A kind of polyphosphate potassium short chain degree of polymerization synthesis control system and method
By acquiring real-time status and optimizing control through predictive models, temperature, flow rate, and rotation speed are dynamically adjusted, solving the multivariate coupling and safety issues in the synthesis of short-chain potassium polyphosphate, and achieving precise control and safety assurance of the degree of polymerization.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHIFANG JINDIYAMEI CHEM ENG
- Filing Date
- 2026-05-06
- Publication Date
- 2026-06-05
AI Technical Summary
Existing short-chain potassium polyphosphate synthesis processes suffer from problems such as strong coupling of multiple variables, frequent operating condition disturbances, and strict safety constraints, leading to chain length distribution shifts, increased side reactions, and high risks of safety accidents.
By employing real-time status acquisition, predictive model aggregation degree prediction, adaptive weight optimization control, risk scoring decision-making, and model updating, multivariate decoupled control is achieved through dynamic adjustment of temperature, flow rate, and rotation speed. A loss function is constructed to update model parameters, and risk factors are identified to implement safety control.
This achieved a high degree of aggregation that closely approximates the target value, improving product consistency, reducing the accident rate, and ensuring production stability and safety.
Smart Images
Figure CN122151552A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of potassium polyphosphate synthesis technology, specifically to a control system and method for the synthesis of potassium polyphosphate short-chain polymerization degree. Background Technology
[0002] Potassium polyphosphate, as an important phosphate product, is widely used in food preservation, agricultural fertilizers, and industrial flame retardants. Its molecular chain length (characterized by the degree of polymerization) directly determines the product performance. Short-chain potassium polyphosphate (degree of polymerization 3-5) has excellent water solubility and bioactivity, making it a core raw material for high-end food additives and high-efficiency water-soluble fertilizers.
[0003] In existing technologies, the synthesis of short-chain potassium polyphosphate typically employs a batch reactor process. The target degree of polymerization is achieved by controlling the condensation reaction conditions (temperature, concentration, pH, reaction time, etc.) of phosphoric acid and potassium hydroxide. However, this process suffers from the following technical problems: First, strong coupling of multiple variables: parameters such as temperature, concentration, and pH influence each other, and adjusting a single parameter can easily lead to chain length distribution shifts or exacerbated side reactions. Second, frequent operational disturbances: fluctuations in raw material purity, uneven heat transfer, and equipment aging cause the reaction trajectory to deviate from the design value, resulting in poor product consistency. Finally, strict safety constraints: under high-temperature or high-concentration conditions, localized overheating or uncontrolled concentration can easily lead to safety accidents such as coking and explosive polymerization. Summary of the Invention
[0004] To address the above problems, the first aspect of this invention provides a method for controlling the degree of polymerization of short-chain potassium polyphosphate, comprising the following steps: Obtain real-time synthesis status; Based on the real-time synthesis status, the degree of aggregation is predicted using a prediction model; Based on the predicted degree of aggregation, a first degree of aggregation threshold, and a second degree of aggregation threshold, control decisions are determined. In response to adjustment control, the optimal control sequence is solved based on the objective function and constraints. Execute the first control input in the optimal control sequence; Risk scoring and decision-making are based on the composite state after the execution of the first step control variable; In response to the model update decision, update the parameters of the prediction model; Respond to security protection decisions and execute security protocols.
[0005] Preferably, the prediction model is: ; in, To predict the degree of aggregation, pH level The reaction temperature, This refers to the phosphate concentration. , , , and These are the model parameters.
[0006] Preferably, the determination of the control execution mode includes: When the predicted aggregation degree is greater than or equal to the first aggregation degree threshold and less than or equal to the second aggregation degree threshold, maintain the current control input; When the predicted aggregation degree is less than the first aggregation degree threshold, adjust the control input to the first control vector; When the predicted aggregation degree is greater than the second aggregation degree threshold, adjust the control input to the second control vector; Wherein, the first control vector is: ; in, The first control vector, For temperature control gain, For flow control gain, For speed control gain; In some embodiments, the second control vector is: ; in, This is the second control vector. For temperature control gain, For flow control gain.
[0007] Preferably, the objective function is: ; in, Let be the objective function. For temperature adjustment amount, Adjust the amount of flow. This is the speed adjustment amount. To control input Predicted aggregation degree , , These are adaptive weighting coefficients; The constraints include: Adjusting rate constraints: , , ; Safety range constraints: , , ; The adaptive weights are: ; in, Temperature weighting, To predict the degree of aggregation; ; in, For traffic weight, This refers to the phosphate concentration. ; in, The rotational speed weight is used.
[0008] Preferably, the risk scoring decision includes: Calculate the comprehensive risk score based on the composite state after the execution of the first step control quantity; The aggregation degree prediction error is calculated based on the actual aggregation degree and the predicted aggregation degree. The decision-making logic is determined based on comprehensive risk scoring and aggregation degree prediction error calculation.
[0009] Preferably, the comprehensive risk score is calculated as follows: ; in, For comprehensive risk scoring, Temperature risk weighting, For temperature risk scoring, For concentration risk weights, For concentration risk scoring, For aggregation degree risk weight, For aggregation degree risk scoring, As the weight for process parameter deviations, Scoring is given for deviations in process parameters; The temperature risk score is as follows: ; in, For temperature risk scoring, This is the highest temperature inside the reactor; The concentration risk score is as follows: ; in, For concentration risk scoring, This represents the maximum concentration of phosphate. The aggregation degree risk score is: ; in, For aggregation degree risk scoring, To predict the degree of aggregation; The process parameter deviation score is: ; in, Scoring for process parameter deviations for Adjustment amount, Adjust the amount of flow. This is the speed adjustment amount.
[0010] Preferably, the method for calculating the degree of aggregation prediction error is as follows: ; in, For the degree of aggregation prediction error, This represents the actual degree of aggregation. To predict the degree of aggregation; The decision-making logic is as follows: ; in, Indicates the decision result, This indicates that control will continue for the next cycle. This indicates that a model update has been triggered. This indicates that security protection has been triggered. As a risk threshold, The aggregation degree prediction error threshold, For the degree of aggregation prediction error, For comprehensive risk scoring.
[0011] Preferably, the updated prediction model parameters include: Update the prediction model parameters based on parameter update rules; The parameter update rule is as follows: ; in, To update the model parameters, These are the current model parameters. For learning rate, This represents the gradient of the loss function with respect to the parameters. The gradient is: ; in, The gradient of the loss function with respect to the parameters. To predict the degree of aggregation, This represents the actual degree of aggregation. Output the partial derivatives of the PLS model with respect to the model parameters. The regularization coefficient is . These are the parameters of the current model.
[0012] Preferably, the security protocol includes: Targeted control strategies are determined based on the identification of risk-dominant factors; Apply safety control measures to the response system to reduce the overall risk score below the threshold; The dominant risk factors were identified as follows: ; in, Index of risk-dominant factors Risk weighting Assess risk level; The targeted control strategy is as follows: ; in, For safety control purposes, For temperature control gain, This is the highest temperature inside the reactor. To control the gain by concentration, This represents the maximum concentration of phosphate. For symbolic functions, Adjust the traffic limit. It is a quadratic programming solver. This represents the actual degree of aggregation. To control the gain for process parameters, for Adjustment amount, Adjust the amount of flow. This is the speed adjustment amount. This indicates that temperature risk is the primary factor. This indicates that concentration risk is the primary factor. This indicates that aggregation degree risk is dominant. This indicates that process parameter deviations are the primary factor.
[0013] A second aspect of the present invention provides a potassium polyphosphate short-chain polymerization degree synthesis control system for executing any of the above-described methods for controlling the polymerization degree of potassium polyphosphate short-chain polymerization, comprising: The acquisition module is used to acquire the real-time synthesis status; The prediction module is used to predict the degree of aggregation based on the real-time synthesis status; The determination module is used to determine control decisions based on the predicted degree of aggregation, a first degree of aggregation threshold, and a second degree of aggregation threshold. The solver module is used to solve for the optimal control sequence based on the objective function and constraints. The first execution module is used to execute the first control input in the optimal control sequence; The scoring module is used to make risk scoring decisions based on the composite state after the execution of the first step control quantity; The update module is used to update the parameters of the prediction model; The second execution module is used to execute the security protocol.
[0014] By adopting the above technical solution, the present invention mainly has the following technical effects: 1. By setting adaptive weights, the optimization priorities of temperature / flow rate / speed are dynamically adjusted, and the optimal combination of ∆T, ∆V, and ∆R is solved through rolling optimization. This solves the technical problem that strong coupling of multiple parameters such as temperature, concentration, and pH can easily lead to chain length distribution shifts or aggravated side reactions due to single adjustments. It achieves the technical effect of multi-variable collaborative decoupling control, making the degree of polymerization accurately approach the target value.
[0015] 2. By constructing a loss function and updating model parameters based on the gradient descent algorithm, the prediction bias is iteratively corrected until convergence. This solves the technical problems of raw material purity fluctuations, uneven heat transfer, and equipment aging causing the reaction trajectory to deviate from the design value and the product consistency to be poor. It achieves the technical effects of adaptive model updating and real-time correction.
[0016] 3. By calculating risk scores and identifying risk-inducing factors, targeted safety control strategies are triggered until the comprehensive risk score is reduced to below the threshold. This solves the technical problem that local overheating or concentration runaway under high temperature or high concentration conditions can easily lead to safety accidents such as coking and explosive agglomeration. Through multi-dimensional real-time risk warning and graded safety handling, the technical effect of reducing the accident rate is achieved. Attached Figure Description
[0017] Figure 1 This is a flowchart of a method for controlling the degree of polymerization of short-chain potassium polyphosphate according to the present invention; Figure 2 To obtain the circuit diagram of the pH detection submodule of the module; Figure 3 To obtain the circuit diagram of the temperature detection submodule of the module; Figure 4 This is the circuit diagram of the motor speed regulation submodule in the execution module. Detailed Implementation
[0018] To enable those skilled in the art to better understand the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. 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.
[0019] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0020] Please see Figure 1 The first aspect of this invention provides a method for controlling the degree of polymerization of short-chain potassium polyphosphate, comprising the following steps: S1. Obtain the real-time synthesis status; In some embodiments, the synthesis information may be a combination of reaction parameters reflecting the synthesis process of potassium polyphosphate. In some embodiments, the synthesis information includes: pH value, which is used to reflect the pH value of the reaction system, and exemplary pH value can be obtained by pH electrode detection; The reaction temperature is used to reflect the real-time temperature inside the reactor. An exemplary reaction temperature can be obtained by a temperature sensor. Phosphate concentration, which reflects the content of phosphate in the reaction solution, can be obtained by ion chromatography or near-infrared spectroscopy.
[0021] S2. Based on the real-time synthesis status, the degree of aggregation is predicted using a prediction model; In some embodiments, the prediction model can be any computational model capable of predicting the degree of aggregation based on reaction environment parameters. In some embodiments, the prediction model is a multiple linear regression prediction model.
[0022] In some embodiments, the prediction model is: ; in, To predict the degree of aggregation, pH level The reaction temperature, This refers to the phosphate concentration. , , , and These are model parameters; In some embodiments, the model parameters represent a set of weights learned from historical experimental data; in other embodiments, the prediction model can be trained. For example, a training sample set can be constructed based on historical reaction data, the training samples including input features such as pH. Reaction temperature and phosphate concentration and the corresponding true aggregation degree Input the input features into the prediction model to obtain the prediction output. Then, based on the true aggregation degree labels of the training samples and the predicted aggregation degree output by the prediction model, a loss function is constructed, and the optimal model parameters are solved by the least squares method or gradient descent algorithm; iteration continues until convergence, and the trained prediction model is obtained.
[0023] In some embodiments, the units in the formula can be kept consistent by setting appropriate units for the model parameters.
[0024] S3. Based on the predicted degree of aggregation, the first degree of aggregation threshold, and the second degree of aggregation threshold, determine the control decision; In some embodiments, the aggregation degree threshold is used to determine whether the predicted aggregation degree is within the target range and to determine the corresponding control execution method.
[0025] In some embodiments, the control execution method can be determined by comparing the predicted aggregation degree, a first aggregation degree threshold, and a second aggregation degree threshold.
[0026] In some embodiments, determining the control execution method includes: When the predicted aggregation degree is greater than or equal to the first aggregation degree threshold and less than or equal to the second aggregation degree threshold, maintain the current control input; When the predicted aggregation degree is less than the first aggregation degree threshold, adjust the control input to the first control vector; When the predicted aggregation degree is greater than the second aggregation degree threshold, adjust the control input to the second control vector; In some embodiments, the first control vector is: ; in, The first control vector, For temperature control gain, For flow control gain, For speed control gain; In some embodiments, the second control vector is: ; in, This is the second control vector. For temperature control gain, For flow control gain; In some embodiments, the control gain is used to represent a fixed adjustment range of the control action. Indicates positive gain. This indicates negative gain.
[0027] As an example, let the control gain be: , , The first aggregation threshold is 3, and the second aggregation threshold is 5. Input real-time parameters (such as...) into the prediction model , , ), to obtain the predicted aggregation degree output by the prediction model. ; like If the condition is met, maintain the current control input. like The degree of polymerization was determined to be too low, and the temperature was increased. Reduce the phosphate dripping flow rate And reduce the speed The control vector is used to promote chain growth; like The polymerization degree was determined to be too high, and cooling was initiated. Increase the phosphate dripping flow rate The control vector is used to suppress the condensation reaction.
[0028] S4. In response to adjustment control, solve for the optimal control sequence based on the objective function and constraints; In some embodiments, the optimal control sequence refers to solving a constrained optimization problem to make the degree of aggregation approach the target value while suppressing control fluctuations; the objective function is used to measure the weighted sum of the degree of aggregation deviation and the control cost, and the constraints are used to ensure the adjustment rate and process safety.
[0029] In some embodiments, the objective function is: ; in, Let be the objective function. For temperature adjustment amount, Adjust the amount of flow. This is the speed adjustment amount. To control input Predicted aggregation degree , , These are adaptive weighting coefficients.
[0030] In some embodiments, the constraints include: Adjusting rate constraints: , , ; Safety range constraints: , , ; In some embodiments, the adaptive weights are used to dynamically adjust the optimization priority of the control variables; In some embodiments, the adaptive weights are: ; in, Temperature weighting, To predict the degree of polymerization, the greater the deviation in the degree of polymerization, the higher the temperature weighting index becomes, and the temperature is adjusted preferentially.
[0031] ; in, For traffic weight, The concentration is phosphate; when the concentration deviates from 45%, the flow rate weight increases linearly, and the flow rate is adjusted preferentially.
[0032] It should be noted that in the above calculations, the phosphate concentration is a percentage value. For example, 45 means 45%, not 0.45.
[0033] ; in, Weighted by rotational speed; In some embodiments, the rotational speed weight is a fixed value to ensure that all three weights are positive and of coordinated magnitude.
[0034] In some embodiments, after determining the objective function and constraints, the optimal adjustment amount can be calculated using a quadratic programming solver (such as OSQP or CVXOPT). , , The initial control input is applied to the reaction system to achieve rolling optimization control.
[0035] As an example, let the initial state be: , , The current control is: , ; The predicted aggregation degree is: The aggregation degree was determined to be too high, triggering adjustment control. One optimization: Adaptive weight calculation: ; ; ; Objective function: ; Solving for the given information, we get: ; Execution control: New temperature: , New traffic: , New speed: ; Re-forecast: ; Secondary optimization: Adaptive weight update: ; ; ; Objective function:
[0036] Solving for the given information, we get: ; Execution control: New temperature: , New traffic: , New speed: ; … S5. Execute the first control input in the optimal control sequence; In some embodiments, this step is used to apply the first control quantity in the optimized control sequence to the reaction system to realize the execution link of rolling time-domain control.
[0037] As an example, the result of a single optimization solution is applied to the reaction system, where the initial control variable is: ; in, This is a temperature regulation amount, indicating cooling. ; The flow rate adjustment indicates an increase in the phosphate dripping flow rate. .
[0038] Apply the initial control input to the reaction system and update the control state: , , ; S6. Make risk scoring decisions based on the synthetic state after the execution of the first step control quantity; In some embodiments, the synthesis state after the execution of the first step control quantity refers to the state of the reaction system after the execution of the first step control in step S5, including temperature, concentration, degree of polymerization and process parameters.
[0039] In some embodiments, the degree of polymerization includes both predicted and actual degree of polymerization. In some embodiments, the degree of polymerization reflects the average length of the potassium polyphosphate molecular chain, and the actual degree of polymerization can be predicted by online spectral analysis combined with a PLS model. As an example, the online spectral analysis includes: installing a Kaiser optical system probe on the sidewall of the reactor, acquiring the spectrum of the reaction liquid at 10-30 second intervals, and correlating the characteristic peaks of the POP bond with offline ³¹P-NMR calibration values using a partial least squares regression model to achieve online soft measurement of the degree of polymerization.
[0040] In some embodiments, the risk scoring decision is used to comprehensively assess the multi-dimensional risks of the response system and determine subsequent operations based on the assessment results.
[0041] In some embodiments, the risk scoring decision includes: S601. Calculate the comprehensive risk score based on the synthetic state after the execution of the first step control quantity; In some embodiments, the comprehensive risk score is calculated as follows: ; in, For comprehensive risk scoring, Temperature risk weighting, For temperature risk scoring, For concentration risk weights, For concentration risk scoring, For aggregation degree risk weight, For aggregation degree risk scoring, As the weight for process parameter deviations, Scoring is given for deviations in process parameters; In some embodiments, the temperature risk score is based on the highest temperature. In some embodiments, the highest temperature refers to the highest temperature inside the reactor. In some embodiments, the temperature risk score is: ; in, For temperature risk scoring, This is the highest temperature inside the reactor; In some embodiments, the concentration risk score is based on the maximum concentration. In some embodiments, the maximum concentration refers to the maximum concentration of phosphate. In some embodiments, the concentration risk score is: ; in, For concentration risk scoring, This represents the maximum concentration of phosphate. It should be noted that in the above calculation process, the maximum phosphate concentration is a percentage value. For example, 45 means 45%, not 0.45.
[0042] In some embodiments, the aggregation degree risk score is calculated based on the predicted aggregation degree. In some embodiments, the aggregation degree risk score is: ; in, For aggregation degree risk scoring, To predict the degree of aggregation; In some embodiments, the process parameter deviation score is calculated based on the adjustment control quantity; in some embodiments, the process parameter deviation score is: ; in, Scoring for process parameter deviations for Adjustment amount, Adjust the amount of flow. This is the speed adjustment amount; In some embodiments, Temperature risk weighting, For concentration risk weights, For aggregation degree risk weight, The weights for process parameter deviations can be obtained by pre-setting them based on historical data.
[0043] As an example, let's say , , , , , ; , , , ; ; ; ; ; ; S602. Calculate the aggregation degree prediction error based on the actual aggregation degree and the predicted aggregation degree. In some embodiments, the actual degree of polymerization is used to obtain the true state of the reaction in order to correct the prediction model.
[0044] In some embodiments, the aggregation degree prediction error is calculated as follows: ; in, For the degree of aggregation prediction error, This represents the actual degree of aggregation. To predict the degree of aggregation.
[0045] S603. Decision logic is determined based on comprehensive risk score and aggregation degree prediction error calculation. In some embodiments, the decision logic can be confirmed by comparing the comprehensive risk score with the risk threshold, and the aggregation degree prediction error with the aggregation degree prediction error threshold.
[0046] In some embodiments, the risk threshold refers to a critical value that reflects the overall safety status of the system and is used to determine whether a safety protection protocol is triggered. In some embodiments, the aggregation degree prediction error threshold refers to the maximum allowable deviation between the model prediction value and the actual value and is used to determine whether the model needs to be updated online.
[0047] In some embodiments, the decision logic is as follows: ; in, Indicates the decision result, This indicates that control will continue for the next cycle. This indicates that a model update has been triggered. This indicates that security protection has been triggered. As a risk threshold, The aggregation degree prediction error threshold, For the degree of aggregation prediction error, For comprehensive risk scoring.
[0048] In some embodiments, the risk threshold and aggregation degree prediction error threshold can be calibrated based on historical data.
[0049] As an example, the risk threshold is: The aggregation degree prediction error threshold is ; like , ,but: , If the risk is deemed acceptable and the model accuracy is satisfactory, the next optimization cycle will continue. like , ,but: , If the risk is deemed acceptable but the model is mismatched, a model update is triggered.
[0050] like , ,but: , The system was deemed high-risk, triggering safety protection mechanisms.
[0051] S7. In response to the model update decision, update the prediction model parameters; In some embodiments, the model parameter update is used to correct the prediction model when a model update is triggered, thereby improving the accuracy of subsequent predictions.
[0052] In some embodiments, the updated prediction model parameters include: S701. Update the prediction model parameters based on the parameter update rules; In some embodiments, the model parameter update represents an iterative optimization process that dynamically adjusts the weights of the prediction model from real-time response data. In some embodiments, the model parameter update can be implemented through online gradient descent.
[0053] In some embodiments, the parameter update rule is as follows: ; in, To update the model parameters, These are the current model parameters. For learning rate, This represents the gradient of the loss function with respect to the parameters. In some embodiments, the gradient is used to guide the direction and step size of model parameter updates.
[0054] In some embodiments, the gradient is: ; in, The gradient of the loss function with respect to the parameters. To predict the degree of aggregation, This represents the actual degree of aggregation. Output the partial derivatives of the PLS model with respect to the model parameters. The regularization coefficient is . These are the current model parameters; In some embodiments, the gradient of the loss function with respect to the model parameters is calculated using the gradient descent algorithm, and the model parameters are updated along the negative gradient direction according to the parameter update rule. This process is iteratively repeated until the model error is reached. Convergence to threshold The updated prediction model parameters are obtained below.
[0055] As an example, let the learning rate be... The regularization coefficient The partial derivatives of the PLS model output with respect to the model parameters The model error threshold ; If a certain model parameter is currently Predicting the degree of aggregation Actual degree of aggregation Model error , This triggers a model update; Calculate the gradient: ; Update model parameters: ; After the update, the model error is reduced, and the iteration continues until convergence. Once the model parameter update is complete, the updated prediction model parameters are obtained.
[0056] S8. Respond to security protection decisions and execute security protocols; In some embodiments, the safety protocol is used to rapidly reduce the risk of the reaction system under high-risk operating conditions and ensure process safety.
[0057] In some embodiments, the execution security protocol includes: S801. Determine targeted control strategies based on the identification of risk-dominant factors; In some embodiments, the risk-dominant factor identification is used to locate the dimension that contributes the most to the overall risk score and activate the corresponding safety control law.
[0058] In some embodiments, the risk-dominant factor is identified as: ; in, Index of risk-dominant factors Risk weighting Assess risk level; In some embodiments, the targeted control strategy is used to implement differentiated security controls based on the type of risk-dominant factor.
[0059] In some embodiments, the targeted control strategy is: ; in, For safety control purposes, For temperature control gain, This is the highest temperature inside the reactor. To control the gain by concentration, This represents the maximum concentration of phosphate. For symbolic functions, Adjust the traffic limit. It is a quadratic programming solver. This represents the actual degree of aggregation. To control the gain for process parameters, for Adjustment amount, Adjust the amount of flow. This is the speed adjustment amount. This indicates that temperature risk is the primary factor. This indicates that concentration risk is the primary factor. This indicates that aggregation degree risk is dominant. This indicates that process parameter deviations are the primary factor. In some embodiments, the temperature control gain, concentration control gain, process parameter control gain, and flow rate adjustment upper limit can be calibrated based on historical data.
[0060] As an example, according to step S602: , , , ; , , , ; but: ; ; ; ; By comparison: ,correspond ; Determine temperature risk as the primary factor and implement targeted temperature control strategies; like , , , ; but: ; ; ; ; By comparison: ,correspond ; The aggregation degree risk is determined to be the dominant factor, and a quadratic programming approach is used to solve for the optimal control quantity.
[0061] As an example, let the temperature control gain be... The concentration control gain The process parameters control the gain. The upper limit of the flow adjustment ; like , ,but: ; The temperature risk was determined to be the primary factor, and cooling control measures were implemented, resulting in a temperature reduction of 1.5℃. like , ,but: ; Based on the assessment that concentration risk is the primary factor, flow restriction control is implemented, reducing the flow rate by 0.9 L / min; like , ,but: ; The aggregation degree is determined to be the dominant risk factor. The optimal control trajectory is solved by quadratic programming to make the aggregation degree converge toward the target value of 4.0. like , , , ,but: ; The process parameter deviation was determined to be the main factor, and a multi-parameter joint correction was performed. The pH was reduced by 0.16 units, the flow rate was reduced by 8 L / min, and the rotation speed was reduced by 64 rpm. The pH, flow rate, and rotation speed were adjusted simultaneously.
[0062] In some embodiments, the pH of the reaction system can be reduced by 0.16 units by decreasing the opening of the alkali feed valve or increasing the opening of the acid feed valve.
[0063] S802. Apply safety control measures to the reaction system to reduce the overall risk score below the threshold. In some embodiments, after the safety control quantity is applied, the comprehensive risk score is recalculated until... Exit the safety protection mode and return to the normal control cycle.
[0064] A second aspect of the present invention provides a potassium polyphosphate short-chain polymerization degree synthesis control system for executing the aforementioned potassium polyphosphate short-chain polymerization degree synthesis control method, comprising: The acquisition module is used to acquire the real-time synthesis status; The prediction module is used to predict the degree of aggregation based on the real-time synthesis status; The determination module is used to determine control decisions based on the predicted degree of aggregation, a first degree of aggregation threshold, and a second degree of aggregation threshold. The solver module is used to solve for the optimal control sequence based on the objective function and constraints. The first execution module is used to execute the first control input in the optimal control sequence; The scoring module is used to make risk scoring decisions based on the composite state after the execution of the first step control quantity; The update module is used to update the parameters of the prediction model; The second execution module is used to execute the security protocol.
[0065] In some embodiments, Figure 2 This is the circuit diagram of the pH detection submodule of the acquisition module. S-201-WP represents the pH sensor, VDD represents the positive power supply, GND represents ground, OUT represents the analog signal output, and NC represents no connection.
[0066] In some embodiments, Figure 3 This is the circuit diagram of the temperature detection submodule of the acquisition module. Here, DS18B20 temperature sensor represents the temperature sensor, VCC represents the positive power supply, IO represents the data line, GND represents ground, and PB11 represents the GPIO pin number of the microcontroller / microcontroller.
[0067] In some embodiments, Figure 4 This is the circuit diagram of the motor speed regulation submodule in the execution module. In the diagram, L298N represents the motor driver chip; IN1 and IN2 represent the motor direction control inputs; ENA represents the motor enable (PWM speed control input); OUT1 and OUT2 represent the motor drive outputs; VCC represents the positive power supply; VS represents the positive motor drive power supply; GND represents ground; SENSA and SENSB represent current detection terminals; D1, D2, D3, and D4 represent freewheeling diodes; 3M represents the DC motor; the photoelectric speed sensor represents the speed detection sensor; IN5 represents the sensor input; OUT represents the sensor output; and P32 represents the GPIO pin number of the microcontroller.
[0068] Finally, it should be noted that the embodiments disclosed in this invention are merely preferred embodiments of this invention and are only used to illustrate the technical solutions of this invention, not to limit it. Although this invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this invention.
Claims
1. A method for controlling the degree of polymerization of short-chain potassium polyphosphate, characterized in that, Includes the following steps: Obtain real-time synthesis status; Based on the real-time synthesis status, the degree of aggregation is predicted using a prediction model; Based on the predicted degree of aggregation, a first degree of aggregation threshold, and a second degree of aggregation threshold, control decisions are determined. In response to adjustment control, the optimal control sequence is solved based on the objective function and constraints. Execute the first control input in the optimal control sequence; Risk scoring and decision-making are based on the composite state after the execution of the first step control variable; In response to the model update decision, update the parameters of the prediction model; Respond to security protection decisions and execute security protocols.
2. The method for controlling the degree of polymerization of short-chain potassium polyphosphate according to claim 1, characterized in that, The prediction model is as follows: ； in, To predict the degree of aggregation, pH level The reaction temperature, This refers to the phosphate concentration. , , , and These are the model parameters.
3. The method for controlling the degree of polymerization of short-chain potassium polyphosphate according to claim 1, characterized in that, The method for determining control execution includes: When the predicted aggregation degree is greater than or equal to the first aggregation degree threshold and less than or equal to the second aggregation degree threshold, maintain the current control input; When the predicted aggregation degree is less than the first aggregation degree threshold, adjust the control input to the first control vector; When the predicted aggregation degree is greater than the second aggregation degree threshold, adjust the control input to the second control vector; Wherein, the first control vector is: ； in, The first control vector, For temperature control gain, For flow control gain, For speed control gain; In some embodiments, the second control vector is: ； in, This is the second control vector. For temperature control gain, For flow control gain.
4. The method for controlling the degree of polymerization of short-chain potassium polyphosphate according to claim 1, characterized in that, The objective function is: ； in, Let be the objective function. For temperature adjustment amount, Adjust the amount of flow. This is the speed adjustment amount. To control input Predicted aggregation degree , , These are adaptive weighting coefficients; The constraints include: Adjusting rate constraints: ， ， ； Safety range constraints: ， ， ； The adaptive weights are: ； in, Temperature weighting, To predict the degree of aggregation; ； in, For traffic weight, This refers to the phosphate concentration. ； in, The rotational speed weight is used.
5. The method for controlling the degree of polymerization of short-chain potassium polyphosphate according to claim 1, characterized in that, The risk scoring decision includes: Calculate the comprehensive risk score based on the composite state after the execution of the first step control quantity; The aggregation degree prediction error is calculated based on the actual aggregation degree and the predicted aggregation degree. The decision-making logic is determined based on comprehensive risk scoring and aggregation degree prediction error calculation.
6. The method for controlling the degree of polymerization of short-chain potassium polyphosphate according to claim 5, characterized in that, The comprehensive risk score is calculated as follows: ； in, For comprehensive risk scoring, Temperature risk weighting, For temperature risk scoring, For concentration risk weights, For concentration risk scoring, For aggregation degree risk weight, For aggregation degree risk scoring, As the weight for process parameter deviations, Scoring is given for deviations in process parameters; The temperature risk score is as follows: ； in, For temperature risk scoring, This is the highest temperature inside the reactor; The concentration risk score is as follows: ； in, For concentration risk scoring, This represents the maximum concentration of phosphate. The aggregation degree risk score is: ； in, For aggregation degree risk scoring, To predict the degree of aggregation; The process parameter deviation score is: ； in, Scoring for process parameter deviations for Adjustment amount, Adjust the amount of flow. This is the speed adjustment amount.
7. The method for controlling the degree of polymerization of short-chain potassium polyphosphate according to claim 6, characterized in that, The method for calculating the degree of aggregation prediction error is as follows: ； in, For the degree of aggregation prediction error, This represents the actual degree of aggregation. To predict the degree of aggregation; The decision-making logic is as follows: ； in, Indicates the decision result, This indicates that control will continue for the next cycle. This indicates that a model update has been triggered. This indicates that security protection has been triggered. As a risk threshold, The aggregation degree prediction error threshold, For the degree of aggregation prediction error, For comprehensive risk scoring.
8. The method for controlling the degree of polymerization of short-chain potassium polyphosphate according to claim 7, characterized in that, The updated prediction model parameters include: Update the prediction model parameters based on parameter update rules; The parameter update rule is as follows: ； in, To update the model parameters, These are the current model parameters. For learning rate, This represents the gradient of the loss function with respect to the parameters. The gradient is: ； in, The gradient of the loss function with respect to the parameters. To predict the degree of aggregation, This represents the actual degree of aggregation. Output the partial derivatives of the PLS model with respect to the model parameters. The regularization coefficient is . These are the parameters of the current model.
9. The method for controlling the degree of polymerization of short-chain potassium polyphosphate according to claim 7, characterized in that, The security protocol includes: Targeted control strategies are determined based on the identification of risk-dominant factors; Apply safety control measures to the response system to reduce the overall risk score below the threshold; The dominant risk factors were identified as follows: ； in, Index of risk-dominant factors Risk weighting Assess risk level; The targeted control strategy is as follows: ； in, For safety control purposes, For temperature control gain, This is the highest temperature inside the reactor. To control the gain by concentration, This represents the maximum concentration of phosphate. For symbolic functions, Adjust the traffic limit. It is a quadratic programming solver. This represents the actual degree of aggregation. To control the gain for process parameters, for Adjustment amount, Adjust the amount of flow. This is the speed adjustment amount. This indicates that temperature risk is the primary factor. This indicates that concentration risk is the primary factor. This indicates that aggregation degree risk is dominant. This indicates that process parameter deviations are the primary factor.
10. A control system for the synthesis of potassium polyphosphate short-chain polymerization degree, characterized in that, The method for controlling the degree of polymerization of potassium polyphosphate short chains according to any one of claims 1-9 includes: The acquisition module is used to acquire the real-time synthesis status; The prediction module is used to predict the degree of aggregation based on the real-time synthesis status; The determination module is used to determine control decisions based on the predicted degree of aggregation, a first degree of aggregation threshold, and a second degree of aggregation threshold. The solver module is used to solve for the optimal control sequence based on the objective function and constraints. The first execution module is used to execute the first control input in the optimal control sequence; The scoring module is used to make risk scoring decisions based on the composite state after the execution of the first step control quantity; The update module is used to update the parameters of the prediction model; The second execution module is used to execute the security protocol.