A self-adjusting reaction kettle temperature control method based on fuzzy Smith-PID

Abstract

The application discloses a kind of self-regulating reaction kettle temperature control methods based on fuzzy Smith-PID, belong to temperature control technical field, its steps are: first, the first-order inertia model of reaction kettle temperature system is obtained, then the temperature error and error change quantity are fuzzed, fuzzy rule is formulated, then the real-time PID parameter value is obtained by fuzzy rule table, self-regulating control is realized, finally, the Smith predictor model of reaction kettle is obtained by offline test method, and is brought into system to obtain the self-regulating reaction kettle temperature control method based on fuzzy Smith-PID.The method of the application is aimed at the large lag, nonlinearity, strong interference of reaction kettle temperature control system, the self-regulating of controller is realized by the method of fuzzy Smith-PID, and the stability and response speed of system are improved, and the application field of reaction kettle system is widened.

Description

Technical Field

[0001] This invention belongs to the field of temperature control technology, specifically relating to a self-regulating reactor temperature control method based on fuzzy Smith-PID. Background Technology

[0002] Reactors are indispensable chemical reaction vessels in pharmaceutical companies, and proper temperature control of these reactors is a critical concern. Reaction temperature directly impacts drug quality, affecting patient health and safety, and sometimes even the company's future. Therefore, designing a system with effective reactor temperature control is of paramount importance. In batch production processes, the chemical reactions within the reactor typically involve two phases: the heating and initial reaction stage, and the exothermic later stage. These two phases have distinct characteristics. If excess heat is not removed promptly or uneven heating occurs, overheating can easily occur, severely impacting production safety. Conversely, excessive heat removal can cause a rapid temperature drop, slowing the reaction and potentially leading to reactor shutdown. Thus, product quality pass rates and production safety are closely related to the accuracy and stability of the reactor temperature control system. However, in actual production, unpredictable external factors can significantly affect the control of the reactor, increasing the difficulty of reactor temperature control. The main challenges in controlling the temperature of the reactor are as follows: (1) Complexity of chemical reactions: The chemical reaction process of raw materials is accompanied by a series of complex physical and chemical reactions and the conversion and transfer of matter and energy. Therefore, the production process is accompanied by large time-varying, high nonlinearity and uncertainty. (2) Difficulty in controlling the temperature of the reactor: The reactors used in industry generally have a large capacity, which makes the reactor walls thick. These structural characteristics determine that the reactor is a controlled object with a large lag and a large heat capacity. The heat in exothermic or endothermic reactions is difficult to remove or supply in time, which exacerbates the lag in control. (3) Modeling difficulties: Due to the complex chemical reaction mechanism of the reaction process and the significant impact of complex external conditions on the system, it is difficult to derive the mechanism model. Furthermore, due to the complexity and nonlinearity of the exothermic and endothermic reactions in the reaction process, and the fact that the irregular changes in the heat transfer coefficient of the heat-conducting medium are highly sensitive to various external disturbances as the reaction process progresses, and the non-repeatability of the process in the batch reaction process, it is difficult to achieve accurate modeling of the reaction process using traditional modeling methods such as the mechanism method and the least squares method. In summary, temperature control in reactors still faces many challenges, such as large hysteresis, nonlinearity, strong interference, and difficulty in modeling, all of which urgently need to be addressed. Currently, industrial applications mainly employ methods like PID control for reactor temperature control, which offers advantages such as simple control principle, ease of tuning, no need for modeling, and high robustness and reliability. However, when faced with phenomena like large hysteresis and strong interference in actual temperature control, PID controllers often struggle to achieve rapid and effective adjustment and control. Therefore, more advanced methods are needed to address these issues in such systems. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a self-regulating reactor temperature control method based on fuzzy Smith-PID. This method combines a fuzzy controller, a Smith predictor compensator, and PID control. On one hand, it utilizes fuzzy inference to adjust the PID controller parameters online, achieving self-adjustment of the PID parameters. On the other hand, it uses the Smith predictor compensator to compensate for the pure time delay of the controlled process, allowing the delayed controlled variable to be reflected in advance at the system deviation input. This causes the PID controller to act earlier, thereby reducing overshoot, accelerating the adjustment process, and improving control performance. The technical solution is as follows: A self-regulating reactor temperature control method based on fuzzy Smith-PID includes the following steps: 1) Considering the large inertia and large time lag characteristics of the controlled object in the self-regulating reactor temperature control system based on fuzzy Smith-PID, a first-order inertial model with an added time delay element is selected to describe the reactor temperature control system model. ,in, For the static gain of the system, Let be the system's time constant. The system's lag time. It is a complex variable; 2) Utilizing PID control algorithm and fuzzy control algorithm, combined with the system's real-time temperature error. and temperature error change rate The PID parameters are updated in real time to obtain the structure diagram of the fuzzy PID controller and realize self-adjusting control. 3) A Smith predictor model for the reactor was obtained through offline testing. This Smith predictor model was then combined with the reactor temperature control system model, and a fuzzy PID controller was used to obtain a time-independent predictor. The impact on the self-regulating reactor temperature control system based on fuzzy Smith-PID.

[0004] Preferably, step 2) specifically involves the following steps: a sets the temperature deviation to The rate of change of temperature deviation is set to PID parameters are used as inputs to the fuzzy control algorithm. , , Change , , As its output, a fuzzy controller with a two-input, three-output structure is obtained; b. Determine the temperature deviation within the temperature control and allowable error range. Temperature change rate Domain of discourse , Discrete points and quantization factors are selected, and based on the determined universe of discourse, discrete points, and quantization factors, a set of... , , Quantization parameters; c. Select the fuzzy set of linguistic variables for input and output variables. For temperature error and temperature error change rate, define the fuzzy subsets on their universe of discourse as: PB (positive large), PM (positive medium), PS (positive small), ZO (zero), NS (negative), NM (negative medium), NB (negative large). The fuzzy subset of the control parameter output is PB (positive large), PM (positive medium), PS (positive small), ZO (zero), NS (negative), NM (negative medium), NB (negative large). d Set temperature deviation value and the rate of change of temperature deviation The quantization interval is E and EC. The temperature deviation value is obtained using the membership function. and the rate of change of temperature deviation The membership degrees are calculated over the quantization interval, and the parameters are normalized based on the obtained membership degrees. Then, based on the established fuzzy rule table, corresponding inference calculations are performed to obtain the correction values ​​for the three parameters of the corresponding PID. , , The fuzzy adjustment value; The fuzzy rule table is established as follows: When the deviation is large, in order to speed up the system's response and prevent the control action from exceeding the range due to possible differential oversaturation caused by the instantaneous increase in deviation at the beginning. A larger value should be selected. and smaller In addition, to prevent integral saturation and avoid large overshoot in the system response, The value should be small. When the deviation is small, in order to make the system have better steady-state performance, it should be increased. , The value should be appropriately selected to avoid oscillation of the output response around the set value and to consider the system's anti-interference capability. The principle is that when the rate of change of deviation is small, Take a larger value; when the rate of change of deviation is large, take a smaller value. Based on the above principles, 7*7=49 fuzzy rules can be formulated. Based on the reactor temperature control rules, fuzzy inference calculations are performed. Fuzzy inference is essentially a fuzzification process, determining the fuzzy quantity corresponding to each precise quantity. The result of fuzzy inference is a fuzzy set, which needs to be defuzzified to obtain a precise value as the control signal to drive the actuator. A weighted average method is used to defuzzify and obtain the corresponding precise value. , , The parameter tuning formula can be expressed as: (1) In the formula, , , Corresponding to , , The initial value. Preferably, step 3) has the following specific steps: To ensure that the equivalent transfer function no longer contains a pure time delay element, a parallel connection is used in the control algorithm. The compensator enabled the reactor temperature control system model to be developed earlier. At that moment, the pure time delay element in the reactor temperature control system model was eliminated. From this, we can obtain the transfer function as:

[0005]

[0006] The compensated transfer function is: (2) From the above equation, we can conclude that after compensation by the predictor, the stability of the system is no longer affected. The impact.

[0007] Compared with the prior art, the present invention has the following advantages: 1. Compared with PID control systems, the method of this invention adopts fuzzy control algorithm to realize the dynamic adjustment function of PID parameters. This method improves the static and dynamic performance of the system and enhances the overall response speed, robustness and stability of the system.

[0008] 2. The method of the present invention addresses the issues of large hysteresis, nonlinearity, and strong interference in the temperature control system of a reactor. By employing the fuzzy Smith-PID method, the controller achieves self-adjustment, thereby improving the stability and response speed of the system and broadening the application range of the reactor system. Attached Figure Description

[0009] Figure 1 This is a diagram of the fuzzy PID control structure in the method of this invention; Figure 2 This is a structural diagram of the Smith predictor in the method of the present invention; Figure 3 This is a structural diagram of the fuzzy Smith-PID controller in the method of the present invention; Figure 4 This is a structural diagram of the reactor temperature control system in Embodiment 1 of the present invention. Detailed Implementation

[0010] The technical solution of the present invention will be clearly and completely described below with reference to specific embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0011] A self-regulating reactor temperature control method based on fuzzy Smith-PID includes the following steps: 1) Considering the large inertia and large time lag characteristics of the controlled object in the self-regulating reactor temperature control system based on fuzzy Smith-PID, a first-order inertial model with an added time delay element is selected to describe the reactor temperature control system model. ,in, For the static gain of the system, Let be the system's time constant. The system's lag time. It is a complex variable; 2) Utilizing PID control algorithm and fuzzy control algorithm, combined with the system's real-time temperature error. and temperature error change rate The PID parameters are updated in real time to obtain the fuzzy PID controller structure diagram, as shown in the attached diagram. Figure 1 As shown, self-regulating control is achieved; 3) A Smith predictor model for the reactor was obtained through offline testing. This Smith predictor model was then combined with the reactor temperature control system model, and a fuzzy PID controller was used to obtain a time-independent predictor. The influence of the fuzzy Smith-PID-based self-regulating reactor temperature control system is shown in the figure below. Figure 3 As shown.

[0012] Preferably, step 2) specifically involves the following steps: a sets the temperature deviation to The rate of change of temperature deviation is set to PID parameters are used as inputs to the fuzzy control algorithm. , , Change , , As its output, a fuzzy controller with a two-input, three-output structure is obtained; b. Determine the temperature deviation within the temperature control and allowable error range. Temperature change rate Domain of discourse , Discrete points and quantization factors are selected, and based on the determined universe of discourse, discrete points, and quantization factors, a set of... , , Quantization parameters; c. Select the fuzzy set of linguistic variables for input and output variables. For temperature error and temperature error change rate, define the fuzzy subsets on their universe of discourse as: PB (positive large), PM (positive medium), PS (positive small), ZO (zero), NS (negative), NM (negative medium), NB (negative large). The fuzzy subset of the control parameter output is PB (positive large), PM (positive medium), PS (positive small), ZO (zero), NS (negative), NM (negative medium), NB (negative large). d Set temperature deviation value and the rate of change of temperature deviation The quantization interval is E and EC. The temperature deviation value is obtained using the membership function. and the rate of change of temperature deviation The membership degrees are calculated over the quantization interval, and the parameters are normalized based on the obtained membership degrees. Then, based on the established fuzzy rule table, corresponding inference calculations are performed to obtain the correction values ​​for the three parameters of the corresponding PID. , , The fuzzy adjustment value; The fuzzy rule table is established as follows: When the deviation is large, in order to speed up the system's response and prevent the control action from exceeding the range due to possible differential oversaturation caused by the instantaneous increase in deviation at the beginning. A larger value should be selected. and smaller In addition, to prevent integral saturation and avoid large overshoot in the system response, The value should be small. When the deviation is small, in order to make the system have better steady-state performance, it should be increased. , The value should be appropriately selected to avoid oscillation of the output response around the set value and to consider the system's anti-interference capability. The principle is that when the rate of change of deviation is small, Take a larger value; when the rate of change of deviation is large, take a smaller value. Based on the above principles, 7*7=49 fuzzy rules can be formulated, and the fuzzy rule tables are shown in Tables 1 to 3: Table 1 Fuzzy rules

[0013] Table 2 Fuzzy rules

[0014] Table 3 Fuzzy rules

[0015] Based on the reactor temperature control rules, fuzzy inference calculations are performed. Fuzzy inference is essentially a fuzzification process, determining the fuzzy quantity corresponding to each precise quantity. The result of fuzzy inference is a fuzzy set, which needs to be defuzzified to obtain a precise value as the control signal to drive the actuator. A weighted average method is used to defuzzify and obtain the corresponding precise value. , , The parameter tuning formula can be expressed as: (1) In the formula, , , Corresponding to , , The initial value. Preferably, step 3) has the following specific steps: The basic principle of the Smith predictor method is to predict the dynamic characteristics of the object and compensate for time lag using a prediction model. The compensator and the controlled object together constitute a controlled object without time lag, as shown in Figure 2. To ensure that the equivalent transfer function no longer has a pure time lag element, a time lag element is connected in parallel in the control algorithm. The compensator enabled the reactor temperature control system model to be developed earlier. At that moment, the pure time delay element in the reactor temperature control system model was eliminated. From this, we can obtain the transfer function as:

[0016]

[0017] The compensated transfer function is: (2) From the above equation, we can conclude that after compensation by the predictor, the stability of the system is no longer affected. The impact. Example

[0018] For the temperature control of a reactor, a self-regulating reactor temperature control system based on an S7 1200 PLC controller and fuzzy Smith-PID was designed. The system uses SCL language to write the fuzzy Smith-PID self-regulating reactor temperature control algorithm. The system consists of a stirring motor, a reactor system, connecting pipes, a signal detection system, a signal processing system, and a display system. A schematic diagram of the system is shown below. Figure 4 As shown.

[0019] The reactor temperature control system mainly consists of hot and cold water valves, a PT100 thermometer, an S7 1200 PLC, and an HMI (Human-Machine Interface). It operates by controlling two valves—a hot water valve and a cooling water valve—as actuators. A mixer agitates the materials to ensure uniformity, improves heat transfer, and maintains a consistent temperature. During the heating phase, the heating water valve is opened to circulate hot water through the coils inside the reactor, raising the reactor temperature. The rate of temperature increase is controlled by adjusting the valve opening (0-100%). Heating stops once the predetermined reaction temperature is reached. Cooling water is circulated through the jacket during the reaction to remove excess heat and maintain a constant temperature. The choice of heat transfer medium depends on the process temperature requirements of each product; common heat transfer media include superheated steam and heat transfer oil. The S7 1200 acts as the processor for online Smith-fuzzy PID calculations, while the touchscreen serves as the HMI for data input, command control, parameter display, and data storage.

[0020] The temperature inside the vessel can be measured by a PT100 sensor, and the signal is transmitted to the PLC through the AI ​​module. After optimization calculation inside the PLC, a control signal is obtained, which is the opening degree adjustment of the hot and cold water valves, thereby achieving temperature regulation.

[0021] The specific implementation of the self-regulating reactor temperature control method based on fuzzy Smith-PID can be obtained through the following control steps: (1) The real-time temperature value is obtained by measuring the PT100, and the real-time temperature difference is calculated based on the difference between the temperature value and the set value. And calculate the rate of change of the real-time temperature difference based on each sampling time. .

[0022] (2) The real-time temperature deviation value and the rate of change of temperature deviation As two inputs to the fuzzy controller, they are compared and calculated in real time using fuzzy rules tables 1, 2, and 3. Then, the corresponding real-time values ​​are obtained by defuzzification using a weighted average method. , , .

[0023] (3) Use formula (1) to obtain the real-time PID parameter values. , , This enables adaptive adjustment and control of the reactor temperature.

[0024] (1) In the formula, , , Corresponding to , , The initial value.

[0025] (4) Predictive compensation processing, as shown in equation (2) and Figure 2 As shown, the transfer function of the controlled object is divided into a linear part and a pure time delay part. By introducing a Smith predictor compensation controller connected in parallel with the controlled object, the pure time delay can be weakened and eliminated, which can be equivalently regarded as reducing the pure time delay part of the controlled object. and its linear part Separate and move it outside the closed-loop system to fundamentally eliminate the effects of pure process time delay; (5) The Smith predictor can be written in the following form

[0026] The method for determining the parameters of the Smith predictor described above is as follows: (1) To keep the system stable, manually apply a step input (such as increasing the valve opening by 5-10%).

[0027] (2) Record the temperature response curve; (3) Extract parameters from the curve, based on: Lag time First, stabilize the system temperature, meaning the temperature no longer changes. The time it takes for the temperature to start changing after the input value (valve opening) is modified is the lag time. As shown in the following formula

[0028] in To modify the timing of the heating valve opening, This represents the time at which the temperature begins to change.

[0029] time constant As shown in the following formula

[0030] in For the reactor temperature to reach At that moment, To modify the initial temperature value when setting the value, This is the temperature value after the temperature stabilizes again.

[0031] Gain As shown in the following formula

[0032] in This is a step input change (such as a 5% increase in valve opening).

[0033] The above description is merely an embodiment of the present invention. It should be noted that, for those skilled in the art, various changes, modifications, substitutions and variations can be made to these embodiments without departing from the technical principles of the present invention. These changes, modifications, substitutions and variations should also be considered within the scope of protection of the present invention.

Claims

1. A self-regulating reactor temperature control method based on fuzzy Smith-PID, characterized in that, Includes the following steps: 1) Considering the large inertia and large time lag characteristics of the controlled object in the self-regulating reactor temperature control system based on fuzzy Smith-PID, a first-order inertial model with an added time delay element is selected to describe the reactor temperature control system model. ,in, For the static gain of the system, Let be the system's time constant. The system's lag time. It is a complex variable; 2) Utilizing PID control algorithm and fuzzy control algorithm, combined with the system's real-time temperature error. and temperature error change rate The PID parameters are updated in real time to obtain the structure diagram of the fuzzy PID controller and realize self-adjusting control. 3) A Smith predictor model for the reactor was obtained through offline testing. This Smith predictor model was then combined with the reactor temperature control system model, and a fuzzy PID controller was used to obtain a time-independent predictor. The impact on the self-regulating reactor temperature control system based on fuzzy Smith-PID.

2. The self-regulating reactor temperature control method based on fuzzy Smith-PID according to claim 1, characterized in that, The specific steps of step 2) are as follows: a sets the temperature deviation to The rate of change of temperature deviation is set to PID parameters are used as inputs to the fuzzy control algorithm. , , Change , , As its output, a fuzzy controller with a two-input, three-output structure is obtained; b. Determine the temperature deviation within the temperature control and allowable error range. Temperature change rate domain of discourse , Discrete points and quantization factors are selected, and based on the determined universe of discourse, discrete points, and quantization factors, a set of... , , Quantization parameters; c. Select the fuzzy set of linguistic variables for input and output variables. For temperature error and temperature error change rate, define the fuzzy subsets on their universe of discourse as: PB (positive large), PM (positive medium), PS (positive small), ZO (zero), NS (negative), NM (negative medium), NB (negative large). The fuzzy subset of the control parameter output is PB (positive large), PM (positive medium), PS (positive small), ZO (zero), NS (negative), NM (negative medium), NB (negative large). d. Set the temperature deviation value and the rate of change of temperature deviation The quantization interval is E and EC. The temperature deviation value is obtained using the membership function. and the rate of change of temperature deviation The membership degrees are calculated over the quantization interval, and the parameters are normalized based on the obtained membership degrees. Then, based on the established fuzzy rule table, corresponding inference calculations are performed to obtain the correction values ​​for the three parameters of the corresponding PID. , , The fuzzy adjustment value; The fuzzy rule table is established as follows: When the deviation is large, in order to speed up the system's response and prevent the control action from exceeding the range due to possible differential oversaturation caused by the instantaneous increase in deviation at the beginning. A larger value should be selected. and smaller In addition, to prevent integral saturation and avoid large overshoot in the system response, The value should be small. When the deviation is small, in order to make the system have better steady-state performance, it should be increased. , The value should be appropriately selected to avoid oscillation of the output response around the set value and to consider the system's anti-interference capability. The principle is that when the rate of change of deviation is small, Take a larger value; when the rate of change of deviation is large, take a smaller value. Based on the above principles, 7*7=49 fuzzy rules can be formulated. Based on the reactor temperature control rules, fuzzy inference calculations are performed. Fuzzy inference is essentially a fuzzification process, determining the fuzzy quantity corresponding to each precise quantity. The result of fuzzy inference is a fuzzy set, which needs to be defuzzified to obtain a precise value as the control signal to drive the actuator. A weighted average method is used to defuzzify and obtain the corresponding precise value. , , The parameter tuning formula can be expressed as: （1）； In the formula, , , Corresponding to , , The initial value.

3. The self-regulating reactor temperature control method based on fuzzy Smith-PID according to claim 1, characterized in that, The specific steps of step 3) are as follows: To ensure that the equivalent transfer function no longer contains a pure time delay element, a parallel connection is used in the control algorithm. The compensator enabled the reactor temperature control system model to be developed earlier. At that moment, the pure time delay element in the reactor temperature control system model was eliminated. From this, we can obtain the transfer function as: ； ； The compensated transfer function is: （2）； From the above equation, we can conclude that after compensation by the predictor, the stability of the system is no longer affected. The impact.

Images