Self-adaptive control method and system for modified asphalt production process

Abstract

The invention discloses a self-adaptive control method and system for a modified asphalt production process, and relates to the technical field of asphalt production. Comprising the following steps: S1, establishing a dynamic coupling relationship model among variables, and obtaining an interaction relationship among the variables; s2, through the dynamic coupling relation model, obtaining an initial set value of each variable, and meanwhile, through an intelligent decision model and a health condition monitoring model, adjusting the initial set values to obtain a final set value; and S3, acquiring a simulation result according to the final set value, comparing the simulation result with a preset target value, and determining a final control strategy according to a comparison result. According to the method, the equipment state information and the health condition monitoring model are combined, so that the production parameters are dynamically adjusted and optimized, the stability and the consistency of the product quality are ensured, the reliability of the production equipment is enhanced, and the service life of the production equipment is prolonged.

Description

Technical Field

[0001] The present invention relates to the technical field of asphalt production, and in particular to a method and system for adaptively controlling a modified asphalt production process. Background Art

[0002] Modified asphalt is a material that improves the performance of base asphalt by adding modifiers such as polymers, sulfur, and rubber. This material has higher durability, better adhesion and anti-rutting ability, and is widely used in road construction and maintenance. However, the production process of modified asphalt is complex and requires strict control of factors such as temperature, mixing ratio and time to ensure the quality of the final product.

[0003] The Chinese invention patent with announcement number CN113867292A discloses a rubber asphalt quality control method, collects data on parameters that affect the performance of modified asphalt, performs Latin hypercube sampling on parameters such as rubber powder fineness, rubber powder dosage, mixing time, reaction temperature, and development time as input variables, and uses the penetration and softening point of asphalt as output variables of the model to establish a neural network model, train the neural network model, and use the trained neural network model to simulate the test set data; set the required asphalt target penetration and target softening point, and continuously correct the parameter value range according to the probability of compliance of each influencing factor, and finally obtain the optimal parameter range that meets the requirements. Through this technical solution, the problems of high cost and low efficiency of rubber asphalt quality control in related technologies are solved.

[0004] In the process of implementing the above and similar quality control methods, there are complex interactions between multiple variables (such as temperature, pressure, mixing time, component ratio, etc.), and these variables not only directly affect the quality of the final product, but also have mutual coupling relationships with each other. In other words, once these coupling variables are handled incorrectly or not properly, it will lead to unstable product quality and low production efficiency, so a control scheme for the modified asphalt production process is needed that can be adaptively adjusted according to the actual changes in the coupling variables. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for adaptively controlling a modified asphalt production process to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: 1. A method for adaptively controlling a modified asphalt production process, comprising:

[0007] S1: Establish a dynamic coupling relationship model: obtain the key variables in the modified asphalt production process, establish a dynamic coupling relationship model between the variables, and obtain the interaction relationship between the variables;

[0008] S2: Determine the set value of each variable: obtain the initial set value of each variable through the dynamic coupling relationship model, and adjust the initial set value through the intelligent decision-making model and the health status monitoring model to obtain the final set value;

[0009] S3: Determine the control strategy: According to the final set value, obtain the simulation result, and compare the simulation result with the preset target value, and determine the final control strategy according to the comparison result, including:

[0010] S3.1: Obtain simulation results: Use the final set value as a production parameter and obtain simulation results through a simulation environment;

[0011] S3.2: Determine the final production parameters: Compare the simulation results with the preset target values, and determine the final production parameters based on the comparison results, specifically:

[0012] When the simulation result is not lower than the preset target value, the production parameter corresponding to the simulation result is the final production parameter, otherwise, step S2 to step S3.2 are repeated until the simulation result is not lower than the preset target value;

[0013] S3.3: Determine the final control strategy: Use the final production parameters as the operating production parameters to operate the equipment.

[0014] Furthermore, a dynamic coupling relationship model between variables is established, including:

[0015] M1: Construct a fast response variable sub-model: Based on the fast-changing variables among the key variables, the nonlinear relationship between temperature and pressure is obtained through a quadratic polynomial regression model, specifically:

[0016]

[0017] in: is the temperature at the next moment, is the temperature at the current moment, For the pressure of the present moment, , , , , is the regression coefficient, is the error term;

[0018] M2: Constructing a slow response variable sub-model: According to the slowly changing variables among the key variables, a slow response variable sub-model is constructed by using the ridge regression linear regression method to obtain the final product quality index;

[0019] M3: Constructing a dynamic coupling relationship model: According to the fast response variable sub-model and the slow response variable sub-model, the dynamic coupling relationship model is obtained, specifically:

[0020]

[0021] in: To predict product quality indicators, is the current mixing time, is the current component ratio, is the mixing time of the previous moment, is the composition ratio at the previous moment, is an intermediate variable.

[0022] Furthermore, the final product quality indicators are obtained, including:

[0023] X1: Determine feature engineering: construct interaction terms and higher-order terms through the mixing time and component ratio at the observation point, and obtain the mixing time and component ratio at the previous time point at the observation point;

[0024] X2: Constructing a slow response variable sub-model: According to the feature engineering and the mixing time and component ratio at the observation point, a slow response variable sub-model is constructed to obtain the final product quality index, specifically:

[0025]

[0026] in: The final product quality indicator is is the current mixing time, is the current component ratio, is the mixing time of the previous moment, is the composition ratio at the previous moment, is the error term of the quality indicator, is the intercept term, , , , , , , is the coefficient.

[0027] Furthermore, the formula for obtaining the regression coefficient is as follows:

[0028]

[0029] in: is the general term for regression coefficients, is the transpose of the design matrix, is the design matrix, is the dependent variable vector.

[0030] Furthermore, the interaction relationship between variables is obtained, including:

[0031] N1: Obtain partial derivatives of each variable: According to the predicted product quality index, obtain the partial derivatives of temperature, pressure, mixing time and component ratio, specifically:

[0032]

[0033] in: is the partial derivative of temperature, is the partial derivative of pressure, is the partial derivative of the mixing time, is the partial derivative of the component ratio, For variables The corresponding predicted product quality indicators, For variables The corresponding predicted product quality indicators, is the temperature at the current moment, is the temperature increment, For variables The corresponding predicted product quality indicators, For variables The corresponding predicted product quality indicators, For the pressure of the present moment, is the pressure increment, For variables The corresponding predicted product quality indicators, For variables The corresponding predicted product quality indicators, is the current mixing time, is the mixing time increment, For variables The corresponding predicted product quality indicators, For variables The corresponding predicted product quality indicators, is the current component ratio, is the increment of the component proportion;

[0034] N2: Obtain interaction effect evaluation: According to the key variables in the modified asphalt production process, set the interaction terms between the variables, and obtain the interaction effect parameters of the interaction terms. At the same time, according to the size of the interaction effect parameters, determine the synergy between the interaction terms, specifically:

[0035] When the interaction effect parameter corresponding to the interaction term between the i-th variable and the j-th variable is less than 0, there is a negative synergy between the i-th variable and the j-th variable; when the interaction effect parameter is greater than 0, there is a positive synergy between the i-th variable and the j-th variable; when the interaction effect parameter is equal to 0, there is a balanced synergy between the i-th variable and the j-th variable;

[0036] N3: Obtain the interaction relationship between variables: According to the partial derivatives and interaction effect evaluation of the variables, the interaction relationship between the variables is determined, specifically:

[0037] Changing any variable alone has limited impact on product quality, but the interaction effect between any two variables can significantly improve product quality.

[0038] Furthermore, the formula for obtaining the interaction effect parameter is as follows:

[0039]

[0040] in: is the interaction effect parameter corresponding to the interaction term between the ith variable and the jth variable, is the predicted product quality index corresponding to the interaction term between the i-th variable and the j-th variable, is the predicted product quality index corresponding to the i-th variable, is the predicted product quality index corresponding to the jth variable, It is the predicted product quality indicator under baseline conditions.

[0041] Furthermore, the final setting value is obtained, specifically:

[0042] S2.1: Obtaining initial setting values: Obtain the predicted product quality index through the dynamic coupling relationship model, and compare the predicted product quality index with the target product quality index range, and determine the initial setting value of each variable based on the comparison result, specifically:

[0043] When the predicted product quality index is within the target product quality index range, the variable setting value corresponding to the predicted product quality index is the initial setting value, otherwise the initial setting value is readjusted until the predicted product quality index is within the target product quality index range;

[0044] S2.2: Obtaining an intermediate set value: According to the current production status and target setting, an optimal control strategy is determined through an intelligent decision-making model, and according to the optimal control strategy, the intermediate set value is determined in the initial set value;

[0045] S2.3: Obtaining the final set value: According to the equipment status information, through the health status monitoring model, the production parameters are adjusted in the intermediate set value, and the adjusted production parameters are the final set values.

[0046] A modified asphalt production process adaptive control system, wherein the adaptive control system uses any of the above-mentioned modified asphalt production process adaptive control methods.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] Firstly, the present invention realizes dynamic adjustment and optimization of production parameters by combining equipment status information with health status monitoring model, thereby not only ensuring the stability and consistency of product quality, but also enhancing the reliability and service life of production equipment;

[0049] Second, the present invention monitors the health status of the equipment and adjusts the production parameters in real time to ensure stable and reliable product quality while optimizing resource utilization efficiency and reducing energy consumption and operating costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is a flow chart of the adaptive control method of the present invention;

[0051] Figure 2 A schematic diagram of a process for obtaining the interaction between variables in the present invention;

[0052] Figure 3 A schematic diagram of the process of obtaining the quality index of the final product in the present invention;

[0053] Figure 4 It is a schematic diagram of the process of obtaining the nonlinear relationship between temperature and pressure in the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] In the process of specific implementation of the existing quality control methods, there are complex interactions between multiple variables (such as temperature, pressure, mixing time, component ratio, etc.), and these variables will not only directly affect the quality of the final product, but also have mutual coupling relationships with each other. Therefore, once these coupled variables are handled incorrectly or not properly, it will lead to unstable product quality and low production efficiency. The technical solution of this application monitors and adjusts the key parameters (temperature, pressure, mixing time and component ratio) in the modified asphalt production process in real time, and dynamically optimizes the control strategy according to the actual operating conditions of the production line to ensure that stable production performance and product quality can be maintained even when the raw material characteristics or external conditions change. , thereby improving the flexibility and response speed of production, reducing human intervention, enhancing the reliability and stability of the system, while reducing energy consumption and operating costs, and promoting the effective use of resources and environmental protection.

[0056] refer to Figure 1-Figure 4 This embodiment provides an adaptive control method for a modified asphalt production process, and the adaptive control method includes the following steps:

[0057] Step S1: Establish a dynamic coupling relationship model. That is, real-time monitoring of key variables (such as temperature, pressure, mixing time, component ratio, etc.) in the modified asphalt production process, and establishing a dynamic coupling relationship model between the variables based on the key variables obtained, and at the same time, through the established dynamic coupling relationship model, determine the interaction relationship between the variables. The details are as follows:

[0058] Step S1.1: Obtain key variables. That is, in the production process of modified asphalt, high-precision sensors are used to monitor key variables in real time, including but not limited to temperature, pressure, mixing time and component ratio. Among them, temperature will affect the performance of modified asphalt, changes in pressure will affect the mixing effect and the physical properties of the final product, mixing time will affect the degree of interaction between the modifier and the base asphalt, and the component ratio will affect the later performance and maturity of the modified asphalt.

[0059] Furthermore, in this embodiment, the key variables obtained are preprocessed, including but not limited to filtering, standardization, missing value processing and anomaly detection, so as to obtain a cleaner and more consistent data set, thereby facilitating the subsequent establishment of a dynamic coupling relationship model.

[0060] Step S1.2: Establish a dynamic coupling relationship model. That is, according to the key variables preprocessed in step S1.1, determine the fast-changing variables and the slow-changing variables. At the same time, according to the fast-changing variables and the slow-changing variables, establish a fast-response variable sub-model and a slow-response variable sub-model, and fuse the fast-response variable sub-model and the slow-response variable sub-model to obtain a dynamic coupling relationship model. The details are as follows:

[0061] Step M1: Construct a fast response variable sub-model. That is, based on the fast-changing variables among the key variables, such as temperature and pressure, the nonlinear relationship between temperature and pressure is obtained through a quadratic polynomial regression model, specifically:

[0062]

[0063] in: is the temperature at the next moment, is the temperature at the current moment, For the pressure of the present moment, , , , , is the regression coefficient, is the error term.

[0064] In this embodiment, the regression coefficient can construct a design matrix based on the rapidly changing variables temperature and pressure, and obtain the specific size of the regression coefficient through the design matrix, which is specifically:

[0065]

[0066] in: is the general term for regression coefficients, is the transpose of the design matrix, is the design matrix, is the dependent variable vector.

[0067] In the process of specific implementation, the temperature and pressure data of the modified asphalt production process are obtained as shown in Table 1 below, specifically:

[0068] Table 1

[0069] According to the data in Table 1 above, the design matrix X is specifically:

[0070]

[0071] Therefore, taking the above data as an example, the design matrix X is specifically:

[0072]

[0073] At the same time, the dependent variable vector is the vector of the temperature at a given time point in this embodiment, that is, [178, 183, 192, 196, 199, 202, 205, 208] T , so the regression coefficient in this embodiment is , , , , They are: -1.748×10 5 , 1, -7.964×10 -2 , -1.374×10 -3 ,3.886×10 -8 ,3.886×10 -5 .

[0074] Furthermore, in this embodiment, the temperature change at 190°C and 2200Pa is obtained by a quadratic polynomial regression model. That is, according to the above obtained data, the predicted temperature is -1.797×10 5 .

[0075] Step M2: Construct a slow response variable sub-model. That is, based on the slowly changing variables among the key variables, such as mixing time, ingredient ratio and long-term quality indicators, and through the ridge regression linear regression method, construct a slow response variable sub-model to obtain the final product quality indicators. The details are as follows:

[0076] Step X1: Determine feature engineering. That is, according to the observation point, add the mixing time and component ratio of the previous time point as new features at each observation point, and construct interaction terms and high-order terms through the mixing time and component ratio at the observation point, that is, the product of mixing time and component ratio, the square of mixing time and the square of component ratio.

[0077] Step X2: Construct a slow response variable sub-model. That is, based on the feature engineering constructed in step X1 and the mixing time and component ratio at the observation point, construct a slow response variable sub-model to obtain the final product quality index, specifically:

[0078]

[0079] in: The final product quality indicator is is the current mixing time, is the current component ratio, is the mixing time of the previous moment, is the composition ratio at the previous moment, is the error term of the quality indicator, is the intercept term, , , , , , , is the coefficient.

[0080] It is worth noting that the coefficients in this embodiment can be determined by a cross-validation method. That is, in the specific implementation process, the coefficients are predefined (specifically set according to actual data, so no specific explanation is given in this embodiment), and the best value is determined from the predefined values ​​through cross-validation, that is, the obtained coefficient size.

[0081] In the process of specific implementation, the data of mixing time, component ratio and final quality index in the production process of modified asphalt shown in Table 2 below were obtained, specifically:

[0082] Table 2

[0083] time 1 2 3 4 5 6 Mixing time / min 30 45 50 40 55 60 Ingredient ratio / % 70 65 80 75 60 85 Final quality index 85 90 95 92 98 100

[0084] According to the data in Table 2 above, we can get the data shown in Table 3 below (including feature engineering), specifically:

[0085] Table 3

[0086] At the same time, the above coefficient , , , , , , are 0.5, 0.8, -0.1, 0.2, 0.01, -0.005 and 0.003 respectively, and the intercept term is 50. Therefore, according to the data in Table 3 above, the quality indicators of the final products obtained are: 110.8, 123.2, 136.55, 127.875, 128.975 and 146.075.

[0087] Step M3: construct a dynamic coupling relationship model. That is, according to the fast response variable sub-model constructed in step M1, the nonlinear relationship between temperature and pressure is obtained, and the slow response variable sub-model constructed in step M2 is used to obtain the predicted product quality index, which is integrated to obtain the dynamic coupling relationship model, which is specifically:

[0088]

[0089] in: To predict product quality indicators, is the current mixing time, is the current component ratio, is the mixing time of the previous moment, is the composition ratio at the previous moment, is an intermediate variable.

[0090] In this embodiment, the intermediate variable The specific formula for obtaining is:

[0091]

[0092] in: is the intermediate variable, is the temperature at the current moment, For the pressure of the present moment.

[0093] Step S1.3: Obtain the interaction relationship between the variables. That is, according to the predicted product quality index obtained in step M3, determine the interaction relationship between the variables. The details are as follows:

[0094] Step N1: Obtain the partial derivatives of each variable. That is, through the predicted product quality index in step M3, obtain the partial derivative of temperature, the partial derivative of pressure, the partial derivative of mixing time and the partial derivative of ingredient ratio. In other words, by setting the increment, when obtaining the partial derivative of a single variable, keep other variables unchanged, change the value of each variable separately, and record the change of the predicted product quality index after each change, and obtain the partial derivative of the variable according to the change of the predicted product quality index, specifically:

[0095]

[0096] in: is the partial derivative of temperature, is the partial derivative of pressure, is the partial derivative of the mixing time, is the partial derivative of the component ratio, For variables The corresponding predicted product quality indicators, For variables The corresponding predicted product quality indicators, is the temperature at the current moment, is the temperature increment, For variables The corresponding predicted product quality indicators, For variables The corresponding predicted product quality indicators, For the pressure of the present moment, is the pressure increment, For variables The corresponding predicted product quality indicators, For variables The corresponding predicted product quality indicators, is the current mixing time, is the mixing time increment, For variables The corresponding predicted product quality indicators, For variables The corresponding predicted product quality indicators, is the current component ratio, is the increment of the component ratio.

[0097] In the specific implementation process, the temperature increment is set to 1°C, the pressure increment is set to 10Pa, the mixing time increment is set to 1min, and the ingredient ratio increment is set to 1%. Specifically, when the temperature increases from 200°C to 201°C, while other variables remain unchanged, the predicted product quality index increases by 0.5%, and the partial derivative of temperature is 0.5% / °C. When the pressure increases from 2100Pa to 2110Pa, while other variables remain unchanged, the predicted product quality index increases by 0.2%, and the partial derivative of pressure is 0.02% / Pa. When the mixing time increases from 46min to 46min, while other variables remain unchanged, the predicted product quality index increases by 0.8%, and the partial derivative of the mixing time is 0.8% / min. When the ingredient ratio increases from 65% to 66%, while other variables remain unchanged, the predicted product quality index increases by 1.5%, and the partial derivative of the ingredient ratio is 1.5% / %.

[0098] Furthermore, according to the above data, the partial derivative of the ingredient ratio is the largest. That is, under the same conditions, a small change in the ingredient ratio will lead to the largest change in product quality. Similarly, the partial derivative of pressure is the smallest, that is, under the same conditions, a small change in pressure will have the smallest impact on product quality.

[0099] Step N2: Obtain interaction effect evaluation. That is, according to the key variables in the modified asphalt production process obtained in step S1.1, set the interaction terms between the variables, that is, the interaction between temperature and component ratio, and the interaction between pressure and mixing time. At the same time, according to the set interaction terms, obtain the impact of the interaction terms on the predicted product quality indicators. In other words, by setting the increment of the interaction term, obtain the predicted product quality indicator corresponding to the interaction term after adding the increment, and compare the predicted product quality indicator corresponding to the added increment with the predicted product quality indicator corresponding to the absence of the added increment, and obtain its interaction effect parameters, specifically:

[0100]

[0101] in: is the interaction effect parameter corresponding to the interaction term between the ith variable and the jth variable, is the predicted product quality index corresponding to the interaction term between the i-th variable and the j-th variable, is the predicted product quality index corresponding to the i-th variable, is the predicted product quality index corresponding to the jth variable, It is the predicted product quality index under the baseline conditions.

[0102] Furthermore, the synergistic effect between different variables is determined based on the size of the obtained interaction effect parameter. Specifically, when the interaction effect parameter corresponding to the interaction term between the i-th variable and the j-th variable is When it is less than 0, it indicates that there is a negative synergy between the i-th variable and the j-th variable. When it is greater than 0, it indicates that there is a positive synergy between the i-th variable and the j-th variable. When it is equal to 0, it indicates that there is a balanced synergy between the i-th variable and the j-th variable.

[0103] That is to say, when the interaction effect parameter corresponding to the interaction term between temperature and component ratio is less than 0, it indicates that there is a negative synergy between temperature and component ratio; when it is greater than 0, it indicates that there is a positive synergy between temperature and component ratio; and when it is equal to 0, it indicates that there is a balanced synergy between temperature and component ratio.

[0104] Similarly, when the interaction effect parameter corresponding to the interaction term between temperature and pressure is less than 0, it indicates that there is a negative synergy between temperature and pressure. When it is greater than 0, it indicates that there is a positive synergy between temperature and pressure. When it is equal to 0, it indicates that there is a balanced synergy between temperature and pressure.

[0105] Similarly, when the interaction effect parameter corresponding to the interaction term between temperature and mixing time is less than 0, it indicates that there is a negative synergy between temperature and mixing time. When it is greater than 0, it indicates that there is a positive synergy between temperature and mixing time. When it is equal to 0, it indicates that there is a balanced synergy between temperature and mixing time.

[0106] Similarly, when the interaction effect parameter corresponding to the interaction term between pressure and mixing time is less than 0, it indicates that there is a negative synergy between pressure and mixing time, when it is greater than 0, it indicates that there is a positive synergy between pressure and mixing time, and when it is equal to 0, it indicates that there is a balanced synergy between pressure and mixing time.

[0107] Similarly, when the interaction effect parameter corresponding to the interaction term between pressure and component ratio is less than 0, it indicates that there is a negative synergy between pressure and component ratio; when it is greater than 0, it indicates that there is a positive synergy between pressure and component ratio; and when it is equal to 0, it indicates that there is a balanced synergy between pressure and component ratio.

[0108] Similarly, when the interaction effect parameter corresponding to the interaction term between the ingredient ratio and the mixing time is less than 0, it indicates that there is a negative synergy between the ingredient ratio and the mixing time. When it is greater than 0, it indicates that there is a positive synergy between the ingredient ratio and the mixing time. When it is equal to 0, it indicates that there is a balanced synergy between the ingredient ratio and the mixing time.

[0109] In the process of specific implementation, the predicted product quality index under the baseline conditions was 90%, and when the temperature increased from 200℃ to 205℃, while other variables remained unchanged, the predicted product quality index increased to 90.5%. When the ingredient ratio increased from 60% to 61%, while other variables remained unchanged, the predicted product quality index increased to 93%. When the temperature and ingredient ratio increased to 205℃ and 61% at the same time, while other variables remained unchanged, the predicted product quality index increased to 95%. Therefore, the interaction effect parameter corresponding to the interaction term between temperature and ingredient ratio is 1.5%. In other words, there is a significant positive synergy between temperature and ingredient ratio.

[0110] Likewise, the predicted product quality index under the baseline condition was 90%, and when the temperature was increased from 200°C to 205°C, while other variables remained unchanged, the predicted product quality index increased to 90.5%. 5 Pa increased to 1.1×10 5 Pa, while other variables remain unchanged, the predicted product quality index increases to 91%. When the temperature and pressure are increased to 205 °C and 1.1×10 5 When the temperature and pressure are changed, and other variables remain unchanged, the predicted product quality index increases to 91.5%. Therefore, the interaction effect parameter corresponding to the interaction term between temperature and pressure is 0%. In other words, there is a balanced synergy between temperature and pressure.

[0111] Similarly, the predicted product quality index under the baseline condition is 90%, and when the temperature is increased from 200℃ to 205℃, while other variables remain unchanged, the predicted product quality index increases to 90.5%. When the mixing time is increased from 10min to 11min, while other variables remain unchanged, the predicted product quality index increases to 93%. When the temperature and mixing time are increased to 205℃ and 11min at the same time, while other variables remain unchanged, the predicted product quality index increases to 93.5%. Therefore, the interaction effect parameter corresponding to the interaction term between temperature and mixing time is 0%. In other words, there is a balanced synergy between temperature and mixing time.

[0112] Similarly, the predicted product quality index under the baseline condition is 90%, and when the pressure is increased from 10 5 Pa increased to 1.1×10 5 When the pressure and mixing time were increased to 1.1×10 5 When the pressure and mixing time are increased by 10.5 and 11 min, and other variables remain unchanged, the predicted product quality index increases to 95%. Therefore, the interaction effect parameter corresponding to the interaction term between pressure and mixing time is 1%. In other words, there is a positive synergy between pressure and mixing time.

[0113] Similarly, the predicted product quality index under the baseline condition is 90%, and when the pressure is increased from 10 5 Pa increased to 1.1×10 5 When the pressure and the ingredient ratio were increased to 1.1×10 5 When Pa and 61%, while other variables remain unchanged, the predicted product quality index increases to 92.5%. Therefore, the interaction effect parameter corresponding to the interaction term between pressure and ingredient ratio is -1.5%. In other words, when there is a negative synergy between pressure and ingredient ratio.

[0114] Similarly, the predicted product quality index under the baseline condition is 90%, and when the ingredient ratio increases from 60% to 61% while other variables remain unchanged, the predicted product quality index increases to 93%. When the mixing time increases from 10min to 11min while other variables remain unchanged, the predicted product quality index increases to 93%. When the ingredient ratio and mixing time are increased to 61% and 11min at the same time while other variables remain unchanged, the predicted product quality index increases to 97.5%. Therefore, the interaction effect parameter corresponding to the interaction term between ingredient ratio and mixing time is 1.5%. In other words, there is a positive synergy between ingredient ratio and mixing time.

[0115] Step N3: Obtain the interaction relationship between variables. That is, according to the partial derivatives of each variable obtained in step N1 and the interaction effect evaluation obtained in step N2, determine the interaction relationship between the variables, specifically:

[0116] Changing any variable (temperature, pressure, mixing time, ingredient ratio) individually has limited effect on product quality, but the interaction effect between any two variables (interaction between pressure and mixing time, interaction between pressure and temperature, interaction between pressure and ingredient ratio, interaction between temperature and ingredient ratio, interaction between temperature and mixing time, interaction between ingredient ratio and mixing time) can significantly improve product quality.

[0117] In the process of specific implementation, according to the above data, it can be seen that: changing the temperature alone predicts that the product quality index will increase by 0.5%, but when it works together with the ingredient ratio, it shows a significant positive synergistic effect, and the predicted product quality index will increase by 4.5%. Further, changing the pressure alone predicts that the product quality index will increase by 1%, but when it works together with the mixing time, it shows a significant positive synergistic effect, and the predicted product quality index will increase by 5%. Further, changing the mixing time and the ingredient ratio alone predicts that the product quality index will only increase by 3%, but when the two work together, they show a great positive synergistic effect, and the predicted product quality index will increase by 7%.

[0118] Step S2: Determine the set value of each variable. That is, the initial set value of each variable is determined through the dynamic coupling relationship model established in step M3, and the intermediate set value is obtained through the intelligent decision-making model based on the initial set value, and then the final set value is determined from the intermediate set value through the health status monitoring model. The details are as follows:

[0119] Step S2.1: Obtaining initial setting values. That is, according to the dynamic coupling relationship model established in step M3, the predicted product quality index is obtained, and the predicted product quality index is compared with the target product quality index range, and according to the comparison result, the initial setting value of each variable is determined, specifically:

[0120] When the predicted product quality index is within the target product quality index range, the variable setting value corresponding to the predicted product quality index is the initial setting value of the variable. Otherwise, the initial setting value of the variable is readjusted until the predicted product quality index is within the target product quality index range.

[0121] In the specific implementation process, the initial temperature is set to 200°C and the initial pressure is set to 10 5 Pa, the initial mixing time is set to 10 min, the initial component ratio is set to 60%, the predicted product quality index is set to 93%, and the target product quality index range is [95%, 100%]. In other words, according to the initial setting values ​​of the current variables, the predicted product quality index obtained is not within the target product quality index range.

[0122] Furthermore, in this embodiment, the mixing time and the ingredient ratio are increased to 11 minutes and 61% respectively. At the same time, the temperature and pressure are kept unchanged. At this time, the predicted product quality index corresponding to the interaction term between the mixing time and the ingredient ratio increases by 7%, that is, the newly obtained predicted product quality index is 97%. Since 97% is within the range of [95%, 100%], the mixing time of 11 minutes and the ingredient ratio of 61% can be used as the initial setting values.

[0123] Step S2.2: Obtain the intermediate set value. That is, according to the current production status and target setting, the optimal control strategy is determined through the intelligent decision-making model. And within the initial set value range obtained in step S2.1, the intermediate set value of each variable is determined according to the optimal control strategy.

[0124] In this embodiment, the intelligent decision model (such as a rule-based expert system, a machine learning model, and Bayesian optimization) determines the optimal control strategy based on the current production status (current temperature, pressure, mixing time, and ingredient ratio) and the target setting (preset product quality index range) through a multi-stage decision analysis method. And based on the determined optimal control strategy, the intermediate setting value of each variable is determined.

[0125] In the specific implementation process, the current temperature is set to 205°C and the current pressure is set to 1.05×10 5 Pa, the current mixing time is set to 11min, the current ingredient ratio is set to 61%, the predicted product quality index is set to 97%, and the target product quality index range is [95%, 100%]. At this time, the predicted product quality index is already within the target product quality index range.

[0126] Furthermore, according to the preset product quality index range, the predicted product quality index is always within the target product quality index range, while minimizing the adjustment cost or other secondary goals (such as energy consumption), and the predicted product quality index is further adjusted. Specifically, in this embodiment, the temperature and ingredient ratio are reduced to 204°C and 60.8% respectively, and the mixing time is increased to 11.2 minutes through the intelligent decision-making model.

[0127] Step S2.3: Obtain the final set value. That is, according to the equipment status information, through the health status monitoring model, the production parameters are adjusted from the intermediate set values ​​determined in step S2.2, that is, the final set value is obtained.

[0128] In this embodiment, a health status monitoring model is established through statistical process control, classifiers in machine learning, or deep learning networks, and features in normal operation mode are identified, while abnormal conditions are detected. That is, through the health status monitoring model, the equipment is always in a healthy operating state when running the current production parameters.

[0129] In the specific implementation process, when the health status monitoring model detects a warning state, the production parameters are adjusted according to actual needs. Specifically, since the temperature has a greater impact on the equipment load, the temperature is adjusted to 203°C. Furthermore, since the mixing time is relatively short and has a smaller impact on the equipment, it can remain unchanged. The adjusted production parameters are the final set values ​​obtained.

[0130] Step S3: Determine the control strategy. According to the final set value obtained in step S2.3, the final set value is used as the production parameter, and a simulation operation is performed to obtain the simulation result. At the same time, the simulation result is compared with the preset target value, and the final control strategy is determined according to the comparison result. The details are as follows:

[0131] Step S3.1: Get simulation results. That is, the final setting value obtained in step S2.3 is used as the production parameter, and the simulation result is obtained through the constructed simulation environment (such as Aspen Plus, MATLAB / Simulink). That is, in the constructed simulation environment, the final setting value is used as the input parameter, and the corresponding simulation result is output.

[0132] In the specific implementation process, the final set values ​​were set to 203°C for temperature and 1.05×10 5Pa, mixing time 11.2min, ingredient ratio 60.8%. At the same time, through the constructed simulation environment, the simulation results obtained are: simulated product quality 97.4%, simulated output 1000kg / h, simulated energy consumption 50kWh / kg, simulated equipment utilization 95%.

[0133] Step S3.2: Determine the final production parameters. That is, according to the actual production needs, set the preset target value (it can be set specifically according to the actual needs, so it is not specifically explained in this embodiment). At the same time, compare the simulation results obtained in step S3.1 with the preset target value, and determine the final production parameters based on the comparison results. Specifically:

[0134] When the simulation result is not lower than the preset target value, the production parameter corresponding to the simulation result is the final production parameter. Otherwise, step S2 to step S3.2 are repeated until the simulation result is not lower than the preset target value.

[0135] In the specific implementation process, the preset target values ​​are set as: simulated product quality 97.5%, simulated output 1000kg / h, simulated energy consumption 48kWh / kg, simulated equipment utilization 95%. According to the simulation results obtained above, it can be seen that: the simulated product quality is close to the target value, the output meets the target value, the energy consumption is slightly higher than the target value, and the equipment utilization reaches the target level. In other words, the production parameters corresponding to the above simulation results meet the production and quality requirements during operation, but the energy consumption needs to be further improved.

[0136] Step S3.3: Determine the final control strategy. That is, according to the final production parameters obtained in step S3.2, the final production parameters are used as the operating production parameters to carry out specific operation of the equipment.

[0137] This embodiment also provides an adaptive control system for a modified asphalt production process, which uses an adaptive control method for a modified asphalt production process described in the above embodiment.

[0138] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is limited by the attached embodiments and their equivalents.

Claims

1. A method for adaptive control of modified asphalt production process, characterized in that: Included are: S1: Establish a dynamic coupling relationship model: obtain the key variables in the modified asphalt production process, establish a dynamic coupling relationship model between the variables, and obtain the interaction relationship between the variables; S2: Determine the set value of each variable: obtain the initial set value of each variable through the dynamic coupling relationship model, and adjust the initial set value through the intelligent decision-making model and the health status monitoring model to obtain the final set value; S3: Determine the control strategy: According to the final set value, obtain the simulation result, and compare the simulation result with the preset target value, and determine the final control strategy according to the comparison result, including: S3.1: Obtain simulation results: Use the final set value as a production parameter and obtain simulation results through a simulation environment; S3.2: Determine the final production parameters: Compare the simulation results with the preset target values, and determine the final production parameters based on the comparison results, specifically: When the simulation result is not lower than the preset target value, the production parameter corresponding to the simulation result is the final production parameter, otherwise, step S2 to step S3.2 are repeated until the simulation result is not lower than the preset target value; S3.3: Determine the final control strategy: Use the final production parameters as the operating production parameters to operate the equipment.

2. The adaptive control method for modified asphalt production process according to claim 1, characterized in that: Establish a dynamic coupling relationship model between variables, including: M1: Construct a fast response variable sub-model: Based on the fast-changing variables among the key variables, the nonlinear relationship between temperature and pressure is obtained through a quadratic polynomial regression model, specifically: ,in: is the temperature at the next moment, is the temperature at the current moment, For the pressure of the present moment, , , , , is the regression coefficient, is the error term; M2: Constructing a slow response variable sub-model: Based on the slowly changing variables among the key variables, a slow response variable sub-model is constructed by using the ridge regression linear regression method to obtain the final product quality index; M3: Constructing a dynamic coupling relationship model: According to the fast response variable sub-model and the slow response variable sub-model, the dynamic coupling relationship model is obtained, specifically: ,in: To predict product quality indicators, is the current mixing time, is the current component ratio, is the mixing time of the previous moment, is the composition ratio at the previous moment, is an intermediate variable.

3. The adaptive control method for modified asphalt production process according to claim 2, characterized in that: Obtain final product quality indicators, including: X1: Determine feature engineering: construct interaction terms and higher-order terms through the mixing time and component ratio at the observation point, and obtain the mixing time and component ratio at the previous time point at the observation point; X2: Constructing a slow response variable sub-model: According to the feature engineering and the mixing time and component ratio at the observation point, a slow response variable sub-model is constructed to obtain the final product quality index, specifically: ,in: It is the quality index of the final product. is the current mixing time, is the current component ratio, is the mixing time of the previous moment, is the composition ratio at the previous moment, is the error term of the quality indicator, is the intercept term, , , , , , , is the coefficient.

4. The adaptive control method for modified asphalt production process according to claim 2, characterized in that: The formula for obtaining the regression coefficient is as follows: ,in: is the general term for regression coefficients, is the transpose of the design matrix, is the design matrix, is the dependent variable vector.

5. The adaptive control method for modified asphalt production process according to claim 1, characterized in that: Obtain the interaction relationship between variables, including: N1: Obtain partial derivatives of each variable: According to the predicted product quality index, obtain the partial derivatives of temperature, pressure, mixing time and component ratio, specifically: ,in: is the partial derivative of temperature, is the partial derivative of pressure, is the partial derivative of the mixing time, is the partial derivative of the component ratio, For variables The corresponding predicted product quality indicators, For variables The corresponding predicted product quality indicators, is the temperature at the current moment, is the temperature increment, For variables The corresponding predicted product quality indicators, For variables The corresponding predicted product quality indicators, For the pressure of the present moment, is the pressure increment, For variables The corresponding predicted product quality indicators, For variables The corresponding predicted product quality indicators, is the current mixing time, is the mixing time increment, For variables The corresponding predicted product quality indicators, For variables The corresponding predicted product quality indicators, is the current component ratio, is the increment of the component proportion; N2: Obtain interaction effect evaluation: According to the key variables in the modified asphalt production process, set the interaction terms between the variables, and obtain the interaction effect parameters of the interaction terms. At the same time, according to the size of the interaction effect parameters, determine the synergy between the interaction terms, specifically: When the interaction effect parameter corresponding to the interaction term between the i-th variable and the j-th variable is less than 0, there is a negative synergy between the i-th variable and the j-th variable; when the interaction effect parameter is greater than 0, there is a positive synergy between the i-th variable and the j-th variable; when the interaction effect parameter is equal to 0, there is a balanced synergy between the i-th variable and the j-th variable; N3: Obtain the interaction relationship between variables: According to the partial derivatives and interaction effect evaluation of the variables, the interaction relationship between the variables is determined, specifically: Changing any variable alone has limited impact on product quality, but the interaction effect between any two variables can significantly improve product quality.

6. The adaptive control method for modified asphalt production process according to claim 5, characterized in that: The formula for obtaining the interaction effect parameter is as follows: ,in: is the interaction effect parameter corresponding to the interaction term between the ith variable and the jth variable, is the predicted product quality index corresponding to the interaction term between the i-th variable and the j-th variable, is the predicted product quality index corresponding to the i-th variable, is the predicted product quality index corresponding to the jth variable, It is the predicted product quality indicator under baseline conditions.

7. The adaptive control method for modified asphalt production process according to claim 1, characterized in that: Get the final setting value, specifically: S2.1: Obtaining initial setting values: Obtain the predicted product quality index through the dynamic coupling relationship model, and compare the predicted product quality index with the target product quality index range, and determine the initial setting value of each variable based on the comparison result, specifically: When the predicted product quality index is within the target product quality index range, the variable setting value corresponding to the predicted product quality index is the initial setting value, otherwise the initial setting value is readjusted until the predicted product quality index is within the target product quality index range; S2.2: Obtaining an intermediate set value: According to the current production status and target setting, an optimal control strategy is determined through an intelligent decision-making model, and according to the optimal control strategy, the intermediate set value is determined in the initial set value; S2.3: Obtaining the final set value: According to the equipment status information, through the health status monitoring model, the production parameters are adjusted in the intermediate set value, and the adjusted production parameters are the final set values.

8. An adaptive control system for a modified asphalt production process, characterized in that: The adaptive control system uses an adaptive control method for a modified asphalt production process as described in any one of claims 1-7.

Images