallest value from the target evaluation indexes as the first evaluation index; Select the relevant chromosome in the target population corresponding to the first evaluation index to obtain the optimal chromosome in the target population.

[0056] In some other embodiments of the present invention, the above-mentioned removal of the linear dependent columns in the matrix to obtain the matrix includes the following steps: The removal of the linear dependent columns in the matrix to obtain the matrix includes the following steps: Step 100: Define a matrix U, a column vector d, and a vector r for storing the column numbers of the linearly independent columns in the matrix , and initialize the matrix U as an empty matrix, initialize d as an empty vector, and initialize r as an empty vector; Step 101: Traverse each column of the matrix [[ID=4,2]] of , for the currently traversed Each column , respectively perform the following steps 1011 to 1024, wherein the column represent The hth column of Step 1011, cloning The hth column of ; Step 1012, judgment matrix Is there a non-zero element in : If so, clone the vector Get vector , the matrix With the matrix The h-th column vector of Multiply to get vector , and the vector Divide each element in by the vector The elements at the corresponding positions in the vector are obtained , then use renew The value of , then go to step 1013 to continue execution; represent The transposed matrix of If not, go to step 1013 to continue execution; Step 1013, calculate Dot product with itself ; Step 1014, determine Is it true: If yes, then the vector After transposition, add as new row to At the end of Add as a new element to the column vector At the end of Add it to the end of vector r as a new element; otherwise, do nothing; is the pre-set allowable error threshold; Step 102, after step 101 is executed, according to vector The serial number stored in The corresponding column number in the generator matrix , where the matrix Representative Matrix No. List, is a vector Middle elements.

[0057] Understandably, the column vector represent The h-th column vector of .

[0058] Understandably, .

[0059] Understandably, ,and is a number close to 0. Alternatively, In this embodiment, .

[0060] In some other embodiments of the present invention, in step 2, according to the matrix Remove the column number of the linearly correlated column removed from The rows with corresponding serial numbers in the matrix are obtained. ,include: Step S51: Based on the matrix stored in the vector r The column numbers of the linearly independent columns in the matrix are obtained. The serial numbers of other columns in the matrix are obtained. The column numbers of the linearly correlated columns are all recorded as target numbers; Step S52: Delete The rows corresponding to the target sequence number in the matrix are obtained. .

[0061] Optionally, step S51 includes: Step S511, traverse the matrix For each column number currently traversed, it is determined whether it exists in the vector r. If it exists, the number is skipped and the matrix is ​​traversed. The column number of the next column in the middle. If it does not exist, the column number of the current traversal is recorded; Step S512: Gather all the column sequence numbers recorded in step S51 to obtain a set of target sequence numbers.

[0062] In some other embodiments of the present invention, the least squares solution of the equation is solved to obtain the corresponding ,include: The matrix obtained after executing the above step 101 is Each row vector of is divided by the corresponding dot product element in vector d (i.e. ), and we get the matrix ,in Representative Matrix No. row vector, Represents a column vector No. row elements, ; in accordance with ,get: ; (1) Multiply both sides of the equation ,get: ; (2) Combining formulas (1) and (2), we get ; (3) Solve the least squares solution of formula (3) and get The least squares solution of ,in, .

[0063] Understandably, the above For the The inverse matrix of .

[0064] In some embodiments of the present invention, the target production process indicator is the total flow rate of the parallel water pump group, the total pressure of the parallel water pump group or the power consumption of a single variable frequency pump in the parallel water pump group; when the target production process indicator is the total flow rate or total pressure of the parallel water pump group, the pre-set modeling independent variable is the operating frequency of each water pump in the parallel water pump group; when the target production process indicator is the power consumption of a single variable frequency pump in the parallel water pump group, the pre-set modeling independent variable is the operating frequency of the corresponding variable frequency pump in the parallel water pump group.

[0065] Taking a parallel water pump group as an example, assuming there are five water pumps, and the five water pumps are connected in parallel, there are five pre-set modeling independent variables, and the five modeling independent variables correspond to the operating frequencies of the five water pumps.

[0066] In some embodiments of the present invention, the target production process indicator is the electricity consumption per ton of steel, steam consumption per ton of steel, or processing time per ton of steel of the steel production line, and the pre-set modeling independent variables are the raw material thickness, raw material width, steel processing thickness, and steel processing width of the steel produced by the steel production line.

[0067] Second, as Figure 2 As shown, the polynomial modeling system provided by the present invention specifically includes: The population construction module 201 is used to construct an initial population, in which the total number of chromosomes is Each chromosome contains genes, each gene was used Initialize with random integers on 、 The highest power and total number of pre-set modeling independent variables, Model the number of terms for a pre-set polynomial; The chromosome selection module 202 is connected to the population construction module 201 and is used to iteratively update the initial population using a genetic algorithm and select the optimal chromosome in the population after the last iterative update as the target chromosome; The model building module 203 is connected to the chromosome selection module 202 and is used to select the target chromosome based on the matrix corresponding to the target chromosome. and the least squares solution , construct a polynomial calculation model for the target production process indicators: , in, is the target production process indicator, represents the independent variable vector composed of each modeling independent variable, is a vector Middle independent variables, , express of Power, is a matrix Middle OK Elements of the column, for Middle elements, .

[0068] Among them, the matrix corresponding to the target chromosome and the least squares solution Methods for obtaining include: Step 1: Construct a matrix based on the gene sequence of the target chromosome , ; Matrix-based and matrix Constructing the Matrix , where the matrix For the pre-collected The independent variable matrix constructed by the historical data of , for The total sample size, , , express of power; Remove Matrix The linearly dependent columns in , we get the matrix ; is a matrix The total number of columns; Matrix-based Constructing the equation ,in , A dependent variable vector constructed using historical production process indicator values ​​corresponding to the historical data collected in advance; Solving the equation for the least squares solution yields The least squares solution of ; Step 2: Based on the matrix Remove the column number of the linearly correlated column removed from The rows with corresponding serial numbers in the matrix are obtained. .

[0069] In addition, the present invention provides a water pump control method, which is applied to a parallel water pump group. The parallel water pump group has a total of water pumps, water pumps are connected in parallel, of which Power frequency pumps and Variable frequency pumps, , 、 、 , 、 、 All are integers; methods include: Step 100: Calculate the power consumption of each power frequency pump in the parallel water pump group using the historical average method; if there is no power frequency pump in the parallel water pump group, skip this step; Step 110: If there is a variable frequency pump in the parallel water pump group, use the method described in any one of claims 1 to 7 to construct a polynomial calculation model for the total flow of the parallel water pump group. , total pressure polynomial calculation model , and the power consumption polynomial calculation model of each variable frequency pump in the parallel pump group; among them, is the independent variable vector composed of the operating frequency of each water pump in the parallel water pump group as the modeling independent variable. If there is no variable frequency pump in the parallel water pump group, skip this step; Step 120 is the Generate all non-idle-start pump plans containing at least one open power frequency pump for each water pump; Step 130: For each of all generated pump start-up plans, execute: The power consumption of each turned-on power frequency pump under the current pump-on scheme is obtained from the power consumptions calculated in step 100, and the calculation expression of the power consumption polynomial calculation model of each turned-on variable frequency pump under the current pump-on scheme is obtained from the power consumption polynomial calculation models constructed in step 110. The power consumption of each turned-on power frequency pump under the current pump-on scheme obtained is added to the calculation expression of the power consumption polynomial calculation model of each turned-on variable frequency pump under the current pump-on scheme, and the total power consumption calculation formula of the parallel water pump group with respect to the working frequency of each turned-on variable frequency pump under the current pump-on scheme is obtained. ; and in the total flow polynomial calculation model And the total pressure polynomial calculation model In the current pump-on scheme, the modeling independent variables corresponding to each closed pump are assigned to 0, and the modeling independent variables corresponding to each open power frequency pump are assigned to its rated operating frequency or the average of its historical operating frequencies. The total flow target calculation formula of the parallel water pump group under the current pump-on scheme is obtained. And the total pressure target calculation formula ,in is the independent variable vector composed of the operating frequencies of all the variable frequency pumps under the current pump start scheme; if there is no closed pump and open power frequency pump under the current pump start scheme, the total flow target calculation formula of the parallel water pump group under the current pump start scheme is for , the total pressure target calculation formula of the parallel water pump group under the current pump start plan is for , for ; Then construct the nonlinear programming problem: ;

[0070] in, is the preset water outlet flow error threshold of the parallel water pump group. is the preset target water flow rate of the parallel pump group. The minimum pressure that the parallel pump group needs to meet is preset. Indicates the first of all variable frequency pumps that are turned on under the current pump start plan. The modeling independent variables corresponding to the variable frequency pumps are: The pre-set The minimum operating frequency of a variable frequency pump, The pre-set The maximum operating frequency of a variable frequency pump, Modeling independent variables In the independent variable vector The serial number in The total number of all variable frequency pumps turned on under the current pump start plan; Then solve the nonlinear programming problem and get A solution to ; Then, substitute the solution back into 、 and , calculate the total power consumption, total flow and total pressure of the parallel pump group under the current pump start scheme; and, based on the above solution and arrangement, get and When is the value of the modeling independent variable corresponding to each closed pump and each open power frequency pump under the current pump start scheme, the operating frequency of each water pump in the parallel water pump group under the current pump start scheme is obtained by summarizing, that is, the target operating frequency of each water pump in the parallel water pump group under the current pump start scheme is obtained; Step 140: read two pump start plans from all generated pump start plans, and then execute step L1; Step L1, obtaining the calculated total power consumption, total flow rate, and total pressure of the parallel water pump group under the two current pump start-up plans, and based on the obtained total power consumption, total flow rate, and total pressure, selecting one of the two current pump start-up plans as a temporary pump start-up plan, and then executing step L2; Methods for selecting one of the two current pump start-up plans as a temporary pump start-up plan include: If both current pump start-up plans satisfy formula (1) and formula (2), the one with the smallest total power consumption of the parallel water pump group under the two current pump start-up plans is used as the temporary pump start-up plan, and then step L2 is executed; If only one of the two current pump start-up plans satisfies both formula (1) and formula (2), the pump start-up plan that satisfies both formula (1) and formula (2) is used as a temporary pump start-up plan, and then step L2 is executed; If the two current pump start-up plans do not satisfy formula (1) and formula (2) at the same time, then the two current pump start-up plans The one with the smallest value is used as the temporary pump start solution, and then step L2 is executed. represents absolute value calculation; Formula (1) is: ; Formula (2) is: ; in, Represents the total flow of the parallel pump group, Represents the total pressure of the parallel pump group; Step L2: Continue to read the next pump start plan from all the generated non-empty pump start plans, and combine it with the most recent temporary pump start plan obtained from the current moment to form two pump start plans. Then continue to execute step L1 for these two pump start plans until all the generated non-empty pump start plans are read, and use the temporary pump start plan obtained last in step L1 as the target pump start plan. Step 150: Obtain the target operating frequency of each water pump in the parallel water pump group under the target pump start-up scheme, and obtain the operating frequency of each water pump in the parallel water pump group under the target pump start-up scheme; Step 160 : Control the operation of each water pump in the parallel water pump group according to the operating frequency of each water pump in the parallel water pump group under the target pump start-up plan.

[0071] It can be understood that when step 110 is executed, the polynomial modeling method described in any embodiment of the above polynomial modeling method can be used to construct a total flow polynomial calculation model and a total pressure polynomial calculation model of the parallel water pump group.

[0072] Among them, when constructing the total flow polynomial calculation model of the parallel water pump group, the dependent variable vector is a vector composed of the collected historical total flow of all water pumps in the parallel water pump group.

[0073] When constructing the total pressure polynomial calculation model of the parallel water pump group, the dependent variable vector is a vector composed of the collected historical total pressures of all water pumps in the parallel water pump group.

[0074] It can be understood that when the method described in any one of claims 1 to 7 is used to construct the total flow polynomial calculation model and the total pressure polynomial calculation model of the parallel water pump group, the matrix used is The highest power of the modeling independent variable set for the same Same, the total number of modeling independent variables set Same, the number of polynomial modeling terms set The difference is: In constructing the total flow polynomial calculation model of parallel pump group When the flow rate on the total outlet pipe of the parallel water pump group is collected, the independent variable vector is constructed. The corresponding dependent variable vector; In constructing the total pressure polynomial calculation model of parallel pump group When the pressure on the total outlet pipe of the parallel water pump group is collected, the independent variable vector The corresponding dependent variable vector.

[0075] It can be understood that the pump start-up plan is used to indicate the type and start and stop status of each water pump in the parallel water pump group. For example, a pump start-up plan is: {power frequency pump: on; power frequency pump: off; variable frequency pump: off; variable frequency pump: on; ...}.

[0076] The order of the pumps in each pumping scheme and the independent variable vector The order of each variable in the corresponding variable vector The water pump corresponding to the j-th independent variable in refers to the same water pump in the parallel water pump group as the j-th water pump in each pump start-up scheme.

[0077] If there is no variable frequency pump (ie, a turned-on variable frequency pump) in the pump start plan and the pump start plan contains a variable frequency pump, then in the pump start plan, all variable frequency pumps are turned off.

[0078] like Figure 3 As shown, the present invention provides a water pump control system, which is applied to a parallel water pump group. The parallel water pump group has a total of water pumps, water pumps are connected in parallel, of which Power frequency pumps and Variable frequency pumps, , 、 、 , 、 、 All are integers; the system includes: The first module 301 is used to calculate the power consumption of each power frequency pump in the parallel water pump group using a historical average method; The second module 302 is used to construct a total flow polynomial calculation model of the parallel water pump group using the polynomial modeling method described in any of the above embodiments. , total pressure polynomial calculation model of parallel pump group , and the power consumption polynomial calculation model of each variable frequency pump in the parallel pump group; among them, is an independent variable vector used to construct the total flow polynomial calculation model and the total pressure polynomial calculation model; The third module 303 is used for Generate all non-idle-start pump plans containing at least one open power frequency pump for each water pump; The fourth module 304 is configured to execute, for each of all generated pump start-up plans: The power consumption of each turned-on power frequency pump under the current pump-on scheme is obtained from the calculated power consumptions, and the calculation expression of the power consumption polynomial calculation model of each turned-on variable frequency pump under the current pump-on scheme is obtained from the constructed power consumption polynomial calculation models. The power consumption of each turned-on power frequency pump under the current pump-on scheme obtained is added to the calculation expression of the power consumption polynomial calculation model of each turned-on variable frequency pump under the current pump-on scheme, and the total power consumption calculation formula of the parallel water pump group for the working frequency of each turned-on variable frequency pump under the current pump-on scheme is obtained. ; and in the total flow polynomial calculation model And the total pressure polynomial calculation model In the current pump-on scheme, the modeling independent variables corresponding to each closed pump are assigned to 0, and the modeling independent variables corresponding to each open power frequency pump are assigned to its rated operating frequency or the average of its historical operating frequencies. The total flow target calculation formula of the parallel water pump group under the current pump-on scheme is obtained. And the total pressure target calculation formula ,in is the independent variable vector composed of the operating frequencies of all the variable frequency pumps under the current pump start scheme; if there is no closed pump and open power frequency pump under the current pump start scheme, the total flow target calculation formula of the parallel water pump group under the current pump start scheme is for , the total pressure target calculation formula of the parallel water pump group under the current pump start plan is for , for ; Then construct the nonlinear programming problem: ;

[0079] in, is the preset water outlet flow error threshold of the parallel water pump group. is the preset target water flow rate of the parallel pump group. The minimum pressure that the parallel pump group needs to meet is preset. Indicates the first of all variable frequency pumps that are turned on under the current pump start plan. The modeling independent variables corresponding to the variable frequency pumps are: The pre-set The minimum operating frequency of a variable frequency pump, The pre-set The maximum operating frequency of a variable frequency pump, Modeling independent variables In the independent variable vector The serial number in The total number of all variable frequency pumps turned on under the current pump start plan; Then solve the nonlinear programming problem and get A solution to ; Then, substitute the solution back into 、 and , calculate the total power consumption, total flow and total pressure of the parallel pump group under the current pump start scheme; and, based on the above solution and arrangement, get and When is the value of the modeling independent variable corresponding to each closed pump and each open power frequency pump under the current pump start scheme, the operating frequency of each water pump in the parallel water pump group under the current pump start scheme is obtained by summarizing, that is, the target operating frequency of each water pump in the parallel water pump group under the current pump start scheme is obtained; The fifth module 305 is configured to read two pump start-up plans from all generated pump start-up plans, and then execute step L1; Step L1, obtaining the calculated total power consumption, total flow rate, and total pressure of the parallel water pump group under the two current pump start-up plans, and based on the obtained total power consumption, total flow rate, and total pressure, selecting one of the two current pump start-up plans as a temporary pump start-up plan, and then executing step L2; Methods for selecting one of the two current pump start-up plans as a temporary pump start-up plan include: If both current pump start-up plans satisfy formula (1) and formula (2), the one with the smallest total power consumption of the parallel water pump group under the two current pump start-up plans is used as the temporary pump start-up plan, and then step L2 is executed; If only one of the two current pump start-up plans satisfies both formula (1) and formula (2), the pump start-up plan that satisfies both formula (1) and formula (2) is used as a temporary pump start-up plan, and then step L2 is executed; If the two current pump start-up plans do not satisfy formula (1) and formula (2) at the same time, then the two current pump start-up plans The one with the smallest value is used as the temporary pump start solution, and then step L2 is executed. represents absolute value calculation; Formula (1) is: ; Formula (2) is: ; in, Represents the total flow of the parallel pump group, Represents the total pressure of the parallel pump group; Step L2: Continue to read the next pump start plan from all the generated non-empty pump start plans, and combine it with the most recent temporary pump start plan obtained from the current moment to form two pump start plans. Then continue to execute step L1 for these two pump start plans until all the generated non-empty pump start plans are read, and use the temporary pump start plan obtained last in step L1 as the target pump start plan. The sixth module 306 is used to obtain the target operating frequency of each water pump in the parallel water pump group under the target pump starting scheme, and obtain the operating frequency of each water pump in the parallel water pump group under the target pump starting scheme; The seventh module 307 is used to control the operation of each water pump in the parallel water pump group according to the operating frequency of each water pump in the parallel water pump group under the target pump start-up plan.

[0080] It can be understood that the power consumption of each power frequency pump in the parallel water pump group is the average of the historical power consumption of the corresponding power frequency pumps in the parallel water pump group.

[0081] It is understandable that when the present invention is used, the method described in any of the above embodiments can be used to construct a polynomial calculation model for the total flow of the parallel water pump group. , total pressure polynomial calculation model of parallel pump group , and the power consumption polynomial calculation model of each variable frequency pump in the parallel water pump group.

[0082] It can be understood that during use, when there are both variable frequency pumps and industrial frequency pumps in the parallel water pump group, the user can use the first module to calculate the power consumption of each industrial frequency pump in the parallel water pump group, and use the second module to construct a total flow polynomial calculation model of the parallel water pump group, a total pressure polynomial calculation model of the parallel water pump group, and a power consumption polynomial calculation model of each variable frequency pump in the parallel water pump group.

[0083] When in use, the third module is used for the The fourth module calculates the total power consumption, total flow and total pressure of the parallel water pump group under each non-idle pump scheme generated by the third module, and obtains the target operating frequency of each water pump in the parallel water pump group under each non-idle pump scheme generated by the third module. The fifth module obtains the target pump-starting scheme and the target operating frequency of each water pump in the parallel water pump group under the target pump-starting scheme. Then, the seventh module controls the operation of each water pump in the parallel water pump group according to the operating frequency of each water pump in the parallel water pump group under the target pump-starting scheme obtained by the fifth module.

[0084] The present invention transforms the multi-pump parallel optimization problem into a nonlinear programming problem, minimizing the total energy consumption while satisfying flow, pressure and frequency constraints.

[0085] The same or similar parts in this specification can be referenced to each other.

[0086] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A polynomial modeling method for production process indicators, characterized in that the method include: S1. Construct an initial population. In this population, the total number of chromosomes is Each chromosome contains genes, each gene was used Initialize with random integers on 、 The highest power and total number of pre-set modeling independent variables, Model the number of terms for a pre-set polynomial; S2, using genetic algorithm to iteratively update the initial population, and selecting the optimal chromosome in the population after the last iterative update as the target chromosome; S3, based on the matrix corresponding to the target chromosome and the least squares solution , construct a polynomial calculation model for the target production process indicators: ， in, is the target production process indicator, represents the independent variable vector composed of each modeling independent variable, is a vector The jth independent variable in , express of Power, is a matrix Middle OK Elements of the column, for Middle elements, ; The matrix corresponding to the target chromosome and the least squares solution The method for obtaining comprises steps S4 to S5: S4. Constructing a matrix based on the gene sequence of the target chromosome , ; Matrix-based and matrix Constructing the Matrix , where the matrix For the pre-collected The independent variable matrix constructed by the historical data of , for The total sample size, , , express of power; Remove Matrix The linearly dependent columns in , we get the matrix ; is a matrix The total number of columns; Matrix-based Constructing the equation ,in , A dependent variable vector constructed using historical production process indicator values ​​corresponding to the historical data collected in advance; Solving the equation for the least squares solution yields The least squares solution of ; S5. Based on the matrix Remove the column number of the linearly correlated column removed from The rows with corresponding serial numbers in the matrix are obtained. .

2. The polynomial modeling method for production process indicators according to claim 1, characterized in that: , is the gene sequence, express Middle A gene.

3. The polynomial modeling method for production process indicators according to claim 1 or 2, characterized in that: The iterative update of the initial population using the genetic algorithm includes the following steps: T1. Perform genetic iteration on the chromosomes of the current latest generation population. In each iteration, each chromosome in the current latest generation population is used as a parent chromosome, and the next generation of child chromosomes of each parent chromosome is obtained. For each parent chromosome, the size relationship between its evaluation index and that of its next generation child chromosome is compared, and the chromosome with the smallest evaluation index is selected as the child chromosome of the corresponding parent chromosome; the child chromosomes are collected to construct the next generation population, and then step T1 is continued until the iteration ends; In step T1, the calculation method of the chromosome evaluation index includes: Use the calculation method of step S4 to calculate the matrix corresponding to the current chromosome and the least squares solution , and get the target matrix and the first least squares solution ; calculate and The Euclidean distance between them is used to obtain the evaluation index of the chromosome.

4. The polynomial modeling method for production process indicators according to claim 3, characterized in that: The implementation method of selecting the optimal chromosome in the population after the last iterative update includes: Aggregating the evaluation indicators of each chromosome in the target population to obtain a set of target evaluation indicators; the target population is the population after the last iterative update; Select an evaluation indicator with the smallest value from each target evaluation indicator as the first evaluation indicator; The relevant chromosomes in the target population corresponding to the first evaluation index are selected to obtain the optimal chromosome in the target population.

5. The polynomial modeling method for production process indicators according to claim 1, characterized in that: The removal matrix The linearly dependent columns in , we get the matrix , including the following steps: Step 100, define a matrix U, a column vector d and a matrix for storing Initialize the matrix U to be an empty matrix, initialize d to be an empty vector, and initialize r to be an empty vector; Step 101, traverse the matrix Each column , for the currently traversed Each column , respectively perform the following steps 1011 to 1024, wherein the column represent The hth column of Step 1011, cloning The hth column of ; Step 1012, judgment matrix Is there a non-zero element in : If so, clone the vector Get vector , the matrix With the matrix The h-th column vector of Multiply to get vector , and the vector Divide each element in by the vector The elements at the corresponding positions in the vector are obtained , then use renew The value of , then go to step 1013 to continue execution; represent The transposed matrix of If not, go to step 1013 to continue execution; Step 1013, calculate Dot product with itself ; Step 1014, determine Is it true: If yes, then the vector After transposition, add as new row to At the end of Add as a new element to the column vector At the end of Add it to the end of vector r as a new element; otherwise, do nothing; is the pre-set allowable error threshold; Step 102, after step 101 is executed, according to vector The serial number stored in The corresponding column number in the generator matrix , where the matrix Representative Matrix No. List, is a vector Middle elements.

6. The polynomial modeling method for production process indicators according to claim 1, characterized in that: The target production process indicator is the total flow rate of the parallel water pump group, the total pressure of the parallel water pump group or the power consumption of a single variable frequency pump in the parallel water pump group; when the target production process indicator is the total flow rate or total pressure of the parallel water pump group, the pre-set modeling independent variable is the operating frequency of each water pump in the parallel water pump group; when the target production process indicator is the power consumption of a single variable frequency pump in the parallel water pump group, the pre-set modeling independent variable is the operating frequency of the corresponding variable frequency pump in the parallel water pump group.

7. The polynomial modeling method for production process indicators according to claim 1, characterized in that: The target production process indicators are electricity consumption per ton of steel, steam consumption per ton of steel or processing time per ton of steel of the steel production line, and the pre-set modeling independent variables are raw material thickness, raw material width, steel processing thickness and steel processing width of steel produced by the steel production line.

8. A polynomial modeling system for production process indicators, characterized in that: The system includes: The population construction module is used to construct the initial population, in which the total number of chromosomes is Each chromosome contains genes, each gene was used Initialize with random integers on 、 The highest power and total number of pre-set modeling independent variables, Model the number of terms for a pre-set polynomial; A chromosome selection module is connected to the population construction module and is used to iteratively update the initial population using a genetic algorithm and select the optimal chromosome in the population after the last iterative update as the target chromosome; Model building module, connected to the chromosome selection module, for the matrix corresponding to the target chromosome and the least squares solution , construct a polynomial calculation model for the target production process indicators: ， in, is the target production process indicator, represents the independent variable vector composed of each modeling independent variable, is a vector The jth independent variable in , express of Power, is a matrix Middle OK Elements of the column, for Middle elements, ; Among them, the matrix corresponding to the target chromosome and the least squares solution Methods for obtaining include: Step 1: Construct a matrix based on the gene sequence of the target chromosome , ; Matrix-based and matrix Constructing the Matrix , where the matrix For the pre-collected The independent variable matrix constructed by the historical data of , for The total sample size, , , express of power; Remove Matrix The linearly dependent columns in , we get the matrix ; is a matrix The total number of columns; Matrix-based Constructing the equation ,in , A dependent variable vector constructed using historical production process indicator values ​​corresponding to the historical data collected in advance; Solving the equation for the least squares solution yields The least squares solution of ; Step 2: Based on the matrix Remove the column number of the linearly correlated column removed from The rows with corresponding serial numbers in the matrix are obtained. .

9. A water pump control method, characterized in that: The method is applied to a parallel pump group with a total of water pumps, water pumps are connected in parallel, of which Power frequency pumps and Variable frequency pumps, , 、 、 , 、 、 All are integers; methods include: Step 100: Calculate the power consumption of each power frequency pump in the parallel water pump group using the historical average method; if there is no power frequency pump in the parallel water pump group, skip this step; Step 110: If there is a variable frequency pump in the parallel water pump group, use the method described in any one of claims 1 to 7 to construct a polynomial calculation model for the total flow of the parallel water pump group. , total pressure polynomial calculation model , and the power consumption polynomial calculation model of each variable frequency pump in the parallel pump group; among them, is the independent variable vector composed of the operating frequency of each water pump in the parallel water pump group as the modeling independent variable. If there is no variable frequency pump in the parallel water pump group, skip this step; Step 120 is the Generate all non-idle-start pump plans containing at least one open power frequency pump for each water pump; Step 130: For each of all generated pump start-up plans, execute: The power consumption of each turned-on power frequency pump under the current pump-on scheme is obtained from the power consumptions calculated in step 100, and the calculation expression of the power consumption polynomial calculation model of each turned-on variable frequency pump under the current pump-on scheme is obtained from the power consumption polynomial calculation models constructed in step 110. The power consumption of each turned-on power frequency pump under the current pump-on scheme obtained is added to the calculation expression of the power consumption polynomial calculation model of each turned-on variable frequency pump under the current pump-on scheme, and the total power consumption calculation formula of the parallel water pump group with respect to the working frequency of each turned-on variable frequency pump under the current pump-on scheme is obtained. ; and in the total flow polynomial calculation model And the total pressure polynomial calculation model In the current pump-on scheme, the modeling independent variables corresponding to each closed pump are assigned to 0, and the modeling independent variables corresponding to each open power frequency pump are assigned to its rated operating frequency or the average of its historical operating frequencies. The total flow target calculation formula of the parallel water pump group under the current pump-on scheme is obtained. And the total pressure target calculation formula ,in is the independent variable vector composed of the operating frequencies of all the variable frequency pumps under the current pump start scheme; if there is no closed pump and open power frequency pump under the current pump start scheme, the total flow target calculation formula of the parallel water pump group under the current pump start scheme is for , the total pressure target calculation formula of the parallel water pump group under the current pump start plan is for , for ; Then construct the nonlinear programming problem: ； in, is the preset water outlet flow error threshold of the parallel water pump group. is the preset target water flow rate of the parallel pump group. The minimum pressure that the parallel pump group needs to meet is preset. Indicates the first of all variable frequency pumps that are turned on under the current pump start plan. The modeling independent variables corresponding to the variable frequency pumps are: The pre-set The minimum operating frequency of a variable frequency pump, The pre-set The maximum operating frequency of a variable frequency pump, Modeling independent variables In the independent variable vector The serial number in The total number of all variable frequency pumps turned on under the current pump start plan; Then solve the nonlinear programming problem and get A solution to ; Then, substitute the solution back into 、 and , calculate the total power consumption, total flow and total pressure of the parallel pump group under the current pump start scheme; and, based on the above solution and arrangement, get and When is the value of the modeling independent variable corresponding to each closed pump and each open power frequency pump under the current pump start scheme, the operating frequency of each water pump in the parallel water pump group under the current pump start scheme is obtained by summarizing, that is, the target operating frequency of each water pump in the parallel water pump group under the current pump start scheme is obtained; Step 140: read two pump start plans from all generated pump start plans, and then execute step L1; Step L1, obtaining the calculated total power consumption, total flow rate, and total pressure of the parallel water pump group under the two current pump start-up plans, and based on the obtained total power consumption, total flow rate, and total pressure, selecting one of the two current pump start-up plans as a temporary pump start-up plan, and then executing step L2; Methods for selecting one of the two current pump start-up plans as a temporary pump start-up plan include: If both current pump start-up plans satisfy formula (1) and formula (2), the one with the smallest total power consumption of the parallel water pump group under the two current pump start-up plans is used as the temporary pump start-up plan, and then step L2 is executed; If only one of the two current pump start-up plans satisfies both formula (1) and formula (2), the pump start-up plan that satisfies both formula (1) and formula (2) is used as a temporary pump start-up plan, and then step L2 is executed; If the two current pump start-up plans do not satisfy formula (1) and formula (2) at the same time, then the two current pump start-up plans The one with the smallest value is used as the temporary pump start solution, and then step L2 is executed. represents absolute value calculation; Formula (1) is: ; Formula (2) is: ; in, Represents the total flow of the parallel pump group, Represents the total pressure of the parallel pump group; Step L2: Continue to read the next pump start plan from all the generated non-empty pump start plans, and combine it with the most recent temporary pump start plan obtained from the current moment to form two pump start plans. Then continue to execute step L1 for these two pump start plans until all the generated non-empty pump start plans are read, and use the temporary pump start plan obtained last in step L1 as the target pump start plan. Step 150: Obtain the target operating frequency of each water pump in the parallel water pump group under the target pump start-up scheme, and obtain the operating frequency of each water pump in the parallel water pump group under the target pump start-up scheme; Step 160 : Control the operation of each water pump in the parallel water pump group according to the operating frequency of each water pump in the parallel water pump group under the target pump start-up plan.

10. A water pump control system, characterized in that: The system is applied to parallel pump groups, which have water pumps, water pumps are connected in parallel, of which Power frequency pumps and Variable frequency pumps, , 、 、 , 、 、 All are integers; the system includes: The first module is used to calculate the power consumption of each power frequency pump in the parallel water pump group using the historical average method; The second module is used to construct a total flow polynomial calculation model of the parallel water pump group using the method described in any one of claims 1 to 7 , total pressure polynomial calculation model of parallel pump group , and the power consumption polynomial calculation model of each variable frequency pump in the parallel pump group; among them, is an independent variable vector used to construct the total flow polynomial calculation model and the total pressure polynomial calculation model; The third module is used for Generate all non-idle-start pump plans containing at least one open power frequency pump for each water pump; The fourth module is used to execute the following steps for each of the generated pump start-up plans: The power consumption of each turned-on power frequency pump under the current pump-on scheme is obtained from the calculated power consumptions, and the calculation expression of the power consumption polynomial calculation model of each turned-on variable frequency pump under the current pump-on scheme is obtained from the constructed power consumption polynomial calculation models. The power consumption of each turned-on power frequency pump under the current pump-on scheme obtained is added to the calculation expression of the power consumption polynomial calculation model of each turned-on variable frequency pump under the current pump-on scheme, and the total power consumption calculation formula of the parallel water pump group for the working frequency of each turned-on variable frequency pump under the current pump-on scheme is obtained. ; and in the total flow polynomial calculation model And the total pressure polynomial calculation model In the current pump-on scheme, the modeling independent variables corresponding to each closed pump are assigned to 0, and the modeling independent variables corresponding to each open power frequency pump are assigned to its rated operating frequency or the average of its historical operating frequencies. The total flow target calculation formula of the parallel water pump group under the current pump-on scheme is obtained. And the total pressure target calculation formula ,in is the independent variable vector composed of the operating frequencies of all the variable frequency pumps under the current pump start scheme; if there is no closed pump and open power frequency pump under the current pump start scheme, the total flow target calculation formula of the parallel water pump group under the current pump start scheme is for , the total pressure target calculation formula of the parallel water pump group under the current pump start plan is for , for ; Then construct the nonlinear programming problem: ； in, is the preset water outlet flow error threshold of the parallel water pump group. is the preset target water flow rate of the parallel pump group. The minimum pressure that the parallel pump group needs to meet is preset. Indicates the first of all variable frequency pumps that are turned on under the current pump start plan. The modeling independent variables corresponding to the variable frequency pumps are: The pre-set The minimum operating frequency of a variable frequency pump, The pre-set The maximum operating frequency of a variable frequency pump, Modeling independent variables In the independent variable vector The serial number in The total number of all variable frequency pumps turned on under the current pump start plan; Then solve the nonlinear programming problem and get A solution to ; Then, substitute the solution back into 、 and , calculate the total power consumption, total flow and total pressure of the parallel pump group under the current pump start scheme; and, based on the above solution and arrangement, get and When is the value of the modeling independent variable corresponding to each closed pump and each open power frequency pump under the current pump start scheme, the operating frequency of each water pump in the parallel water pump group under the current pump start scheme is obtained by summarizing, that is, the target operating frequency of each water pump in the parallel water pump group under the current pump start scheme is obtained; A fifth module is configured to read two pump start-up plans from all generated pump start-up plans, and then execute step L1; Step L1, obtaining the calculated total power consumption, total flow rate, and total pressure of the parallel water pump group under the two current pump start-up plans, and based on the obtained total power consumption, total flow rate, and total pressure, selecting one of the two current pump start-up plans as a temporary pump start-up plan, and then executing step L2; Methods for selecting one of the two current pump start-up plans as a temporary pump start-up plan include: If both current pump start-up plans satisfy formula (1) and formula (2), the one with the smallest total power consumption of the parallel water pump group under the two current pump start-up plans is used as the temporary pump start-up plan, and then step L2 is executed; If only one of the two current pump start-up plans satisfies both formula (1) and formula (2), the pump start-up plan that satisfies both formula (1) and formula (2) is used as a temporary pump start-up plan, and then step L2 is executed; If the two current pump start-up plans do not satisfy formula (1) and formula (2) at the same time, then the two current pump start-up plans The one with the smallest value is used as the temporary pump start solution, and then step L2 is executed. represents absolute value calculation; Formula (1) is: ; Formula (2) is: ; in, Represents the total flow of the parallel pump group, Represents the total pressure of the parallel pump group; Step L2: Continue to read the next pump start plan from all the generated non-empty pump start plans, and combine it with the most recent temporary pump start plan obtained from the current moment to form two pump start plans. Then continue to execute step L1 for these two pump start plans until all the generated non-empty pump start plans are read, and use the temporary pump start plan obtained last in step L1 as the target pump start plan. The sixth module is used to obtain the target operating frequency of each water pump in the parallel water pump group under the target pump start-up plan, and obtain the operating frequency of each water pump in the parallel water pump group under the target pump start-up plan; The seventh module is used to control the operation of each water pump in the parallel water pump group according to the operating frequency of each water pump in the parallel water pump group under the target pump start-up plan.

Images