CN-121979339-A - Self-adaptive temperature control method for polyphenylene sulfide polymerization kettle

CN121979339ACN 121979339 ACN121979339 ACN 121979339ACN-121979339-A

Abstract

The invention relates to the technical field of process control, in particular to a self-adaptive temperature control method of a polyphenylene sulfide polymerization kettle, which comprises the steps of collecting melt viscosity, heat release power and residence time deviation of the polymerization kettle in real time and constructing a normalized multidimensional time sequence input vector by utilizing a sliding window technology. And then, analyzing and outputting self-adaptive weight coefficients aiming at different deviation dimensions in real time by combining a weight control model of a long-term memory network and an attention mechanism and combining static working condition characteristics. Meanwhile, based on the shearing heat effect and the reaction accumulation effect, a specific temperature compensation component for the inlet of the heat medium jacket is calculated, and independent amplitude limiting processing is performed on each component. And finally, the target set temperature is weighted and synthesized, and the precise action of the executing mechanism is driven by using the cascade PID control and the valve flow characteristic inverse mapping model. The invention realizes the precise steady-state control of the polymerization reaction through deep fusion and mechanism compensation.

Inventors

FAN HONGJIE
YUE XILONG
XIAO SHIJIAN
LIU XIN
SHEN SIYU
Xu Wuting

Assignees

扬州循曜聚合科技有限公司

Dates

Publication Date: 20260505
Application Date: 20251229

Claims (8)

1. The self-adaptive temperature control method for the polyphenylene sulfide polymerization kettle is characterized by comprising the following steps of: Collecting melt viscosity, liquid level height, jacket inlet temperature, jacket outlet temperature and heat medium circulation flow of a polymerization kettle; calculating the viscosity deviation of the real-time melt viscosity and the target viscosity, calculating the real-time heat release power according to the jacket inlet temperature, the jacket outlet temperature and the heat medium circulation flow, obtaining the power deviation degree relative to the standard heat release power, and calculating the actual residence time deviation according to the change rate of the real-time liquid level height; The method comprises the steps of generating a first temperature compensation value, a second temperature compensation value and a third temperature compensation value based on viscosity deviation, power deviation and actual residence time deviation respectively; Inputting the multidimensional time sequence sliding window input vector into a weight control model, and outputting a first weight coefficient, a second weight coefficient and a third weight coefficient; And weighting and summing the first temperature compensation value, the second temperature compensation value and the third temperature compensation value by using the first weight coefficient, the second weight coefficient and the third weight coefficient, adding the calculation result to the reference temperature to obtain a final target set temperature, and adjusting the opening of the heating medium regulating valve according to the final target set temperature.
2. The method for adaptively controlling the temperature of a polyphenylene sulfide polymerization kettle according to claim 1, wherein the steps of collecting the melt viscosity, the liquid level height, the jacket inlet temperature, the jacket outlet temperature and the heat medium circulation flow rate of the polymerization kettle comprise: The method comprises the steps of detecting vibration damping signals of a probe in a reaction material by using an online torsional vibration type viscometer, mapping and converting the vibration damping signals into melt viscosity by using a preset viscosity-damping corresponding relation table, collecting real-time pressure difference values between a bottom flange of a polymerization kettle and a gas phase space by using a double-flange differential pressure transmitter, dividing the real-time pressure difference values by the product of preset melt density and gravity acceleration based on a hydrostatic principle, converting the products to obtain the liquid level height, collecting resistance value signals changing along with temperature by using an armoured platinum resistance temperature sensor, converting the collected resistance value signals into jacket inlet temperature and jacket outlet temperature according to a resistance-temperature characteristic curve of platinum resistance, detecting phase difference signals generated by a measuring tube due to fluid flow by using a mass flowmeter, and obtaining the heat medium circulation flow by linear proportion conversion processing of the phase difference signals.
3. The self-adaptive control method for the temperature of the polyphenylene sulfide polymerization kettle according to claim 1 is characterized in that the calculation process of the viscosity deviation, the power deviation degree and the actual residence time deviation comprises the steps of performing numerical difference operation on the real-time melt viscosity and the target viscosity to obtain an absolute viscosity difference with positive and negative signs, dividing the absolute viscosity difference by the target viscosity to obtain the normalized viscosity deviation, reading a pre-stored heat capacity parameter of a heating medium, calculating a temperature difference value between the jacket outlet temperature and the jacket inlet temperature, calculating a product of the temperature difference value, the heating medium circulation flow and the heating medium heat capacity to obtain real-time heat release power at the current moment, comparing the real-time heat release power with a central value of the standard heat release power, calculating a difference ratio between the real-time heat release power and the standard heat release power as the power deviation degree, performing first-order difference operation on the continuously collected real-time liquid level height, calculating a liquid level height difference between the current sampling moment and the last sampling moment, dividing the liquid level height difference by a sampling period to obtain a liquid level lifting rate in unit time, and mapping the liquid level lifting rate to the fluctuation time as the actual residence time deviation value according to the effective volume parameter of the polymerization kettle.
4. The self-adaptive temperature control method for the polyphenylene sulfide polymerization kettle according to claim 1, wherein the process of generating a first temperature compensation value, a second temperature compensation value and a third temperature compensation value comprises the steps of obtaining a preset coefficient required for calculation, reading a preset viscosity proportionality coefficient, a power proportionality coefficient, a time proportionality coefficient and a preset shear thermal coupling coefficient and reaction accumulation coupling coefficient from a controller storage unit, multiplying the viscosity deviation and the viscosity proportionality coefficient to obtain a viscosity basic component only responding to viscosity change, multiplying the power deviation degree and the power proportionality coefficient to obtain a power basic component only responding to exothermic change, multiplying the actual residence time deviation and the time proportionality coefficient to obtain a time basic component only responding to time change, calculating a product of the viscosity deviation and the power deviation degree, multiplying the product and the shear thermal coupling coefficient to obtain a shear thermal correction term for compensating high-viscosity friction heat generation, calculating a product of the power deviation degree and the actual residence time deviation, multiplying the product and the reaction accumulation coupling coefficient to obtain a viscosity basic component only responding to viscosity change, and obtaining a second temperature basic component for compensation value, and subtracting a temperature basic component from a first temperature basic component for compensation value, and directly obtaining a second temperature basic component for compensation value.
5. The self-adaptive temperature control method for the polyphenylene sulfide polymerization kettle according to claim 1, wherein the process of constructing the input vector of the multi-dimensional time sequence sliding window comprises the steps of setting a time depth parameter and a sampling step size parameter of the sliding window; the method comprises the steps of creating a first-in first-out queue, a second first-in first-out queue and a third first-in first-out queue in a memory, respectively storing historical sequence data of viscosity deviation, power deviation degree and actual residence time deviation, sequentially pressing data of current time and past continuous N sampling time into the corresponding first-in first-out queues according to the sampling step length, removing earliest data, maintaining the length of the queues to be fixed, reading preset physical limit amplitude values of each parameter, respectively executing maximum and minimum normalization operation on all data in the first, second and third first-in first-out queues, mapping all values into a zero-to-one dimensionless interval, extracting normalized first-in first-out queue data as a viscosity characteristic dimension, extracting normalized second-in first-out queue data as a power characteristic dimension, extracting normalized third-in first-out queue data as a time characteristic dimension, splicing the viscosity characteristic dimension, the power characteristic dimension and the time characteristic dimension according to a preset channel sequence, and generating a multi-dimensional parallel input window by means of splicing.
6. The method for adaptively controlling the temperature of a polyphenylene sulfide polymerization kettle according to claim 1, wherein the weight control model comprises: The dynamic trend encoder receives the input vector of the multi-dimensional time sequence sliding window as input, processes the input vector by adopting a long-period memory network, eliminates short-period random noise by an internal forgetting gate and input gate mechanism, captures the historical evolution rules of viscosity deviation, power deviation degree and residence time deviation, and generates a dynamic trend feature vector containing reaction dynamic time-varying information; the static working condition encoder receives preset reference temperature, target viscosity and standard heat release power as inputs, adopts a fully connected embedded network for processing, maps process setting parameters of a physical layer into high-dimensional sparse semantic vectors, and generates a static working condition feature vector representing the characteristics of the current polymerization reaction stage; The multisource attention fusion unit is used for splicing the static working condition feature vector and the dynamic trend feature vector, inputting the spliced composite state vector into an attention generation network, generating attention weight vectors aiming at different reaction parameters, applying the attention weight vectors to the dynamic trend feature vector output by the dynamic trend encoder, dynamically weighting the dynamic trend feature vector through a feature recalibration mechanism and outputting global context perception features; and the self-adaptive decision output layer receives the global context perception feature as input, adopts a multi-layer perception machine to carry out dimension reduction and linear combination of high-dimensional features, maps an operation result to a numerical value interval from 0 to 1 through an S-shaped nonlinear activation function, and finally outputs a first weight coefficient, a second weight coefficient and a third weight coefficient in parallel.
7. The self-adaptive control method for the temperature of the polyphenylene sulfide polymerization kettle according to claim 1 is characterized in that the obtaining process of the final target set temperature comprises the steps of calculating the product of a first weight coefficient and a first temperature compensation value, the product of a second weight coefficient and a second temperature compensation value and the product of a third weight coefficient and a third temperature compensation value respectively, adding the three products to obtain a total dynamic compensation amount, algebraically superposing the total dynamic compensation amount and the reference temperature to generate an original target set temperature, reading a preset highest allowable temperature and a preset lowest maintenance temperature of the polymerization process, judging whether the original target set temperature exceeds a safety interval formed by the highest allowable temperature and the lowest maintenance temperature, forcedly clamping the original target set temperature to a corresponding boundary value if the original target set temperature exceeds the safety interval, calculating the difference between the original target set temperature of a current calculation period and the final target set temperature of a previous period, carrying out peak clipping treatment on the original target set temperature if the absolute value is larger than a preset maximum rate threshold, and outputting the obtained difference as the final target set temperature after the final rate smoothing.
8. The self-adaptive control method for the temperature of the polyphenylene sulfide polymerization kettle according to claim 1 is characterized by comprising the steps of taking the final target set temperature as a set value of a main control loop, taking the collected real-time temperature of the polymerization kettle as a feedback value of the main control loop, calculating to obtain a jacket temperature set value through a main PID controller, taking the jacket temperature set value as a set value of a secondary control loop, taking the collected jacket inlet temperature as a feedback value of the secondary control loop, calculating to obtain a standard control output signal through the secondary PID controller, taking a preset heat medium regulating valve flow characteristic curve, wherein the characteristic curve represents a nonlinear relation between a valve opening and a heat medium flow, reversely mapping the standard control output signal by utilizing an inverse function model of the flow characteristic curve, calculating a corrected linearization valve opening command, converting the linearization valve opening command into a current signal, driving an actuating mechanism of the heat medium regulating valve to act, and enabling a valve core position to reach a corresponding opening.

Description

Self-adaptive temperature control method for polyphenylene sulfide polymerization kettle Technical Field The invention relates to the technical field of process control, in particular to a self-adaptive temperature control method for a polyphenylene sulfide polymerization kettle. Background Continuous polymerization systems, particularly those used to produce high performance polymers such as polyphenylene sulfide, typically employ multiple polymerization vessels operated in series. Since the polymerization process itself is typically a highly exothermic, high viscosity, large hysteresis complex system, the quality index (e.g., melt viscosity) of the reaction product is extremely sensitive to small temperature fluctuations. Therefore, in the whole continuous production process, accurate, stable and robust temperature control is realized, and the key for guaranteeing the quality qualification rate of products and the production operation safety is provided. In the prior art, the temperature control of the polymerizer is mainly based on the traditional PID (proportional-integral-derivative) control method or a cascade control strategy extended on the basis of the traditional PID (proportional-integral-derivative) control method. The methods collect real-time temperature of materials in the kettle through the sensor and generate a control signal according to deviation between the temperature and a set value, so that circulation flow of heating medium (such as heat conducting oil) in the temperature adjusting jacket is adjusted. This control scheme relies on preset fixed parameters and works well when the system is operating in steady state. However, in a practically complex industrial production environment, such a control strategy based on fixed parameters faces a number of challenges. First, strong coupling within the polymerization system is a core challenge. For example, the change of the melt viscosity of the material can directly influence the shearing heat (i.e. internal heat generation) generated in the stirring process so as to change the actual exothermic power of the reaction, and meanwhile, the fluctuation of the liquid level of the reaction material can influence the actual residence time of the material in the kettle, so that the reaction depth and the instantaneous heating value at the current moment are directly related. Conventional control loops have difficulty effectively decoupling these nonlinear and strong parameter interactions. Secondly, the existing system has insufficient self-adaptive capacity and generalization performance for dynamic working conditions. In the continuous operation process, factors such as small fluctuation of raw material components, attenuation of catalyst activity, scale formation of a kettle wall and the like can cause slow drift of the optimal operation condition point of the system. The conventional control algorithm lacks a set of effective mechanisms to recognize these operating condition changes and dynamically adjusts its control gain or model parameters according to the new operating context, resulting in a decrease in control accuracy as the operating time is extended. Finally, existing control systems have limitations in the use of multidimensional information. While a large amount of multi-source and high-value data such as melt viscosity, inlet and outlet temperatures of a heating medium jacket, circulation flow of the heating medium and the like can be acquired in real time in the production process, the prior art mainly uses only a few parameters such as temperature, liquid level and the like as feedback signals. This makes the system lacking the ability to efficiently and deeply fuse and translate these heterogeneous multidimensional information into high-precision, adaptive control decisions. The core problem faced by the temperature control technology of the existing continuous polymerization reaction system is how to establish a control method capable of fully utilizing multi-source and multi-dimensional real-time operation data and effectively aiming at strong coupling of reaction material parameters and dynamic change of operation conditions, thereby realizing high-precision self-adaptive adjustment of the temperature of a polymerization kettle. Therefore, a self-adaptive temperature control method for a polyphenylene sulfide polymerization kettle is provided. Disclosure of Invention The invention aims to provide a self-adaptive control method for the temperature of a polyphenylene sulfide polymerization kettle, which realizes precise steady-state control of polymerization reaction through deep fusion and mechanism compensation. The method comprises the steps of collecting melt viscosity, exothermic power and residence time deviation of a polymerization kettle in real time, and constructing a normalized multidimensional time sequence input vector by utilizing a sliding window technology. And then, analyzing and outputting self-adaptive weight