Decision control system and method based on temperature field and atmosphere analysis

Abstract

The invention discloses a decision control system and method based on temperature field and atmosphere analysis, and relates to the technical field of intelligent control. A monitoring module in the system obtains a temperature set and a gas concentration set; the trend analysis module constructs a three-dimensional model and establishes a coordinate system, a variation function is combined with a Kriging equation to supplement a temperature set to generate an expanded temperature set, the three-dimensional model is discretized into a plurality of nodes and is matched with points in the expanded temperature set through nearest neighbor search, a heat conduction equation and a convection boundary are set, and a finite element matrix equation set is constructed; solving to obtain a temperature field model and carrying out statistics to generate a temperature trend curve; a decision module compares the temperature trend curve, the gas concentration set and a process standard to obtain an abnormal section number, and analyzes a section temperature difference and a concentration error vector of the abnormal section number by adopting a fuzzy PID control algorithm to generate an adjustment vector; and the execution module performs automatic adjustment based on the adjustment vector, so that intelligent decision adjustment based on temperature field and atmosphere conjoint analysis is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent control, and particularly to a decision control system and method based on temperature field and atmosphere analysis. Background Art

[0002] In the field of modern industrial production, many technological processes highly rely on the precise control of the temperature field and atmosphere of functional equipment. Precise temperature and suitable gas environment not only relate to product quality, but also directly affect production efficiency and equipment life. Therefore, it is urgent to develop an efficient monitoring and control system.

[0003] Traditional monitoring and control systems have defects. Traditional control systems often conduct trend analysis and intelligent regulation based on temperature data or gas components collected by sensors, lacking means for analyzing the spatial variation characteristics of temperature, having inaccurate temperature trend analysis, ignoring the associated effects of temperature field and atmosphere in industrial production, and having one-sided regulation means. Summary of the Invention

[0004] Aiming at the deficiencies of the existing technology, the present invention proposes a decision control system and method based on temperature field and atmosphere analysis, realizing the fine modeling of the temperature field model and the intelligent decision adjustment based on temperature field and atmosphere analysis.

[0005] The technical solution for achieving the object of the present invention is as follows:

[0006] A decision control system based on temperature field and atmosphere analysis includes a monitoring module, a trend analysis module, a decision module, and an execution module;

[0007] The monitoring module periodically sets monitoring times, and at each monitoring time, it monitors and obtains a temperature set and a gas concentration set through a temperature measurement device group and a gas detection device, where the temperature set and the gas concentration set respectively include temperature sequences of functional segments and gas concentration vectors of

[0008] functional segments, The trend analysis module constructs a three-dimensional model of the functional equipment and establishes a Cartesian space coordinate system, uses a variogram to describe the spatial variation characteristics of temperature, performs spatial expansion on the measured temperatures in the temperature set to generate an expanded temperature set , based on the thermal conductivity of the functional device Set up the heat conduction equation and define the convection boundary , use the finite element method to construct the finite element matrix equation and solve it by the conjugate gradient method to obtain the temperature field model , for the temperature field model At The statistical analysis is carried out in the axial direction to generate the temperature trend curve , where is the total number of nodes, and the temperature field model includes the temperatures of the nodes, and the temperature trend curve reflects the temperature change trends of the functional segments, is the coordinate variable in the axial direction;

[0009] The decision-making module compares the temperature trend curve and the gas concentration set with the corresponding process standards respectively to obtain the abnormal segment numbers, takes the segment temperature difference and the concentration error vector of each abnormal segment number as the input of the fuzzy PID control algorithm, uses fuzzy logic to dynamically adjust the parameters of the PID control algorithm, and outputs the adjustment vector corresponding to each abnormal segment number. Among them, the process standards include the standard temperature trend curve and the gas standard concentration set , and the adjustment vector includes the fuel valve adjustment amount and the air valve adjustment amount;

[0010] The execution module automatically adjusts the heating equipment of the corresponding functional segment based on the adjustment vector of each abnormal segment number.

[0011] Furthermore, the monitoring module includes a temperature measurement unit and a concentration detection unit;

[0012] The temperature measurement unit obtains the measured temperatures of the temperature measurement devices, sorts the measured temperatures with the same segment number according to the device number order to construct a temperature sequence, integrates the temperature sequences of the functional segments to construct a temperature set and transmits it to the trend analysis module, is the total number of functional segments, is the total number of temperature measurement devices;

[0013] The concentration detection unit obtains the gas concentration vectors measured by the gas detection devices in the functional segments, integrates the gas concentration vectors of the functional segments to construct a gas concentration set and transmits it to the decision-making module. Among them, the gas concentration vector of the th functional segment including the oxygen concentration, carbon dioxide concentration, carbon monoxide concentration, and nitrogen concentration of the function segments.

[0014] Specifically, the gas detection device uses a semiconductor laser gas analyzer. Multiple semiconductor lasers with different wavelengths respectively emit lasers that match the specific absorption spectra of oxygen, carbon dioxide, carbon monoxide, and nitrogen. By detecting the laser intensity of each wavelength of laser after passing through the mixed gas, the oxygen concentration, carbon dioxide concentration, carbon monoxide concentration, and nitrogen concentration are respectively determined according to the Beer-Lambert law.

[0015] Specifically, the three-dimensional model of the functional device needs to be pre-constructed according to the geometric parameters of the function segments through a three-dimensional modeling method. The three-dimensional modeling method includes surface modeling, point cloud reverse modeling, image reverse modeling, and three-dimensional scanning modeling. The central axis of the three-dimensional model is used as the

[0016] axis to construct a Cartesian space coordinate system. Furthermore, the variogram is used to describe the spatial variation characteristics of temperature. Based on the Kriging equation, the measured temperatures in the temperature set are spatially extended to generate an extended temperature set

[0017] including the following specific steps: in the temperature measurement device coordinates of the temperature measurement devices in the temperature set are and the temperature measurement device coordinates constitute a temperature measurement device pair and the corresponding spatial distance is calculated. A total of temperature measurement device pairs and the corresponding spatial distances are obtained. is the temperature measurement device coordinates of the

[0018] Calculate the variogram value of the temperature measurement device pair The variogram value is the semi-variance of the temperatures at the temperature measurement device coordinates and the temperature measurement device coordinates used to quantitatively describe the variability of the spatial distance The variogram is obtained by fitting the existing spatial distances and variogram values through the least squares method;

[0019] Based on the principles of unbiasedness and minimum variance estimation for the coordinates to be estimated Establish the Kriging equations to obtain the optimal weight coefficient vector The optimal weight coefficient vector Satisfies that the coordinates of the point to be estimated With the coordinates of the nth temperature measuring device The variogram value is equal to the inner product of the optimal weight coefficient vector With the coordinates of the mth temperature measuring device The sequence of variogram values The optimal weight coefficient vector Satisfies that the sum of the optimal weight coefficients is 1;

[0020] Solve to obtain the optimal weight coefficient vector Of the coordinates of the point to be estimated The coordinates of the point to be estimated The estimated temperature Is the inner product of the known temperature vector formed by arranging the In the temperature set n measured temperatures and the optimal weight coefficient vector ;

[0021] Uniformly select k points to be estimated on the surface of the three-dimensional model and calculate the estimated temperature And merge with the temperature set To generate an extended temperature set The extended temperature set Includes the temperatures of the m points on the surface of the functional device, m is the total number of points.

[0022] Specifically, the grid segmentation technology eliminates the detailed structures in the three-dimensional model that have little impact on the temperature field modeling, generates a simplified three-dimensional model, translates each coordinate axis of the coordinate system along the directions of the other two coordinate axes to generate a set of parallel lines. The lines in the three sets of parallel lines of the three coordinate axes intersect each other, and the intersection point of any two lines is defined as a node. A total of n nodes are obtained and a hollow hexahedron surrounded by 8 nodes is defined as a unit. A total of e units are obtained, e is the total number of units. For the temperature of each point in the extended temperature set Perform the nearest neighbor search according to the point coordinates, retrieve the node with the smallest spatial distance from the point coordinates and assign the corresponding point temperature. Define all the nodes with assigned temperatures as known surface nodes.

[0023] Furthermore, based on the thermal conductivity of the functional device Set up the heat conduction equation and define the convection boundary , the finite element method is used to construct the finite element matrix equation and the conjugate gradient method is used to solve the temperature field model , including the following specific steps:

[0024] Determine thermal conductivity , set up the heat conduction equation to describe the heat flux density With temperature gradient The ratio is equal to the thermal conductivity The opposite of

[0025] Defining Convection Boundaries , the heat flux density of convective heat transfer is described by Newton's law of cooling and surface temperature and ambient temperature The difference is proportional to the convective heat transfer coefficient. ;

[0026] consider The thermal balance inside each unit is calculated based on Fourier's law. The heat conduction equation in each unit is analyzed and the weighted residual method is used to construct the The finite element equations of the unit describe the The second-order temperature gradient within the unit volume Integrals on the cell boundary Heat flux At the unit boundary Area The integral on the graph shows a proportional relationship, which is equal to the thermal conductivity , ;

[0027] Use the The interpolation of the shape function of each node in the element describes the The spatial coordinates within a unit are Temperature indication , the shape function is determined according to the geometric characteristics and node distribution of the element, 、 and They are Axis direction, Axis direction and Coordinate variables in the axis direction;

[0028] The first The temperature of each unit is represented Substitute The finite element equations of the th element are derived through integral operations and matrix transformations, and are organized into finite element matrix equations to describe the stiffness matrix of the th element, which is equal to the load vector

[0029] when multiplied by the nodal temperature vector ; The finite element matrix equations of the th element are updated according to the assembly rules, which are established based on the nodal sharing relationship of adjacent elements. The values of the stiffness matrices and load vectors of adjacent elements corresponding to the shared nodes are summed and replaced, and integrated into a system of finite element matrix equations and solved using the conjugate gradient method to obtain the nodal temperature vector of the

[0030] th element, generating a temperature field model Specifically, statistical analysis of the temperature field model in the axis direction can reflect the temperature change trend of different functional segments. The temperatures of the same coordinate variable in the temperature field model are averaged to generate the average temperature corresponding to the coordinate variable . All coordinate variables are traversed to obtain the corresponding average temperatures, and the temperature trend curve

[0031] is obtained by least squares fitting. Furthermore, by comparing the temperature trend curve with the standard temperature trend curve

[0032] the abnormal segment numbers are obtained, including the following specific steps: Determine the range of segment coordinate variables of the th functional segment in the

[0033] axis direction according to the three-dimensional model; - Calculate the mean square error of the temperature trend curve and the standard temperature trend curve in the th functional segment. The mean square error of the th functional segment is the definite integral of the square of the error between the temperature trend curve and the standard temperature trend curve divided by the length of the segment coordinate variable range in the axis direction, , and are respectively the th functional segment in the The starting coordinate variable and the ending coordinate variable in the axial direction ;

[0034] If the temperature mean square error is greater than or equal to the temperature mean square error threshold, mark it as the abnormal segment number. If the temperature mean square error is less than the temperature mean square error threshold, no processing is performed.

[0035] Furthermore, compare the gas concentration set with the gas standard concentration set to obtain the abnormal segment number, including the following specific steps:

[0036] Obtain the gas concentration vector of the nth functional segment in the gas concentration set and the gas standard concentration vector ;

[0037] Calculate the concentration mean square error of the nth functional segment. The concentration mean square error is the square of the two-norm of the concentration error vector of the nth functional segment. The concentration error vector of the [[ID=…]]nth

[0038] functional segment is equal to the gas concentration vector minus the gas standard concentration vector ; If the concentration mean square error

[0039] is greater than or equal to the concentration mean square error threshold, mark it as the abnormal segment number. If the concentration mean square error is less than the concentration mean square error threshold, no processing is performed. Specifically, the segment temperature difference of the abnormal functional segment is the definite integral of the error between the temperature trend curve

[0040] and the standard temperature trend curve within the corresponding segment independent variable range divided by the segment independent variable length. The concentration error vector of the abnormal functional segment is the gas concentration vector of the abnormal functional segment minus the gas standard concentration vector. The concentration error vector includes oxygen concentration error, carbon dioxide concentration error, carbon monoxide concentration error, and nitrogen concentration error.

[0040] Specifically, the fuzzy PID control algorithm includes fuzzy inference. The input variables of the fuzzy inference include the temperature difference between segments of the abnormal function segment, oxygen concentration error, carbon dioxide concentration error, carbon monoxide concentration error, and nitrogen concentration error. The output variables of the fuzzy inference include the proportional factor, integral factor, and differential factor of the PID control algorithm. Determine the domain of each input variable or output variable, and define fuzzy sets including "negative large", "negative small", "zero", "positive medium", and "positive large". Based on experience, formulate fuzzy rules, and establish the corresponding relationship between the input fuzzy set group and the output fuzzy set group. For each fuzzy rule, calculate the membership degree of each input variable belonging to the corresponding fuzzy set in the input fuzzy set group through the Gaussian membership function, and obtain the activation strength of the fuzzy rule through the AND operation. Perform the OR operation on the activation strengths of all fuzzy rules to determine the membership degrees of the output fuzzy set group and each output variable. Use the centroid method for defuzzification, and output the proportional factor, integral factor, and differential factor of the PID control algorithm.

[0041] Specifically, the fuzzy PID control algorithm includes the PID control algorithm. The temperature difference between segments of the abnormal function segment, oxygen concentration error, carbon dioxide concentration error, carbon monoxide concentration error, and nitrogen concentration error are weighted and summed to generate a comprehensive error. Obtain the proportional factor, integral factor, and differential factor. Multiply the comprehensive error by the proportional factor to generate a proportional control quantity and determine the remaining comprehensive error. Multiply the integral operation result of the remaining comprehensive error by to generate an integral control quantity and update the remaining comprehensive error. Repeat the integral adjustment until the remaining comprehensive error is less than the comprehensive error threshold and then stop. The differential adjustment multiplies the change rate of the comprehensive error between two adjacent integral adjustments by the differential factor to generate a differential correction control quantity. The number of differential adjustment times is the same as the number of integral adjustment times. The sum of the proportional control quantity, multiple integral control quantities, and multiple differential control quantities is used as the control quantity. Consider the stoichiometric ratio to split and obtain the fuel valve adjustment quantity and air valve adjustment quantity, and combine them into the adjustment vector corresponding to the abnormal function segment.

[0042] The decision-making control method based on temperature field and atmosphere analysis, and the implementation of the decision-making control system based on temperature field and atmosphere analysis, include the following specific steps:

[0043] At each monitoring moment, obtain the temperature set and gas concentration set through the temperature measurement equipment group and gas detection equipment;

[0044] Based on the geometric parameters of the function segments, construct a three-dimensional model and establish a Cartesian space coordinate system, where is the total number of function segments;

[0045] Use the variogram to describe the spatial variation characteristics of temperature, and based on the Kriging equation, for the temperature set Spatially expand the measured temperature in it to generate an expanded temperature set ;

[0046] Adopt the grid segmentation technology to simplify and discretize the three-dimensional model into nodes, and assign the temperature of each point in the expanded temperature set to the nodes matching the corresponding point coordinates through nearest neighbor search, where

[0047] is the total number of nodes; Obtain the ambient temperature , set the heat conduction equation based on the thermal conductivity of the functional device , and define the convection boundary, ;

[0048] Build a finite element matrix equation set through the finite element method and solve it through the conjugate gradient method to obtain the temperature field model ; Perform statistical analysis on the temperature field model in the axis direction to generate a temperature trend curve and compare it with the standard temperature trend curve , and determine the abnormal segment number in combination with the comparison result of the gas concentration set

[0049] Use the segment temperature difference and concentration error vector of each abnormal segment number as the input of the fuzzy PID control algorithm, dynamically adjust the parameters of the fuzzy PID control algorithm by using fuzzy logic, output the adjustment vector corresponding to each abnormal segment number and perform adjustment.

[0050] Compared with the prior art, the present invention constructs a three-dimensional model of a functional device and establishes a coordinate system, uses a variogram to describe the spatial variation of the temperature in the temperature set on the surface of the functional device, supplements and generates an expanded temperature set through the Kriging equation to support temperature field modeling, simplifies and discretizes the three-dimensional model into several nodes and matches them with the points in the expanded temperature set through nearest neighbor search, sets the heat conduction equation based on the thermal conductivity and defines the convection boundary in combination with the ambient temperature, constructs a finite element matrix equation set through the finite element method, solves the temperature of the unmatched nodes through the conjugate gradient method, constructs a detailed temperature field model of the functional device and fits to obtain the temperature trend curve, providing accurate data support for subsequent intelligent decision-making; compares the temperature trend curve and the gas concentration set with the process standard to determine the abnormal functional segment, and uses the fuzzy PID control algorithm to analyze the segment temperature difference and concentration error vector of the abnormal functional segment to generate the corresponding adjustment vector and perform real-time adjustment, realizing the accurate analysis of the temperature trend and the intelligent decision-making adjustment based on the joint analysis of the temperature field and the atmosphere. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1It is the principle structure diagram of the decision control system based on temperature field and atmosphere analysis in the present invention;

[0052] Figure 2 It is the flowchart for generating the extended temperature set in the present invention;

[0053] Figure 3 It is the flowchart for obtaining the abnormal segment number by comparing the gas concentration set with the gas standard concentration set in the present invention;

[0054] Figure 4 It is the flowchart of the decision control method based on temperature field and atmosphere analysis in the present invention. Specific embodiments

[0055] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0056] Embodiment 1

[0057] As Figure 1 shown, a specific embodiment of the present invention discloses a decision control system based on temperature field and atmosphere analysis, including a monitoring module, a trend analysis module, a decision module and an execution module;

[0058] The monitoring module sets the monitoring time periodically at the monitoring interval and monitors and obtains the temperature set and the gas concentration set through the temperature measurement equipment group and gas detection equipment arranged in segments at each monitoring time. Among them, the temperature set and the gas concentration set respectively include temperature sequences of functional segments and gas concentration vectors of

[0059] functional segments. is the total number of functional segments. For example, the functional segments of a continuous annealing furnace include a heating section, a heat preservation section and a slow cooling section, and the corresponding segment numbers can be marked as 1, 2 and 3 in sequence; The trend analysis module constructs a three-dimensional model of the functional equipment and establishes a Cartesian space coordinate system based on the geometric parameters of ​​​​Set up the heat conduction equation and define the convection boundary , construct a finite element matrix equation by the finite element method, and solve the finite element matrix equation by the conjugate gradient method to obtain the temperature field model , for the temperature field model In the axial direction, perform statistical analysis to generate a temperature trend curve , where is the total number of nodes, the coordinates of the temperature measurement device are determined according to the installation parameters of the temperature measurement device and the Cartesian space coordinate system, and the temperature field model includes the temperatures of the nodes, and the temperature trend curve reflects the temperature change trends of the functional segments, is the coordinate variable in the axial direction. For example, the temperature trend curve of a continuous annealing furnace reflects the temperature change trends of the heating section, the heat preservation section and the slow cooling section;

[0060] The decision-making module compares the temperature trend curve with the standard temperature trend curve and compares the gas concentration set with the gas standard concentration set to obtain the abnormal segment numbers, and takes the segment temperature difference and the concentration error vector of each abnormal segment number as the input of the fuzzy PID control algorithm, uses fuzzy logic to dynamically adjust the parameters of the PID control algorithm, and outputs the adjustment vector corresponding to each abnormal segment number. Among them, the standard temperature trend curve and the gas standard concentration set are the process standards predefined for the functional equipment. The adjustment vector includes the fuel valve adjustment amount and the air valve adjustment amount. For example, if the heating section, the heat preservation section and the slow cooling section of a continuous annealing furnace are heated by burning natural gas, the fuel valve adjustment amount and the air valve adjustment amount are actually the adjustment amounts of the valve opening degrees to control the combustion ratio of natural gas and air;

[0061] The execution module automatically adjusts the fuel valve opening degree and the air valve opening degree of the heating equipment of the corresponding functional segment based on the adjustment vector of each abnormal segment number. For example, the heating equipment of the heating section, the heat preservation section and the slow cooling section of a continuous annealing furnace includes double regenerative burners and high-speed burner burners in the heating section.

[0062] Furthermore, the monitoring module includes a temperature measurement unit and a concentration detection unit;

[0063] The temperature measurement unit obtains the measured temperatures of each temperature measurement device in the functional segments, arranges the measured temperatures with the same segment number in the order of the device numbers to construct a temperature sequence, and integrates Construct a temperature set from the temperature sequences of each functional segment and transmit it to the trend analysis module, where is the temperature sequence of the th functional segment is the th measurement temperature collected by the temperature measuring device with the measurement group serial number in the temperature measuring device group of the th functional segment, is the corresponding device number, and are the maximum device numbers counted for the first functional segments and the first functional segments respectively, , is the maximum measurement group serial number of the temperature measuring device group of the th functional segment;

[0064] The concentration detection unit obtains the gas concentration vectors measured by the gas detection devices in functional segments, integrates the gas concentration vectors of functional segments to construct a gas concentration set and transmits it to the decision-making module, where is the gas concentration vector of the th functional segment, 、 、 and are the oxygen concentration, carbon dioxide concentration, carbon monoxide concentration and nitrogen concentration measured by the gas detection device of the th functional segment respectively. The device number of the gas detection device of the th functional segment is , is the total number of temperature measuring devices, therefore, temperature measuring devices of the th functional segment have occupied device numbers from 1 to , the device numbers of gas detection devices need to start from Each functional segment requires 1 gas detection device. The concentration detection unit includes a total of [[ID=7The device numbers occupy from to .

[0065] Specifically, when applied to a continuous annealing furnace, it is known that the corresponding segment numbers of the heating section, heat preservation section and slow cooling section are 1, 2 and 3 respectively, and 4 temperature sensors are installed at the socket joints of the heating section, heat preservation section and slow cooling section. Then the total number of measurement group serial numbers corresponding to the first functional segment to the third functional segment 、 and All are 4. At this time, the temperature measurement unit includes a total of 12 temperature measurement devices, that is , and the corresponding device numbers occupy 1 to 12. Since gas detection devices are required in both the heating section, the heat preservation section, and the slow cooling section, the concentration detection unit includes a total of 3 gas detection devices, and the corresponding device numbers occupy 13 to 15. Taking the temperature sensor with the temperature measurement group number 3 in the temperature measurement device group of the third functional section as an example, the corresponding device number is , and the gas detection device number corresponding to the gas detection device in the second functional section is .

[0066] Specifically, the gas detection device uses a semiconductor laser gas analyzer. Through multiple semiconductor lasers with different wavelengths, lasers that precisely match the specific absorption spectra of oxygen, carbon dioxide, carbon monoxide, and nitrogen are emitted respectively. When the laser passes through the mixed gas including these gases, each gas absorbs the laser of the corresponding wavelength respectively, causing the laser intensity to weaken. By detecting the changes in the intensities of the lasers of each wavelength, the oxygen concentration, carbon dioxide concentration, carbon monoxide concentration, and nitrogen concentration are determined respectively according to the Beer-Lambert law.

[0067] Specifically, the three-dimensional model of the functional device needs to be pre-constructed by a three-dimensional modeling method based on the geometric parameters of the functional sections. The three-dimensional model can accurately reflect the actual physical structure of the functional device. The three-dimensional modeling method includes surface modeling, point cloud reverse modeling, image reverse modeling, and three-dimensional scanning modeling. After obtaining the three-dimensional model, a Cartesian space coordinate system is established. Among them, the axis is defined as the central axis of the three-dimensional model, and the other two coordinate axes and the origin can be defined by oneself.

[0068] As Figure 2 shown, further, the variogram is used to describe the spatial variation characteristics of the temperature in the functional device, and the measured temperatures in the temperature set are spatially extended on the surface of the three-dimensional model based on the Kriging equation to generate an extended temperature set , including the following specific steps:

[0069] Define the temperature measurement device coordinates of the temperature measurement devices in the temperature set as . Combine the temperature measurement device coordinate of the th temperature measurement device and the temperature measurement device coordinate of the th temperature measurement device to form a temperature measurement device pair and calculate the corresponding spatial distance . A total of temperature measurement device pairs and the corresponding spatial distances are obtained. ,The coordinates of the temperature measuring device are determined based on the installation parameters of the temperature measuring device and the Cartesian space coordinate system;

[0070] Calculate the temperature measurement equipment The variance function value of , the variation function value For the Temperature measuring device coordinates Hedi Temperature measuring device coordinates The semivariance of the temperature at , used to quantitatively describe the spatial distance Variability, variogram value The larger the distance, the The more drastic the temperature change under the scale, the greater the value of the variation function. The specific formula is as follows:

[0071] ;

[0072] in, and Temperature measuring equipment Middle Temperature measuring device coordinates Hedi Temperature measuring device coordinates The temperature at the location is obtained by fitting a polynomial using the least squares method based on the existing spatial distance and variogram value. ;

[0073] The coordinates to be estimated Establish the Kriging equation system and solve the optimal weight coefficient vector based on the unbiased and minimum variance estimation principles , the optimal weight coefficient vector Satisfy the coordinates of the point to be estimated With the Temperature measuring device coordinates The variance function value is equal to the optimal weight coefficient vector With the Temperature measuring device coordinates The sequence of variance function values The inner product of Including Temperature measuring device coordinates With temperature set middle The variation function value of the temperature measuring device coordinates, that is, , and the optimal weight coefficient vector The sum of the sequence is 1, so the coordinates of the point to be estimated The Kriging equations include The equations are sorted into matrix form as follows:

[0074] ;

[0075] where are the coordinates of the point to be estimated and are the variogram values corresponding to the spatial distances between the coordinates of thermometric devices, are the optimal weights from the 1st optimal weight to the

[0076] The optimal weight coefficient vector for the coordinates of the point to be estimated is obtained by solving. Then, the estimated temperature of the coordinates of the point to be estimated is the inner product of the known temperature vector formed by arranging the measured temperatures of thermometric devices in the temperature set and the optimal weight coefficient vector ;

[0077] Since the thermometric devices are installed on the surface of the functional device, points to be estimated are evenly selected on the surface of the three-dimensional model of the functional device, and the corresponding estimated temperatures are estimated. The estimated temperatures are combined with the temperature set to generate an extended temperature set . The extended temperature set includes the temperatures of points on the surface of the functional device, being the total number of points.

[0078] Specifically, the grid segmentation technology first simplifies the three-dimensional model of the functional device, removes the detailed structures in the three-dimensional model that have little impact on the temperature field modeling, generates a simplified three-dimensional model, translates each axis of the Cartesian space coordinate system along the directions of the other two axes respectively to generate a set of parallel lines. The lines in the set of parallel lines of the three axes intersect with each other, and the simplified three-dimensional model is cut. The intersection point of any two lines is a node, and nodes are discretely generated. A hollow hexahedron surrounded by 8 nodes is defined as a unit. Then, nodes can construct units. Among them, a hollow hexahedron means that there are no nodes inside the hexahedron. and are the total number of nodes and the total number of units respectively, and the total number of nodes is much larger than the total number of points The distance between two adjacent parallel lines in a parallel line group is defined based on the spatial distance between each position and the heat source in the simplified three-dimensional model. For example, the distance between adjacent parallel lines close to the burner in the heating section, holding section and slow cooling section of a continuous annealing furnace is defined as the distance between adjacent parallel lines away from the burner. , for the expanded temperature set The temperature of each point in the surface is obtained by performing a nearest neighbor search based on the corresponding point coordinates, retrieving the node with the smallest spatial distance from the point coordinates and assigning the temperature of the corresponding point. All nodes that have been assigned temperatures are defined as known surface nodes.

[0079] Furthermore, based on the thermal conductivity of the functional device Set up the heat conduction equation and define the convection boundary , construct the finite element equation by finite element method and solve it by conjugate gradient method to obtain the temperature field model , including the following specific steps:

[0080] Determine thermal conductivity based on the material of the functional device , set up the heat conduction equation to describe the heat flux density With temperature gradient The ratio is equal to the thermal conductivity The opposite number of ;

[0081] Defining Convection Boundaries , considering the convection heat transfer between the surface of the functional device and the ambient air, the heat flux density of the convection heat transfer is described by Newton's cooling law and surface temperature and ambient temperature The relationship between the convection boundary On, satisfied , is the convective heat transfer coefficient, which reflects the convective heat transfer intensity between the surface and the ambient air. Includes temperatures of all known surface nodes;

[0082] consider Nodes built The heat balance inside each unit is The heat conduction equation within each unit is analyzed. Based on Fourier's law, heat is transferred from high temperature to low temperature, and the transfer rate is related to the temperature gradient. Proportional to the first The finite element equations of each unit are as follows:

[0083] ;

[0084] in, Denote the volume variable, and are the unit volume and unit boundary of the th unit respectively, and are the heat flux density and area of the unit boundary respectively, , and are the second-order temperature gradients in the th unit in the axis direction, axis direction and axis direction of the Cartesian coordinate system respectively. The second-order temperature gradient reflects the change of the temperature gradient , ; ;

[0085] Use the interpolation representation of the shape functions of each node in the th unit to describe the temperature representation in the th unit with spatial coordinates , , and are the coordinate variables in the axis direction, axis direction and axis direction respectively, as follows:

[0086] ;

[0087] where, and are the shape function and nodal temperature of the th node in the th unit respectively. The shape function , , and are the nodal coordinates of the th node in the th unit respectively, ;

[0088] Substitute the temperature representation of the th unit into the finite element equation of the th unit. After integral operation and matrix transformation derivation, it is sorted into a finite element matrix equation in matrix form , where,[[ID=9 + 6]] , and are the The stiffness matrix, nodal temperature vector, and load vector of a unit. The stiffness matrix and the load vector are used to describe the influence between the eight nodal temperatures within the th unit and the influence of the boundary heat flux on the eight nodal temperatures within the th unit;

[0089] Update the finite element matrix equation of the units according to the assembly rules. The assembly rules are established based on the nodal sharing relationship of adjacent units. When the th unit and the th unit are adjacent, there are four shared nodes between the th unit and the th unit. For a single shared node, extract the stiffness matrix of the th unit and the stiffness matrix of the th unit and accumulate and replace the influence vectors related to the single shared node. Extract the load vector of the th unit and the load vector of the th unit and accumulate and replace the influence related to the single shared node. After the finite element matrix equation of the units is updated according to the assembly rules, it is integrated into a system of finite element matrix equations, and the conjugate gradient method is used to solve for the nodal temperature vector of the units. Assign the temperatures in the nodal temperature vector of each unit to the corresponding nodes to generate a temperature field model .

[0090] Specifically, since the axis of the Cartesian coordinate system is the central axis of the functional device and exactly passes through all functional segments of the functional device, statistical analysis of the temperature field model in the axis direction can effectively reflect the temperature change trend of different functional segments. For the temperature field model in the axis direction with the same coordinate variable , take the average value of the temperatures to generate the average temperature corresponding to the coordinate variable . Traverse all the coordinate variables in the temperature field model in turn to obtain a series of average temperatures, and use the least squares fitting polynomial to obtain the temperature trend curve .

[0091] Furthermore, compare the temperature trend curve With the standard temperature trend curve Obtain the abnormal segment number, including the following specific steps:

[0092] Determine according to the three-dimensional model Functional segments in the Axis direction segment coordinate variable range, where the Functional segment's segment coordinate variable range is , And Are respectively the starting coordinate variable and the ending coordinate variable of the Functional segment;

[0093] Calculate the temperature trend curve And the standard temperature trend curve In the Functional segment's temperature mean square error , The Functional segment's temperature mean square error Is the definite integral of the square of the error between the temperature trend curve And the standard temperature trend curve Divided by the segment coordinate variable length In the axis direction segment coordinate variable range And then divided by the segment coordinate variable length , The specific formula is as follows:

[0094] ;

[0095] Among them, Represents Axis direction coordinate variable. If the Functional segment's temperature mean square error Is greater than or equal to the temperature mean square error threshold, then it is determined that the Functional segment is abnormal, Mark it as the abnormal segment number. If the Functional segment's temperature mean square error Is less than the temperature mean square error threshold, then it is determined that the Functional segment is normal.

[0096] As Figure 3 Shown, further, compare the gas concentration set With the gas standard concentration set To obtain the abnormal segment number, including the following specific steps:

[0097] Obtain the gas concentration set And the gas standard concentration set In the Functional segment's gas concentration vector And the gas standard concentration vector ;

[0098] Calculate the gas concentration vector and the gas standard concentration vector of the mean square error of concentration , which is specifically as follows:

[0099] ; <()

[0100] Among them, is the concentration error vector of the th functional segment;

[0101] If the mean square error of concentration of the th functional segment is greater than or equal to the mean square error threshold of concentration, it is determined that the th functional segment is abnormal, marked as the abnormal segment number. If the mean square error of concentration of the th functional segment is less than the mean square error threshold of concentration, it is determined that the th functional segment is normal. <()()

[0102] Specifically, assuming that the abnormal segment number is , the segment temperature difference of the th functional segment is the definite integral of the error between the temperature trend curve and the standard temperature trend curve within the range of the segment coordinate variable in the axis direction divided by the segment coordinate variable length , and the specific formula is as follows:

[0103] ;

[0104] The concentration error vector of the th functional segment, , , and are respectively the oxygen concentration error, carbon dioxide concentration error, carbon monoxide concentration error and nitrogen concentration error of the th functional segment.

[0105] Specifically, the fuzzy PID control algorithm includes fuzzy inference. The input variables of the fuzzy inference include the temperature difference between segments of the abnormal function segment, the oxygen concentration error, the carbon dioxide concentration error, the carbon monoxide concentration error, and the nitrogen concentration error. The output variables of the fuzzy inference include the proportional factor, the integral factor, and the derivative factor of the PID control algorithm. For each input variable or output variable, the corresponding universe of discourse is determined, and the universe of discourse is the value range corresponding to the input variable or output variable. The defined fuzzy sets include "negative large", "negative small", "zero", "positive medium", and "positive large". Fuzzy rules are formulated based on experience, and the fuzzy rules establish the corresponding relationship between the input fuzzy set group and the output fuzzy set group. Among them, the input fuzzy set group is composed of the fuzzy sets to which the 5 input variables belong. Since there are 5 types of input variables in total, each input variable can belong to 5 different fuzzy sets, and the input fuzzy set group has a total of ones. The output fuzzy set group is composed of the fuzzy sets to which the 3 output variables belong, and the output fuzzy set group has a total of ones. For each fuzzy rule, the membership degrees of the 5 input variables belonging to the corresponding fuzzy sets in the input fuzzy set group are calculated respectively through the Gaussian membership function, and the activation strength of the fuzzy rule is obtained through the AND operation. The activation strengths of all fuzzy rules are subjected to the OR operation to determine the final output fuzzy set group and the membership degrees of the 3 output variables therein. The centroid method is used to perform defuzzification in combination with the universes of discourse of the 3 output variables, and the proportional factor, the integral factor, and the derivative factor of the PID control algorithm are output and further assigned to the PID control algorithm.

[0106] Specifically, the fuzzy PID control algorithm includes the PID control algorithm. The temperature difference between segments of the abnormal function segment, the oxygen concentration error, the carbon dioxide concentration error, the carbon monoxide concentration error, and the nitrogen concentration error are respectively multiplied by the corresponding weight factors and summed to generate a comprehensive error. The proportional factor, the integral factor, and the derivative factor determined by fuzzy inference are obtained. The proportional adjustment uses the product of the comprehensive error and the proportional factor as the proportional control quantity. The proportional adjustment is used to reduce the comprehensive error at one time from a macroscopic level to generate a remaining comprehensive error. The integral adjustment performs an integral operation on the remaining comprehensive error and multiplies it by the integral factor to generate an integral control quantity and updates the remaining comprehensive error. The integral adjustment is repeated multiple times until the remaining comprehensive error is less than the comprehensive error threshold and then stops. The multiple integral adjustments are used to continuously reduce the remaining comprehensive error from a microscopic perspective. The derivative adjustment multiplies the change rate of the comprehensive error between two adjacent integral adjustments by the derivative factor to generate a derivative correction control quantity to avoid overshoot phenomena in the PID control algorithm. The number of derivative adjustments is the same as the number of integral adjustments. The control quantity of the PID control algorithm is equal to the sum of the proportional control quantity, the multiple integral control quantities, and the multiple derivative control quantities. Considering the fuel and air stoichiometric ratio, the control quantity is split into the fuel valve adjustment quantity and the air valve adjustment quantity, and combined into the adjustment vector corresponding to the abnormal function segment.

[0107] Example 2

[0108] As Figure 4 shown, the present invention also discloses a decision-making control method based on temperature field and atmosphere analysis, and the implementation of a decision-making control system based on temperature field and atmosphere analysis, including the following specific steps:

[0109] At each monitoring moment, a temperature set is monitored and obtained through a temperature measuring device group and a gas detection device arranged in segments and a gas concentration set ;

[0110] Based on the geometric parameters of the functional segments, a three-dimensional model of the functional device is constructed and a Cartesian space coordinate system is established, where is the total number of functional segments;

[0111] The variogram is used to describe the spatial variation characteristics of temperature in the functional device, and based on the Kriging equation, the measured temperatures in the temperature set are spatially extended on the surface of the three-dimensional model to generate an extended temperature set ;

[0112] The grid cutting technology is used to simplify and discretize the three-dimensional model into nodes, and the temperature of each point in the extended temperature set is assigned to the surface nodes matching the corresponding point coordinates through the nearest neighbor search, where is the total number of nodes;

[0113] The ambient temperature is obtained, and based on the thermal conductivity of the functional device, a heat conduction equation is set and a convection boundary is defined, and a finite element matrix equation set is constructed by the finite element method and solved by the conjugate gradient method to obtain a temperature field model ;

[0114] The temperature field model is statistically analyzed in the axis direction to generate a temperature trend curve and compared with the standard temperature trend curve , and combined with the comparison result of the gas concentration set and the gas standard concentration set to determine the abnormal segment number;

[0115] The segment temperature difference and concentration error vector of each abnormal segment number are used as the input of the fuzzy PID control algorithm, the parameters of the fuzzy PID control algorithm are dynamically adjusted by using fuzzy logic, and an adjustment vector corresponding to each abnormal segment number is output and the fuel valve opening and air valve opening of the corresponding functional segment are automatically adjusted.

[0116] The present invention discloses a decision control system and method based on temperature field and atmosphere analysis. The decision control system includes a monitoring module, a trend analysis module, a decision module and an execution module; the monitoring module obtains a temperature set and a gas concentration set; the trend analysis module constructs a three-dimensional model and establishes a coordinate system, combines a variogram with a Kriging equation to supplement the temperature set to generate an extended temperature set, discretizes the three-dimensional model into multiple nodes and matches the points in the extended temperature set through nearest neighbor search, sets a heat conduction equation and a convection boundary and constructs a finite element matrix equation set, solves to obtain a temperature field model and statistically generates a temperature trend curve; the decision module compares the temperature trend curve, the gas concentration set and the process standard to obtain an abnormal segment number, and uses a fuzzy PID control algorithm to analyze the segment temperature difference and the concentration error vector of the abnormal segment number to generate an adjustment vector; the execution module automatically adjusts based on the adjustment vector, realizing intelligent decision adjustment based on the joint analysis of the temperature field and the atmosphere.

[0117] The above is only the preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions falling within the concept of the present invention belong to the protection scope of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A decision-making control system based on temperature field and atmosphere analysis, characterized in that: Includes trend analysis module and decision-making module; The trend analysis module constructs a three-dimensional model and establishes a coordinate system, obtains a temperature set, uses a variation function to describe the spatial variation of temperature, adds temperatures of different points to the temperature set through the Kriging equation to generate an expanded temperature set, and uses a grid segmentation technique to simplify and discretize the three-dimensional model into Nodes, assign the temperature of each point in the expanded temperature set to the matching node through nearest neighbor search, set the heat conduction equation based on thermal conductivity and define the convection boundary in combination with the ambient temperature, use the finite element method to construct the finite element matrix equation group, solve the generated temperature field model through the conjugate gradient method and obtain the temperature trend curve by fitting. The temperature field model includes The temperature of the node, is the total number of nodes; The decision module obtains the temperature trend curve and gas concentration set and compares them with the process standards to determine the abnormal segment number, inputs the segment temperature difference and concentration error vector of each abnormal segment number into the fuzzy PID control algorithm, uses fuzzy logic to adjust the PID control algorithm parameters, and outputs the adjustment vector corresponding to each abnormal segment number, wherein the adjustment vector includes the fuel valve adjustment amount and the air valve adjustment amount.

2. The decision-making control system based on temperature field and atmosphere analysis according to claim 1, characterized in that: The step of adding the temperatures of different points to the temperature set by the Kriging equation to generate the expanded temperature set includes: Uniformly select multiple points to be estimated on the surface of the three-dimensional model; For each point to be estimated, a Kriging equation system is established based on the unbiasedness and minimum variance estimation principles to obtain an optimal weight coefficient vector, wherein the inner product of the optimal weight coefficient vector and the sequence of variance function values of the coordinates of a single temperature measuring device in the temperature set is equal to the variance function value of the coordinates of the point to be estimated and the coordinates of the single temperature measuring device, and the sum of the optimal weight coefficients in the optimal weight coefficient vector is 1; Solve to obtain the optimal weight coefficient vector of the point to be estimated. The estimated temperature of the point to be estimated is the temperature concentration The inner product of the known temperature vector arranged by the measured temperatures and the optimal weight coefficient vector; Add the temperatures of all points to be estimated into the temperature set to generate an expanded temperature set.

3. The decision-making control system based on temperature field and atmosphere analysis according to claim 1, characterized in that: The use of a variation function to describe the spatial variation of temperature includes: From the temperature set Randomly select two temperature measuring device coordinates from the temperature measuring device coordinates to form a temperature measuring device pair, and calculate the spatial distance. is the total number of temperature measuring devices; The variogram value of the temperature measuring device pair is defined as the temperature semi-variance of the coordinates of the two temperature measuring devices, which is used to describe the temperature variability of spatial distance. The variogram is obtained by fitting the spatial distance and variogram value of all temperature measuring device pairs in the temperature set using the least squares method.

4. The decision-making control system based on temperature field and atmosphere analysis according to claim 1, characterized in that: Setting the heat conduction equation based on thermal conductivity and defining the convection boundary in combination with the ambient temperature includes: Determine thermal conductivity and set up the heat conduction equation to describe that the ratio of heat flux density to temperature gradient is equal to the inverse of thermal conductivity; The convection boundary is defined based on the convection heat transfer between the surface of the functional equipment and the ambient air. The heat flux density of the convection heat transfer is described by Newton's law of cooling, which is equal to the difference between the surface temperature and the ambient temperature multiplied by the convection heat transfer coefficient.

5. The decision-making control system based on temperature field and atmosphere analysis according to claim 1, characterized in that: The method of constructing a finite element matrix equation group by using the finite element method and solving the temperature field model by the conjugate gradient method includes: The heat conduction equation of a single unit is analyzed based on Fourier's law, and the weighted residual method is used to construct the finite element equation of a single unit. The finite element equation describes that the integral of the second-order temperature gradient in a single unit over the unit volume is proportional to the integral of the heat flux density at the unit boundary over the area of the unit boundary, and the ratio is equal to the thermal conductivity; The temperature representation of any spatial coordinate in a single unit is described by interpolating the shape functions of all nodes in the unit. The shape functions are determined based on the geometric characteristics and node distribution of the unit. Substituting the temperature representation of a single unit into the finite element equation of the single unit, and organizing it into a finite element matrix equation through integration operation and matrix transformation, the finite element matrix equation describes that the product of the stiffness matrix of the single unit and the node temperature vector is equal to the load vector, the stiffness matrix is used to describe the mutual influence between the temperatures of the nodes in the single unit, the node temperature vector records the temperature of all nodes in the single unit, and the load vector is used to describe the influence of the boundary heat flow on each node in the single unit; The finite element matrix equations of all units are updated and integrated into a finite element matrix equation group according to the assembly rules. The conjugate gradient method is used to obtain the node temperature vectors of all units and establish a temperature field model. The assembly rules are established based on the node sharing relationship of adjacent units. The values of the shared nodes in the stiffness matrices and load vectors of adjacent units are accumulated and replaced respectively.

6. The decision-making control system based on temperature field and atmosphere analysis according to claim 1, characterized in that: The acquisition of the temperature trend curve and the gas concentration set and comparison with the process standards to determine the abnormal section number includes: Determine each functional segment based on the three-dimensional model The segment coordinate variable range in the axial direction is used to obtain the process standard, wherein the process standard includes a standard temperature trend curve and a gas standard concentration set; Calculate the temperature mean square error between the temperature trend curve and the standard temperature trend curve in a single functional segment. The temperature mean square error of a single functional segment is the definite integral of the square of the error between the temperature trend curve and the standard temperature trend curve within the corresponding segment coordinate variable range divided by the segment coordinate variable length. If the temperature mean square error is greater than or equal to the temperature mean square error threshold, the segment number of the single functional segment is marked as an abnormal segment number. If the temperature mean square error is less than the temperature mean square error threshold, no processing is performed. Extracting the gas concentration vector and gas standard concentration vector of a single functional segment in the gas concentration set and the gas standard concentration set; Calculate the concentration mean square error of a single functional segment. The concentration mean square error is the square of the two-norm of the concentration error vector of a single functional segment. , the concentration error vector is equal to the absolute value of the gas concentration vector of a single functional segment minus the gas standard concentration vector; If the concentration mean square error is greater than or equal to the concentration mean square error threshold, the segment number of the single functional segment is marked as an abnormal segment number. If the concentration mean square error is less than the concentration mean square error threshold, no processing is performed.

7. The decision-making control system based on temperature field and atmosphere analysis according to claim 1, characterized in that: The grid segmentation technology uses multiple straight lines parallel to the three coordinate axes to segment the three-dimensional model, defines the intersection of the straight lines as nodes, and defines the empty hexahedron surrounded by 8 nodes as a unit. The empty hexahedron refers to a hexahedron without any nodes inside. A nearest neighbor search is performed based on the coordinates of each point in the expanded temperature set to determine the node with the smallest spatial distance to the point coordinate and assign the temperature corresponding to the point coordinate.

8. The decision-making control system based on temperature field and atmosphere analysis according to claim 1, characterized in that: The fuzzy PID control algorithm includes fuzzy reasoning, and the input variables of the fuzzy reasoning include the temperature difference of the abnormal functional segment and the oxygen concentration error, carbon dioxide concentration error, carbon monoxide concentration error, and nitrogen concentration error in the concentration error vector. The output variables of the fuzzy reasoning include the proportional factor, integral factor, and differential factor of the PID control algorithm. The domain of each input variable or output variable is determined, the fuzzy set is defined, and fuzzy rules are formulated. The fuzzy rules establish the corresponding relationship between the input fuzzy set group and the output fuzzy set group. For a single fuzzy rule, the membership of each input variable to the corresponding fuzzy set in the input fuzzy set group is calculated by a Gaussian membership function, and the activation strength of the single fuzzy rule is obtained by an AND operation. The activation strengths of all fuzzy rules are ANDed to determine the membership of the output fuzzy set group and each output variable. The center of gravity method is used for defuzzification, and the proportional factor, integral factor, and differential factor are output.

9. The decision-making control system based on temperature field and atmosphere analysis according to claim 1, characterized in that: The decision control system further includes a monitoring module and an execution module; The monitoring module periodically sets monitoring moments and collects temperature sets and gas concentration sets at each monitoring moment, wherein the temperature sets and gas concentration sets respectively include The temperature sequence of each functional segment and The gas concentration vector of each functional segment, the temperature sequence of a single functional segment includes the measured temperatures of all temperature measuring devices in the single functional segment, is the total number of functional segments; The execution module automatically adjusts based on the adjustment vector of each abnormal segment number.

10. The decision-making control method based on temperature field and atmosphere analysis is characterized in that: The specific steps include: Obtain temperature sets and gas concentration sets at each monitoring moment; Construct a three-dimensional model based on the geometric parameters of each functional segment and establish a coordinate system; The spatial variation of temperature is described by using the variation function, and the temperature in the temperature set is spatially expanded based on the Kriging equation to generate an expanded temperature set; The structured grid technology is used to simplify and discretize the three-dimensional model into nodes, assign the temperature of each point in the expanded temperature set to the matching node through nearest neighbor search. is the total number of nodes; The heat conduction equation is set based on thermal conductivity and the convection boundary is defined by combining the obtained ambient temperature. The finite element matrix equation group is constructed by the finite element method and the conjugate gradient method is used to solve it to obtain the temperature field model. The temperature field model Statistical analysis is performed in the axial direction to fit and generate a temperature trend curve. The temperature trend curve and gas concentration set are compared with the corresponding process standards to determine the abnormal section number; The segment temperature difference and concentration error vector of each abnormal segment number are input into the fuzzy PID control algorithm, and the PID control algorithm parameters are adjusted using fuzzy logic. The adjustment vector corresponding to each abnormal segment number is output and automatically adjusted.

Images