Continuous rolling process comprehensive monitoring method considering rolling characteristics

Abstract

The invention discloses a continuous rolling process comprehensive monitoring method considering rolling characteristics. The continuous rolling process comprehensive monitoring method comprises the steps that rolling schedule data, equipment data and rolling process parameters of a rolling site are collected; according to the rolling theoretical equation, the field rolling schedule data and the equipment data, a rolling process state space equation is constructed to simulate the rolling process; calculating the Hurst index of each rolling process parameter; calculating an LQG index of each rolling process parameter; the CUSUM offset index of each rolling process parameter is calculated; calculating a comprehensive evaluation index of each rolling process parameter; calculating the importance coefficient of each rolling process parameter; and comprehensive evaluation indexes of the continuous rolling process are calculated, and the continuous rolling process is evaluated. According to the continuous rolling process comprehensive monitoring method considering the rolling characteristics, comprehensive evaluation of the continuous rolling process is achieved, rolling characteristic analysis is introduced, the fidelity of evaluation indexes of the continuous rolling process is greatly improved, and the continuous rolling process comprehensive monitoring method can be widely applied to rolling production.

Description

Technical Field

[0001] The invention belongs to the technical field of rolling control and relates to a comprehensive monitoring method for a continuous rolling process taking rolling characteristics into consideration. Background Art

[0002] As the demands for steel products continue to rise across various industries, higher requirements are being placed on the control level of the continuous rolling process. The continuous rolling process involves numerous process parameters and is accompanied by various unknown interferences. Due to a lack of continuous rolling process monitoring and system maintenance, the control accuracy of the rolling process has declined, failing to meet the control requirements set by the target. Implementing comprehensive rolling process monitoring to promptly identify and control the causes of poor performance is crucial to ensuring rolling process control accuracy.

[0003] At present, with the development of computer technology and informatization, rolling process monitoring and control performance evaluation have been applied. Chinese patent "CN117806263A" proposes a rolling process thickness control performance evaluation method based on multi-factor covariance ratio, which monitors multiple rolling process parameters simultaneously during the rolling process to achieve real-time evaluation of each pass and overall performance in the rolling process; Chinese patent "CN117983668B" proposes a hot rolling process thickness looper tension optimization control method based on performance evaluation, which uses the Hurst index to evaluate the performance of the thickness-looper-tension control system in real time, and uses the crayfish optimization algorithm to optimize the control parameters of the control system; Chinese patent "CN116433085A" proposes a rolling process control system performance evaluation method, which defines a Hurst index based on strip production data. Performance index, which is used to evaluate the performance of the current rolling process control system; Chinese patent "CN117519067B" proposes a method for evaluating the control performance of multiple stands in a continuous rolling process. The detrended fluctuation analysis algorithm is used to solve the corresponding s value of the multi-variable time series of multiple stands in the continuous rolling process, and then the Hurst index of the control system is obtained. The performance status of the controller is evaluated using the multi-stand control performance index; Chinese patent "CN116274421A" proposes a remote monitoring system and monitoring method for key process actions in the rolling process based on sound sensors. By collecting audio signals from important on-site parts during the rolling process, the system solves the problem of operators not being able to hear on-site sounds after remote centralized control, and realizes monitoring and analysis of on-site conditions.

[0004] While the aforementioned methods for evaluating rolling process monitoring and control performance have achieved some success, they still have limitations. Firstly, the continuous rolling process involves numerous process parameters and complex operating conditions, and these studies, employing only a single evaluation metric, make it difficult to comprehensively assess the rolling process. Secondly, these studies either fail to consider the evaluation of the continuous rolling process or construct evaluation metrics solely through a linear weighting approach, failing to consider the actual rolling characteristics of the process and thus failing to effectively evaluate the process. Summary of the Invention

[0005] In order to solve the above technical problems, the purpose of the present invention is to provide a comprehensive monitoring method for the continuous rolling process taking into account the rolling characteristics. By introducing rolling characteristics and a comprehensive evaluation system, a comprehensive evaluation of the continuous rolling process is achieved, which makes up for the lack of evaluation benchmark indicators for continuous rolling process monitoring in the existing evaluation system.

[0006] The present invention provides a comprehensive monitoring method for a continuous rolling process taking rolling characteristics into consideration, comprising:

[0007] Step 1: Collect rolling procedure data, equipment data and rolling process parameters at the rolling site;

[0008] Step 2: Based on the rolling theory equation, on-site rolling procedure data and equipment data, the rolling process state space equation is constructed to simulate the rolling process;

[0009] Step 3: Calculate the Hurst index of each rolling process parameter;

[0010] Step 4: Calculate the LQG index of each rolling process parameter;

[0011] Step 5: Calculate the CUSUM deviation index of each rolling process parameter;

[0012] Step 6: Calculate the comprehensive evaluation index of each rolling process parameter;

[0013] Step 7: Calculate the importance coefficient of each rolling process parameter;

[0014] Step 8: Calculate the comprehensive evaluation index of the continuous rolling process and evaluate the continuous rolling process according to the continuous rolling process health evaluation standard.

[0015] Furthermore, the step 2 is specifically as follows:

[0016] Step 2.1: Establish the state space equation of the rolling process:

[0017]

[0018] Where A, B, C, D, K are the state space matrices of the rolling process, x is the state vector of the rolling process, is the first-order derivative of x, u is the rolling process control vector, e0 is the noise vector, y is the output vector, and t is the time;

[0019] Step 2.2: Based on rolling theory, establish the rolling force increment equation for each stand, the forward slip increment equation for each stand, the thickness increment equation for each stand outlet, and the tension increment equation for each stand outlet;

[0020] Step 2.3: Calculate the partial differential coefficients of each incremental equation based on the rolling procedure data and equipment data;

[0021] Step 2.4: According to the partial differential coefficients of each incremental equation, calculate the state space matrix A, B, C, D, K of the rolling process, and substitute A, B, C, D, K into the state space equation of step 2.1.

[0022] Furthermore, the step 3 is specifically as follows:

[0023] Step 3.1: Determine the autocorrelation sequence of each rolling process parameter:

[0024] Step 3.2: Divide the autocorrelation sequence into M windows of window length l, and calculate the root mean square fluctuation f(l) of the autocorrelation sequence when the window length is l:

[0025]

[0026] Where Y m (i) is the autocorrelation sequence of the mth window, is the least squares fitting curve in the mth window, a m and b m are the slope and intercept of the least squares fitting curve, respectively, and i is the i-th data point in the m-th window;

[0027] Step 3.3: Calculate the logarithm of multiple groups of l, log l, and the logarithm of the RMS fluctuation of l, log f(l);

[0028] Step 3.4: Fit the slopes ρ of multiple sets of log l and log f(l) using the first-order least squares method;

[0029] Step 3.5: Calculate the Hurst exponent of each rolling process parameter:

[0030]

[0031] Among them, F Hurst The Hurst index represents the rolling process parameters.

[0032] Furthermore, the step 4 is specifically as follows:

[0033] Step 4.1: Define the objective function of LQG:

[0034]

[0035] Among them, u t is the control vector at time t, y t is the output vector at time t, E is the expectation, and λ is the weighting coefficient for controlling the input variance;

[0036] Step 4.2: Based on the rolling process state space equation, define the output matrix of rolling process parameters:

[0037]

[0038] Rewrite the output matrix into subspace matrix equation form:

[0039]

[0040] Among them, y 0|N-1 is the output matrix of rolling process parameters, u 0|N-1 is the input matrix of rolling process parameters, is the input partial differential coefficient matrix, is the partial differential coefficient matrix of noise;

[0041] Substituting the output matrix of rolling process parameters into the LQG objective function, we get:

[0042]

[0043] Step 4.3: Obtain the rolling process parameter output matrix under the optimal control rate. The steps are as follows:

[0044] Make the objective function J to u 0|N-1 Find the partial derivative:

[0045]

[0046] The control rate of minimizing the objective function is obtained as follows:

[0047]

[0048] Then the rolling process parameter output matrix under the optimal control rate is obtained:

[0049]

[0050] Where I is the identity matrix, λ is the weighting coefficient for controlling the input variance;

[0051] Step 4.4: Define the coefficient matrix γ:

[0052]

[0053] Step 4.5: Find the optimal output sequence:

[0054]

[0055] Among them, e t-j is the noise at time tj;

[0056] Step 4.6: Find the covariance of the optimal output sequence:

[0057]

[0058] Among them, Cov[e t-j ] is the noise e t-j The covariance of γ j The transpose of , N is the number of sampling points;

[0059] Step 4.7: Obtain the LQG performance benchmark curve. The steps are as follows:

[0060] Calculate LQG performance benchmark:

[0061]

[0062] By changing the value of λ from 0 to ∞, the optimal input-output variance of the rolling process under the LQG benchmark is obtained, and then the LQG benchmark performance curve of the rolling process is obtained;

[0063] Step 4.8: Obtain the LQG index F using the LQG performance benchmark curve LQG :

[0064]

[0065] Among them, V y is the actual variance of the rolling process parameters at the current moment t, is the expected variance of the LQG performance benchmark corresponding to the rolling process parameters at the current time t.

[0066] Furthermore, the step 5 is specifically as follows:

[0067] Step 5.1: Calculate the standard deviation σ of the rolling process parameters;

[0068] Step 5.2: Calculate the mean of rolling process parameters

[0069] Step 5.3: Calculate the CUSUM offset index F CUSUM :

[0070]

[0071] Among them, y n is the actual value of the nth rolling process parameter, is the average value of the rolling process parameters.

[0072] Furthermore, in step 6, the comprehensive evaluation index of each rolling process parameter is calculated according to the following formula:

[0073]

[0074] Among them, F k Comprehensive evaluation index of the kth rolling process parameter, is the Hurst index of the kth rolling process parameter, is the LQG index of the kth rolling process parameter, is the CUSUM offset index of the kth rolling process parameter.

[0075] Furthermore, the step 7 is specifically as follows:

[0076] Step 7.1: Based on the constructed rolling process state space equation, apply 5% noise to each control variable in the rolling process control vector in turn, and calculate the change ΔW of the kth rolling process parameter caused by the noise of the pth control variable pk ;

[0077] Step 7.2: Calculate the relative change of the kth rolling process parameter

[0078]

[0079] in, is the set value of the kth rolling process parameter;

[0080] Step 7.3: Calculate the importance coefficient of the kth rolling process parameter

[0081]

[0082] Where P is the number of control mechanisms.

[0083] Furthermore, in step 8, the comprehensive evaluation index of the continuous rolling process is calculated according to the following formula:

[0084] Step 8: Construct comprehensive evaluation indicators for the continuous rolling process and evaluate the continuous rolling process, specifically:

[0085] Step 8.1: Calculate the comprehensive evaluation index of the continuous rolling process:

[0086]

[0087] Where K is the number of rolling process parameters.

[0088] Step 8.2: Establish the evaluation criteria for the health of the continuous rolling process based on the comprehensive evaluation indicators:

[0089] When the comprehensive evaluation index is in the range of [0.9, 1], the continuous rolling process is in very good condition; when the comprehensive evaluation index is in the range of [0.7, 0.9), the continuous rolling process is degraded, but no maintenance is required; when the comprehensive evaluation index is in the range of [0, 0.7), the continuous rolling process fails and maintenance measures must be taken to maintain the continuous rolling process.

[0090] The present invention provides a comprehensive monitoring method for a continuous rolling process that takes rolling characteristics into consideration. First, the Hurst index, LQG index, and CUSUM offset index of each rolling process parameter of the continuous rolling process are calculated respectively, and then the three indexes are averaged to calculate the comprehensive evaluation index of each process parameter of the continuous rolling process. Then, the importance coefficient of each process parameter of the continuous rolling process is calculated through rolling characteristic analysis. Finally, the comprehensive evaluation index and importance coefficient of each process parameter are weighted and summed to obtain the comprehensive evaluation index of the continuous rolling process. The comprehensive monitoring method for a continuous rolling process that takes rolling characteristics into consideration proposed by the present invention not only realizes a comprehensive and comprehensive evaluation of the continuous rolling process, but also introduces rolling characteristic analysis, greatly improving the fidelity of the continuous rolling process evaluation index, and can be widely used in rolling production. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] Figure 1 This is a flow chart of a comprehensive monitoring method for a continuous rolling process taking rolling characteristics into consideration according to the present invention;

[0092] Figure 2 The process data of thickness at the exit of each stand and post-tension during the cold rolling process collected during the implementation of the present invention;

[0093] Figure 3 Schematic diagram of the LQG index calculation process of the present invention. DETAILED DESCRIPTION

[0094] The specific implementation of the present invention is further described in detail below with reference to the accompanying drawings and examples.

[0095] In order to verify the effectiveness of the comprehensive monitoring method for tandem rolling process considering rolling characteristics, a cold rolling production line was taken as an example to conduct a comprehensive evaluation of the thickness-tension control system of cold rolling.

[0096] like Figure 1 As shown, the present invention provides a comprehensive monitoring method for a continuous rolling process taking rolling characteristics into consideration, comprising:

[0097] Step 1: Collect rolling procedure data, equipment data and rolling process parameters at the rolling site.

[0098] The on-site rolling procedure data and equipment data include: work roll radius, rolling mill stiffness, friction coefficient, stand spacing, entry thickness setting value, exit thickness setting value, rolling force setting value, front tension setting value, back tension setting value, initial roll gap setting value, roll speed setting value, strip entry speed setting value, and inertia time constant setting value, as shown in Table 1.

[0099] The rolling process parameters include the actual value of the post-tension and the actual value of the exit thickness. Figure 2 shown.

[0100] Table 1 Rolling procedure data and equipment data

[0101]

[0102] Step 2: Based on the rolling theory equation, on-site rolling procedure data and equipment data, the rolling process state space equation is constructed to simulate the rolling process. Specifically:

[0103] Step 2.1: Establish the state space equation of the rolling process:

[0104]

[0105] Where A, B, C, D, K are the state space matrices of the rolling process, x is the state vector of the rolling process, is the first-order derivative of x, u is the rolling process control vector, e0 is the noise vector, y is the output vector, and t is the time.

[0106] Step 2.2: Based on rolling theory, establish the rolling force increment equation for each stand, the forward slip increment equation for each stand, the outlet thickness increment equation for each stand, and the forward tension increment equation for each stand.

[0107] The rolling force increment equation is:

[0108] ΔF=α1Δh in +α2Δh out +α3Δσ in +α4Δσ out

[0109] Where, is the partial differential coefficient of rolling force to inlet thickness, F is rolling force, h in is the inlet thickness, Δh in is the inlet thickness increment, is the partial differential coefficient of rolling force to outlet thickness, h out is the outlet thickness, Δh out is the outlet thickness increment, is the partial differential coefficient of rolling force to back tension, σin is the post-tension, Δσ in is the post-tension increment, is the partial differential coefficient of rolling force to front tension, σ out is the front tension, Δσ out is the front tension increment.

[0110] Forward sliding increment equation:

[0111] Δf=γ1Δh in +γ2Δh out +γ3Δσ in +γ4Δσ out

[0112] Where, is the partial differential coefficient of the front slip with respect to the inlet thickness, f is the front slip value, is the partial differential coefficient of the forward sliding with respect to the outlet thickness, is the partial differential coefficient of the forward sliding force on the backward tension, is the partial differential coefficient of forward slip on forward tension.

[0113] Exit thickness control increment equation:

[0114]

[0115] Where, is the partial differential coefficient of the outlet thickness to the roll gap, ΔS is the roll gap increment, is the partial differential coefficient of the outlet thickness to the inlet thickness, is the partial differential coefficient of outlet thickness to post-tension, is the partial differential coefficient of outlet thickness to front tension, M m is the rolling mill stiffness.

[0116] Pre-tension control increment equation:

[0117]

[0118] b S =-Vγ2a S

[0119]

[0120] Where, E is the elastic modulus of the strip steel, L is the rack spacing, is the partial differential coefficient of the rate of change of the front tension to the rear tension, is the partial differential coefficient of the front tension change rate to the front tension of the next rack, +1 represents the next rack, is the partial differential coefficient of the front tension change rate to the front tension, is the partial differential coefficient of the front tension change rate to the next rack entrance thickness, is the partial differential coefficient of the front tension change rate with respect to the inlet thickness, b S is the partial differential coefficient of the front tension change rate to the roll gap, b is the partial differential coefficient of the front tension change rate to the next stand roll gap, V is the partial differential coefficient of the front tension change rate to the rolling speed, V is the rolling speed, f is the front slip value, b is the back slip value, is the rate of change of the front tension, and β2 is the partial differential coefficient of the strip deformation resistance to the angular velocity of the working roll.

[0121] Step 2.3: Calculate the partial differential coefficients of each incremental equation based on the rolling procedure data and equipment data.

[0122] Step 2.4: According to the partial differential coefficients of each incremental equation, calculate the state space matrix A, B, C, D, K of the rolling process, and substitute A, B, C, D, K into the state space equation of step 2.1.

[0123] In this embodiment, the state space matrices A, B, C, D, and K are expressed as follows:

[0124]

[0125] Step 3: Calculate the Hurst index of each rolling process parameter to evaluate the trend of the production process, specifically:

[0126] Step 3.1: Determine the autocorrelation sequence of each rolling process parameter.

[0127]

[0128] Where y(i) is the rolling process parameter, is the average value of the rolling process parameter, Y(j) is the autocorrelation sequence, N is the number of data points, i is the i-th data point, and j is the intermediate variable for determining the autocorrelation sequence.

[0129] Step 3.2: Divide the autocorrelation sequence into M windows of window length l, and calculate the root mean square fluctuation f(l) of the autocorrelation sequence when the window length is l:

[0130]

[0131] Where Y m (i) is the autocorrelation sequence of the mth window, is the least squares fitting curve in the mth window, a m and b m are the slope and intercept of the least squares fitting curve, respectively, and i is the i-th data point in the m-th window.

[0132] Step 3.3: Calculate the logarithm of multiple groups of l, log l, and the logarithm of the RMS fluctuation of l, log f(l).

[0133] Step 3.4: Fit the slopes ρ of multiple sets of log l and log f(l) using the first-order least squares method.

[0134] Step 3.5: Calculate the Hurst exponent of each rolling process parameter:

[0135]

[0136] Among them, F Hurst The Hurst index represents the rolling process parameters.

[0137] In this implementation, the Hurst index of the actual value of the outlet thickness and the actual value of the back tension of each rack is shown in Table 2;

[0138] Table 2 Hurst index of actual values ​​of outlet thickness and post-tension of each rack

[0139]

[0140] Step 4: Calculate the LQG index of each rolling process parameter to evaluate the volatility of the production process, specifically:

[0141] Step 4.1: Define the objective function of LQG:

[0142]

[0143] Among them, u t is the control vector at time t, y t is the output vector at time t, E is the expectation, and λ is the weighted coefficient for controlling the input variance.

[0144] Step 4.2: Based on the rolling process state space equation, define the output matrix of rolling process parameters:

[0145]

[0146] Rewrite the output matrix into subspace matrix equation form:

[0147]

[0148] Among them, y 0|N-1 is the output matrix of rolling process parameters, u 0|N-1 is the input matrix of rolling process parameters, is the input partial differential coefficient matrix, is the partial differential coefficient matrix of the noise.

[0149] Substituting the output matrix of rolling process parameters into the LQG objective function, we obtain:

[0150]

[0151] Step 4.3: Obtain the rolling process parameter output matrix under the optimal control rate. The steps are as follows:

[0152] Make the objective function J to u 0|N-1 Find the partial derivative:

[0153]

[0154] The control rate of minimizing the objective function is obtained as follows:

[0155]

[0156] Then the rolling process parameter output matrix under the optimal control rate is obtained:

[0157]

[0158] Where I is the identity matrix and λ is the weighting coefficient that controls the input variance.

[0159] Step 4.4: Define the coefficient matrix γ:

[0160]

[0161] Step 4.5: Find the optimal output sequence:

[0162]

[0163] Among them, e t-j is the noise at time tj.

[0164] Step 4.6: Find the covariance of the optimal output sequence:

[0165]

[0166] Among them, Cov[e t-j ] is the noise e t-j The covariance of γ j is the transpose of , and N is the number of sampling points.

[0167] Step 4.7: Obtain the LQG performance benchmark curve, such as Figure 3 As shown, the steps are as follows:

[0168] Calculate LQG performance benchmark:

[0169]

[0170] By changing the value of λ from 0 to ∞ in sequence, the optimal input-output variance of the rolling process under the LQG benchmark is obtained, and then the LQG benchmark performance curve of the rolling process is obtained.

[0171] Step 4.8: Obtain the LQG index F using the LQG performance benchmark curve LQG :

[0172]

[0173] Among them, V y is the actual variance of the rolling process parameters at the current moment t, is the expected variance of the LQG performance benchmark corresponding to the rolling process parameters at the current time t.

[0174] In this implementation, the LQG index of the actual values ​​of the outlet thickness and the actual values ​​of the post-tension of each rack is shown in Table 3;

[0175] Table 3 Actual strip exit thickness and LQG index of post-tension of each stand

[0176]

[0177] Step 5: Calculate the CUSUM deviation index of each rolling process parameter to evaluate the centralization of the production process, specifically:

[0178] Step 5.1: Calculate the standard deviation σ of the rolling process parameters.

[0179] Step 5.2: Calculate the mean of rolling process parameters

[0180] Step 5.3: Calculate the CUSUM offset index F CUSUM :

[0181]

[0182] Among them, y n is the actual value of the nth rolling process parameter.

[0183] In this implementation, the CUSUM deviation index of the actual value of the outlet thickness and the actual value of the back tension of each rack is shown in Table 4.

[0184] Table 4 CUSUM deviation index of actual values ​​of outlet thickness and post-tension of each stand in cold rolling mill

[0185]

[0186] Step 6: Calculate the comprehensive evaluation index of each rolling process parameter.

[0187] In specific implementation, the comprehensive evaluation index of each rolling process parameter is calculated according to the following formula:

[0188]

[0189] Among them, F k Comprehensive evaluation index of the kth rolling process parameter, is the Hurst index of the kth rolling process parameter, is the LQG index of the kth rolling process parameter, is the CUSUM offset index of the kth rolling process parameter.

[0190] In this implementation, the comprehensive evaluation index of the actual value of the outlet thickness and the actual value of the back tension of each rack is shown in Table 5.

[0191] Table 5 Comprehensive evaluation index F of actual values ​​of outlet thickness and post-tension of each stand in cold rolling mill k

[0192]

[0193] Step 7: Calculate the importance coefficient of each rolling process parameter, specifically:

[0194] Step 7.1: Based on the constructed rolling process state space equation, apply 5% noise to each control variable in the rolling process control vector in turn, and calculate the change ΔW of the kth rolling process parameter caused by the noise of the pth control variable pk .

[0195] Step 7.2: Calculate the relative change of the kth rolling process parameter

[0196]

[0197] in, is the set value of the kth rolling process parameter.

[0198] Step 7.3: Calculate the importance coefficient of the kth rolling process parameter

[0199]

[0200] Where P is the number of control mechanisms.

[0201] In this implementation, the importance coefficients of the actual values ​​of the outlet thickness and the actual values ​​of the back tension of each rack are shown in Table 6.

[0202] Table 6 Importance coefficients of actual values ​​of outlet thickness and post-tension of each stand in cold rolling mill

[0203]

[0204] Step 8: Calculate the comprehensive evaluation index of the continuous rolling process and evaluate the continuous rolling process according to the continuous rolling process health evaluation standard, specifically:

[0205] Step 8.1: Calculate the comprehensive evaluation index of the continuous rolling process:

[0206]

[0207] Where K is the number of rolling process parameters.

[0208] In this implementation, the comprehensive evaluation indicators of the continuous rolling process are:

[0209]

[0210] Step 8.2: Establish the evaluation criteria for the health of the continuous rolling process based on the comprehensive evaluation indicators:

[0211]

[0212] When the comprehensive evaluation index is in the range of [0.9, 1], the rolling process is operating very well. When it is in the range of [0.7, 0.9), the rolling process is operating degraded, but maintenance is not yet required. When it is in the range of [0, 0.7), the rolling process has experienced a failure and maintenance measures must be taken. Based on the comprehensive evaluation index of 0.854, the rolling process is operating degraded, but maintenance is not yet required.

[0213] This example uses the thickness-tension system of a tandem cold rolling mill as an example. Comprehensive evaluation indicators for thickness and tension of each stand in the tandem cold rolling mill are calculated based on the Hurst index, LQG index, and CUSUM offset index. The importance index of thickness and tension for each stand is calculated through rolling characteristics analysis and standardization. Finally, a weighted summation is performed to determine the final comprehensive score for the tandem cold rolling thickness-tension system. The proposed comprehensive tandem rolling process monitoring method, which considers rolling characteristics, addresses the shortcomings of existing methods and improves the detection accuracy of the tandem rolling process.

[0214] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A comprehensive monitoring method for continuous rolling process considering rolling characteristics, characterized in that: include: Step 1: Collect rolling procedure data, equipment data and rolling process parameters at the rolling site; Step 2: Based on the rolling theory equation, on-site rolling procedure data and equipment data, the rolling process state space equation is constructed to simulate the rolling process; Step 3: Calculate the Hurst index of each rolling process parameter; Step 4: Calculate the LQG index of each rolling process parameter; Step 5: Calculate the CUSUM deviation index of each rolling process parameter; Step 6: Calculate the comprehensive evaluation index of each rolling process parameter; Step 7: Calculate the importance coefficient of each rolling process parameter; Step 8: Calculate the comprehensive evaluation index of the continuous rolling process and evaluate the continuous rolling process according to the continuous rolling process health evaluation standard.

2. The method for comprehensive monitoring of a continuous rolling process taking rolling characteristics into consideration according to claim 1, characterized in that: The step 2 is specifically as follows: Step 2.1: Establish the state space equation of the rolling process: Where A, B, C, D, K are the state space matrices of the rolling process, x is the state vector of the rolling process, is the first-order derivative of x, u is the rolling process control vector, e0 is the noise vector, y is the output vector, and t is the time; Step 2.2: Based on rolling theory, establish the rolling force increment equation for each stand, the forward slip increment equation for each stand, the thickness increment equation for each stand outlet, and the tension increment equation for each stand outlet; Step 2.3: Calculate the partial differential coefficients of each incremental equation based on the rolling procedure data and equipment data; Step 2.4: According to the partial differential coefficients of each incremental equation, calculate the state space matrix A, B, C, D, K of the rolling process, and substitute A, B, C, D, K into the state space equation of step 2.

1.

3. The method for comprehensive monitoring of a continuous rolling process taking rolling characteristics into consideration according to claim 1, wherein: The step 3 is specifically as follows: Step 3.1: Determine the autocorrelation sequence of each rolling process parameter: Step 3.2: Divide the autocorrelation sequence into M windows of window length l, and calculate the root mean square fluctuation f(l) of the autocorrelation sequence when the window length is l: Where Y m (i) is the autocorrelation sequence of the mth window, is the least squares fitting curve in the mth window, a m and b m are the slope and intercept of the least squares fitting curve, respectively, and i is the i-th data point in the m-th window; Step 3.3: Calculate the logarithm logl of multiple groups of l and the logarithm of the RMS fluctuation log f(l); Step 3.4: Fit the slopes ρ of multiple sets of log l and logf(l) using the first-order least squares method; Step 3.5: Calculate the Hurst exponent of each rolling process parameter: Among them, F Hurst The Hurst index represents the rolling process parameters.

4. The method for comprehensive monitoring of a continuous rolling process taking rolling characteristics into consideration according to claim 2, wherein: The step 4 is specifically as follows: Step 4.1: Define the objective function of LQG: Among them, u t is the control vector at time t, y t is the output vector at time t, E is the expectation, and λ is the weighting coefficient for controlling the input variance; Step 4.2: Based on the rolling process state space equation, define the output matrix of rolling process parameters: Rewrite the output matrix into subspace matrix equation form: Among them, y 0|N-1 is the output matrix of rolling process parameters, u 0|N-1 is the input matrix of rolling process parameters, is the input partial differential coefficient matrix, is the partial differential coefficient matrix of noise; Substituting the output matrix of rolling process parameters into the LQG objective function, we obtain: Step 4.3: Obtain the rolling process parameter output matrix under the optimal control rate. The steps are as follows: Make the objective function J to u 0|N-1 Find the partial derivative: The control rate of minimizing the objective function is obtained as follows: Then the rolling process parameter output matrix under the optimal control rate is obtained: Where I is the identity matrix, λ is the weighting coefficient for controlling the input variance; Step 4.4: Define the coefficient matrix γ: Step 4.5: Find the optimal output sequence: Among them, e t-j is the noise at time tj; Step 4.6: Find the covariance of the optimal output sequence: Among them, Cov[e t-j ] is the noise e t-j The covariance of γ j The transpose of , N is the number of sampling points; Step 4.7: Obtain the LQG performance benchmark curve. The steps are as follows: Calculate LQG performance benchmark: By changing the value of λ from 0 to ∞, the optimal input-output variance of the rolling process under the LQG benchmark is obtained, and then the LQG benchmark performance curve of the rolling process is obtained; Step 4.8: Obtain the LQG index F using the LQG performance benchmark curve LQG : Among them, V y is the actual variance of the rolling process parameters at the current moment t, is the expected variance of the LQG performance benchmark corresponding to the rolling process parameters at the current time t.

5. The method for comprehensive monitoring of continuous rolling process taking rolling characteristics into consideration according to claim 1, characterized in that: The step 5 is specifically as follows: Step 5.1: Calculate the standard deviation σ of the rolling process parameters; Step 5.2: Calculate the mean of rolling process parameters Step 5.3: Calculate the CUSUM offset index F CUSUM : Among them, y n is the actual value of the nth rolling process parameter, is the average value of the rolling process parameters.

6. The method for comprehensive monitoring of a continuous rolling process taking rolling characteristics into consideration according to claim 1, wherein: In step 6, the comprehensive evaluation index of each rolling process parameter is calculated according to the following formula: Among them, F k Comprehensive evaluation index of the kth rolling process parameter, is the Hurst index of the kth rolling process parameter, is the LQG index of the kth rolling process parameter, is the CUSUM offset index of the kth rolling process parameter.

7. The method for comprehensive monitoring of a continuous rolling process taking rolling characteristics into consideration according to claim 1, wherein: The step 7 is specifically as follows: Step 7.1: Based on the constructed rolling process state space equation, apply 5% noise to each control variable in the rolling process control vector in turn, and calculate the change ΔW of the kth rolling process parameter caused by the noise of the pth control variable pk ; Step 7.2: Calculate the relative change of the kth rolling process parameter in, is the set value of the kth rolling process parameter; Step 7.3: Calculate the importance coefficient of the kth rolling process parameter Where P is the number of control mechanisms.

8. The method for comprehensive monitoring of a continuous rolling process taking rolling characteristics into consideration according to claim 1, wherein: In step 8, the comprehensive evaluation index of the continuous rolling process is calculated according to the following formula: Step 8: Construct comprehensive evaluation indicators for the continuous rolling process and evaluate the continuous rolling process, specifically: Step 8.1: Calculate the comprehensive evaluation index of the continuous rolling process: Where K is the number of rolling process parameters. Step 8.2: Establish the evaluation criteria for the health of the continuous rolling process based on the comprehensive evaluation indicators: When the comprehensive evaluation index is in the range of [0.9, 1], the continuous rolling process is in very good condition; when the comprehensive evaluation index is in the range of [0.7, 0.9), the continuous rolling process is degraded, but no maintenance is required; when the comprehensive evaluation index is in the range of [0, 0.7), the continuous rolling process fails and maintenance measures must be taken to maintain the continuous rolling process.

Images