Wide vermicelli production control system with safety monitoring function

Abstract

The invention discloses a wide vermicelli production control system with a safety monitoring function, and relates to the technical field of automatic control, and the wide vermicelli production control system comprises an online state sensing unit which is used for collecting diffuse reflection spectrum vectors and core temperature in real time; the state estimation and prediction unit is used for receiving the diffuse reflection spectrum vector, calculating a predicted glass transition temperature, calculating a safety margin and generating a dynamic temperature set point and a core safety constraint; the dynamic optimization control unit is used for receiving the core temperature, the dynamic temperature set point and the core security constraint, and performing rolling optimization calculation according to a multi-objective optimization function to generate an optimal control instruction; and the process parameter execution unit is used for receiving the optimal control instruction and adjusting the microwave power and the vacuum degree. Through real-time product state sensing, online prediction and scientific safety boundary construction, complete self-adaptive control of the production process is realized, the drying efficiency and the product quality are cooperatively improved, and the production stability and the yield are greatly enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of automation control technology, in particular to a wide powder production control system with safety monitoring. BACKGROUND

[0002] In the modern production process of low-temperature rapid rehydration wide powder, microwave vacuum combined drying is a key technology that plays a core role. The purpose of this process is to efficiently remove water from the product while maintaining a low temperature, thereby forming a final product with a precise microscopic porous network structure. This special structure is the physical basis for the product's ability to rapidly rehydrate in a low-temperature environment. Specifically, when the wide powder is dehydrated, its glass transition temperature will dynamically increase as the water content decreases. The glass transition temperature is a critical threshold. Once the actual core temperature of the product exceeds the glass transition temperature at any time during the drying process, the high molecular chain segments that make up the product's skeleton will change from a hard, stable glass state to a soft, easily flowing rubber state. This phase change will immediately cause the microscopic porous structure to collapse inward, permanently and irreversibly damaging the product's rapid rehydration function. Therefore, the core temperature of the product must be strictly controlled below its real-time glass transition temperature throughout the drying process.

[0003] However, existing production technologies have fundamental technical flaws in this regard. Current control strategies generally rely on a fixed, pre-set process program, which cannot respond to various unavoidable disturbance factors during production, especially the natural differences in physical properties of different batches of raw materials. These differences directly result in different glass transition temperature curves for each batch of product. Due to the lack of effective means to sense the internal state of the product in real time and accurately obtain this dynamically changing glass transition temperature curve, existing technologies are caught in a dilemma: operators can only set a very low universal processing temperature and power that applies to all possible situations to avoid the risk of product structure collapse, but this greatly sacrifices production efficiency and prolongs the drying cycle. Conversely, if more aggressive process parameters are used to pursue production efficiency, the risk of temperature exceeding the limit will be significantly increased, resulting in unguaranteed product yield and quality stability. SUMMARY

[0004] The purpose of the present application is to provide a wide powder production control system with safety monitoring, which solves the problems in the background art.

[0005] To solve the above technical problems, the present application provides a wide powder production control system with safety monitoring, comprising: an online state sensing unit for real-time acquisition of the diffuse reflectance spectrum vector and core temperature of the wide powder during the drying process. a state estimation and prediction unit configured to receive the diffuse reflectance spectrum vector collected by the online state perception unit, and calculate a predicted glass transition temperature based on the diffuse reflectance spectrum vector and a preset glass transition temperature prediction model; the state estimation and prediction unit is further configured to calculate a safety margin, and generate a dynamic temperature set point and a core safety constraint based on the predicted glass transition temperature and the safety margin; a dynamic optimization control unit configured to receive the core temperature collected by the online state perception unit, and the dynamic temperature set point and the core safety constraint generated by the state estimation and prediction unit; and perform a rolling optimization calculation according to a preset multi-objective optimization function based on the core temperature, the dynamic temperature set point and the core safety constraint, to generate an optimal control instruction; a process parameter execution unit configured to receive the optimal control instruction generated by the dynamic optimization control unit, and adjust the microwave power and the vacuum degree according to the optimal control instruction.

[0006] Preferably, the glass transition temperature prediction model comprises a preset regression coefficient vector and a preset intercept constant; the process of calculating the predicted glass transition temperature by the state estimation and prediction unit comprises: multiplying the diffuse reflectance spectrum vector by the regression coefficient vector, and adding the result of the operation to the intercept constant to obtain the predicted glass transition temperature.

[0007] Preferably, the preset regression coefficient vector and the preset intercept constant are obtained by solving a partial least squares regression algorithm based on actual glass transition temperature reference values of a plurality of wide powder standard samples and near-infrared spectra corresponding thereto through offline calibration experiments.

[0008] Preferably, the process of calculating the safety margin by the state estimation and prediction unit comprises: performing a linear superposition calculation on a prediction root mean square error of the glass transition temperature prediction model, a measurement noise standard deviation of the core temperature sensor and a preset fixed compensation margin to generate the safety margin.

[0009] Preferably, the prediction root mean square error is obtained by statistically evaluating the prediction results of a verification set in offline calibration experiments.

[0010] Preferably, the process of generating the dynamic temperature set point by the state estimation and prediction unit comprises: subtracting the safety margin from the predicted glass transition temperature.

[0011] Preferably, the core safety constraint is that the future core temperature predicted by the preset process model built-in the dynamic optimization control unit is required to be less than or equal to the value obtained by subtracting the safety margin from the predicted glass transition temperature at any time within the prediction time domain.

[0012] Preferably, the preset multi-objective optimization function includes a temperature tracking objective term and a control input objective term; the temperature tracking objective term is used to minimize the deviation between the future core temperature and the dynamic temperature set point; and the control input objective term is used to drive the control input to tend to the rated maximum value of the equipment.

[0013] Preferably, the optimal control instruction includes a microwave power instruction and a vacuum degree instruction; the process parameter execution unit includes a microwave power adjustment unit and a vacuum degree control unit; the microwave power adjustment unit adjusts the output of the microwave generator according to the microwave power instruction; and the vacuum degree control unit adjusts the opening degree of the vacuum valve according to the vacuum degree instruction.

[0014] Advantages Compared with the prior art, the present application has the following advantages: 1. By constructing an advanced closed-loop control system, high adaptability of the production process is achieved. The online state perception unit is used to capture the spectral information of the product in real time, and the built-in state estimation and prediction model is used to instantly solve the dynamically changing glass transition temperature that cannot be obtained in the prior art, accurately understand the core physical state of each batch of product, and automatically match the optimal processing path for different characteristics of the raw materials.

[0015] 2. By means of strict mathematical logic, the dynamic safety boundary in the production process is scientifically and quantitatively established. The safety margin calculated with high precision makes the dynamically generated temperature set point and the core safety constraint have high reliability, so that the dynamic optimization control unit can boldly and continuously push the process parameters such as microwave power to the limit allowed by the equipment under the premise of ensuring the integrity of the microstructure of the product, and finally convert the contradictory efficiency and quality in the traditional process into an optimal objective that can be achieved in coordination.

[0016] 3. By means of real-time data feedback and forward-looking prediction, potential risks that may lead to collapse of the product structure can be actively foreseen and avoided, thereby significantly improving the robustness and final yield of the production process. The dynamic optimization control unit always regards the core safety constraint as an inviolable rule when formulating the control strategy. This risk avoidance mechanism based on prediction fundamentally eliminates the disastrous batch rejection accidents caused by mismatch between the process and the raw materials, brings high stability and repeatability to the production, and has vital value for guaranteeing the reliable delivery of the supply chain. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only show some of the embodiments of the present application, and all other drawings obtained by those of ordinary skill in the art without creative effort based on these drawings also belong to the protection scope of the present application. Figure 1 The figure is a logic diagram of the system of the present application. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort also belong to the protection scope of the present application.

[0019] Embodiment 1: Please refer to Figure 1 The present application provides a wide powder production control system with safety monitoring, comprising: An online state sensing unit is configured to collect a diffuse reflectance spectrum vector and a core temperature of the wide powder in a drying process in real time. A state estimation and prediction unit is configured to receive the diffuse reflectance spectrum vector collected by the online state sensing unit, and calculate a predicted glass transition temperature based on the diffuse reflectance spectrum vector and a preset glass transition temperature prediction model. The state estimation and prediction unit is further configured to calculate a safety margin, and generate a dynamic temperature set point and a core safety constraint based on the predicted glass transition temperature and the safety margin. A dynamic optimization control unit is configured to receive the core temperature collected by the online state sensing unit, and the dynamic temperature set point and the core safety constraint generated by the state estimation and prediction unit. The dynamic optimization control unit is further configured to perform a rolling optimization calculation according to a preset multi-objective optimization function based on the core temperature, the dynamic temperature set point and the core safety constraint, so as to generate an optimal control instruction. A process parameter execution unit is configured to receive the optimal control instruction generated by the dynamic optimization control unit, and adjust a microwave power and a vacuum degree according to the optimal control instruction.

[0020] Aiming at the technical bottleneck that the existing fixed process program cannot adapt to the batch fluctuation of raw materials, resulting in the structural collapse of military wide powder in the pursuit of drying efficiency, the embodiment provides a fundamental solution; through the orderly cooperation of the online state perception unit, the state estimation and prediction unit, the dynamic optimization control unit and the process parameter execution unit, a complete information-physical closed-loop control system is constructed; the system discards the fixed experience parameters, and instead perceives the molecular scale state of the product in real time through the online state perception unit, calculates the phase change safety boundary determining the product quality through the state estimation and prediction unit, and optimizes all process parameters dynamically based on the safety boundary through the dynamic optimization control unit; the direct technical result of this design concept is that the system can independently plan an optimal drying path for each batch of wide powder product, maximize production efficiency while ensuring the integrity of the internal porous network of the product with high certainty, ensuring that the core function of low-temperature rapid rehydration in military scenarios can be realized, and finally achieving the coordinated progress of efficiency and quality.

[0021] Embodiment 2: The glass transition temperature prediction model includes a preset regression coefficient vector and a preset intercept constant; the process of calculating the predicted glass transition temperature by the state estimation and prediction unit includes: multiplying the diffuse reflectance spectrum vector by the regression coefficient vector, and adding the operation result to the intercept constant to obtain the predicted glass transition temperature; The source of the preset regression coefficient vector and the preset intercept constant is: through offline calibration experiments, based on the actual glass transition temperature reference value of multiple groups of wide powder standard samples and the near-infrared spectrum corresponding thereto, the regression coefficient vector and the intercept constant are solved by using the partial least squares regression algorithm; specifically, the actual glass transition temperature reference value is obtained by precisely measuring the wide powder standard samples with different water contents by differential scanning calorimetry, thereby establishing a reliable offline calibration dataset; since the offline calibration dataset covers multiple groups of representative wide powder standard samples, the regression coefficient vector and the intercept constant solved have good universality and robustness; in actual online application, the diffuse reflectance spectrum vector of the current batch of wide powder is collected in real time as the model input, so that the unique physical characteristics of the batch of raw materials are reflected in the calculation result of the predicted glass transition temperature in real time, thereby realizing adaptive processing of the natural differences of different batches of raw materials; The state estimation and prediction unit performs its core calculation function, and its internal logic is rooted in a glass transition temperature prediction model based on the experience modeling theory of chemical metrology.

[0022] In the complex system of wide powder drying, the glass transition temperature as a key threshold determining whether the structure collapses cannot be directly measured online; in order to overcome this technical problem, the application establishes a mathematical bridge from the spectrum information that can be measured online to the glass transition temperature that cannot be directly measured; the theoretical basis is that near-infrared spectrum can sensitively capture the change of hydrogen-containing group molecular vibration inside the product, and the change is directly related to the moisture content and the motion state of the polymer chain segment, thereby there is a stable mapping relationship with the glass transition temperature; the fundamental technical motivation of constructing this model is to make the implicit physical correlation explicit and quantitative, thereby giving the system the ability of predictive control; The model is represented as a refined multiple linear regression formula: ; represents time; represents the predicted glass transition temperature calculated by the model at the moment , which is the core output defining the dynamic safety upper limit of all subsequent control behaviors; represents the diffuse reflectance spectrum data column vector collected by the online state perception unit at the moment and preprocessed, which is the real-time input of the model; represents the regression coefficient vector of the model, which is derived from the offline calibration experiment, and is the optimal weight set solved by the partial least squares regression algorithm to minimize the prediction error; represents the intercept constant of the model, which has the same source as the regression coefficient vector, and is the system bias item solved in the regression calculation; In actual operation, the state estimation and prediction unit continuously inputs the diffuse reflectance spectrum vector collected in real time into this formula, and instantaneously calculates the current predicted glass transition temperature; the profound effect of this application mode is that it successfully converts an invisible, dynamically changing, microphase transition boundary that determines the success or failure of military wide powder products into a clear, quantified safety index that can be used for engineering control; this enables the control system to first have the ability to understand and avoid the risk of structural collapse, ensuring that each wide powder can maintain its perfect internal porous structure at the fastest drying speed.

[0023] Embodiment 3: The process of calculating the safety margin by the state estimation and prediction unit includes: performing linear superposition calculation based on the prediction root mean square error of the glass transition temperature prediction model, the measurement noise standard deviation of the core temperature sensor, and a preset fixed compensation margin, to generate the safety margin; The source of the predicted root mean square error is obtained by statistically evaluating the prediction results of the validation set in the offline calibration experiment.

[0024] The state estimation and prediction unit performs another key calculation immediately after calculating the predicted glass transition temperature, i.e. generating the safety margin; In view of the inherent prediction error of the glass transition temperature prediction model and the noise interference in the measurement process of the core temperature sensor, an uncorrected prediction value cannot be directly used as an absolute boundary for control. Therefore, a robust buffer zone that can cover all known sources of uncertainty in the control loop must be constructed. The technical motivation of the safety margin is not an arbitrary empirical setting, but the result of quantitative evaluation and logical construction of various uncertain factors in the system based on the error superposition principle, aiming to ensure that the actual product core temperature will never touch the real glass transition temperature under any disturbance; The generation logic of this safety margin is precisely defined by the following formula, ensuring that those skilled in the art can implement it accordingly: ; represents the final generated safety margin, which is the key input for setting the dynamic temperature target and core safety constraints subsequently; represents the predicted root mean square error of the glass transition temperature prediction model, which is obtained by statistically evaluating the prediction results of the validation set in the offline calibration experiment, and accurately quantifies the uncertainty of the model itself; represents the measurement noise standard deviation of the core temperature sensor, which is obtained from the sensor factory specifications or through static measurement experiments, and represents the uncertainty of the physical measurement process; and is a dimensionless confidence coefficient, which is selected according to the desired control reliability level, for example, if a confidence level of 99.7% is required, it is selected as 3 according to the normal distribution theory, to ensure extremely high safety; is a pre-set fixed compensation margin, which is an empirical value set by engineers to cover non-ideal effects not described by the model, to enhance the robustness of the system; Through the calculation of this formula, the state estimation and prediction unit no longer relies on ambiguous experience, but generates a scientific, dynamic safety buffer area directly linked to the performance of the system itself; its application effect is revolutionary: it allows the dynamic optimization control unit to drive the production parameters to approach the physical limit with higher confidence, because this carefully calculated safety margin constitutes a highly reliable safety mechanism; this makes the drying efficiency of military wide powder reach an unprecedented height, and the yield of its low-temperature rapid rehydration performance reaches a nearly perfect level.

[0025] Embodiment 4: The process of generating the dynamic temperature set point by the state estimation and prediction unit includes: subtracting the safety margin from the predicted glass transition temperature.

[0026] This step reveals the adaptability of the control logic of the present application, that is, the generation mechanism of the dynamic temperature set point; the logical relationship of this mechanism is: ; represents the dynamic temperature set point, which is not a fixed preset value, but the control target pursued by the system at the current time k, and its generation is adaptive and will change with the change of the product state; represents the predicted glass transition temperature calculated at the current time k, which is calculated according to the spectrum vector collected by the online state perception unit in real time and the preset prediction model, and it defines the dynamic safety upper limit under the current state of the product; represents the final generated safety margin, which is in units of degrees Celsius, and is the key input for subsequent setting of the dynamic temperature target and core safety constraints; This formula clearly shows that the control target of the system is not a fixed, pre-programmed value, but is dynamically synthesized by two real-time calculated variables: the predicted glass transition temperature and the safety margin; the direct technical effect of this process is that the target pursued by the control system is always the optimal solution under the current product state; when the product causes the glass transition temperature to rise due to dehydration, the control target also rises, thereby allowing the system to apply greater energy input to accelerate drying; on the contrary, when the initial glass transition temperature is low due to raw material differences, the control target is automatically adjusted to be low, ensuring the absolute safety of the production process; this adaptive target setting is the key to intelligent and efficient production.

[0027] Embodiment 5: The core safety constraint is that the future core temperature predicted by the preset process model embedded in the dynamic optimization control unit is required to be less than or equal to the value obtained by subtracting the safety margin from the predicted glass transition temperature at any time within the prediction horizon.

[0028] After the dynamic pursuit target is established, the present application further sets a hard constraint that must be strictly followed, i.e., the core safety constraint; the mathematical expression of this constraint is: ; represents the future core temperature predicted by the preset process model embedded in the dynamic optimization control unit; represents the dynamic safety boundary, i.e., the value obtained by subtracting the safety margin from the predicted glass transition temperature; is the current time; is the future step within the prediction horizon; Specifically, the preset process model can be described by the following discrete-time state-space model: ; ; is the state vector of the system, which includes key state variables such as the core temperature; is the control input vector, i.e., the microwave power and the vacuum degree; is the system output, i.e., the core temperature predicted by the model ; is the state-space matrix, the parameters of which are obtained by performing system identification experiments on the drying process and fitting the collected input-output data using algorithms such as subspace identification; Specifically, those skilled in the art can obtain the state-space matrix through the following conventional experimental steps: First, stabilize the wide powder drying process near a typical operating point; Then, apply specially designed disturbance signals with sufficient excitation characteristics, such as pseudo-random binary sequence signals, on the control input end, i.e., the microwave power and the vacuum degree; At the same time, continuously record the changes in the system output, i.e., the core temperature; Finally, input the collected multiple sets of input-output data pairs into the system identification toolbox and fit them using standard subspace identification algorithms to solve the value of the state-space matrix This process can be completed without excessive experiments; Based on this model, the dynamic optimization control unit can determine the control sequence to be implemented in the future according to the current system state and the future planned control sequence , to predict the future core temperature sequence within the prediction horizon This constraint, together with the aforementioned dynamic temperature setpoint, constitutes a hierarchical control philosophy: the dynamic temperature setpoint is the target that the system strives to achieve, while the core safety constraint is an insurmountable boundary; when the dynamic optimization control unit performs optimization solving, it can flexibly adjust the strategy to approach the target, but any control sequence that may cause the predicted future core temperature to touch this boundary will be directly rejected; the technical effect of this design is to establish a forward-looking safety guarantee mechanism, completely eliminating the possibility of temperature overshoot caused by system disturbance or model deviation, thereby providing ultimate protection for the microstructure integrity of military wide-powder products.

[0029] Embodiment 6: The preset multi-objective optimization function includes a temperature tracking target term and a control input target term; the temperature tracking target term is used to minimize the deviation between the future core temperature and the dynamic temperature setpoint; and the control input target term is used to drive the control input to tend to the rated maximum value of the equipment.

[0030] Specifically, the preset multi-objective optimization function can be expressed as a cost function in the following form : ; is the prediction horizon; is the control horizon, and ; is the predicted future core temperature of the process model; is the dynamic temperature setpoint sequence at the future time, which can generally be assumed to remain unchanged at the current value within the prediction horizon, i.e. ; is the increment of the control input, i.e., the change amount of the microwave power and the vacuum degree, ; and are positive definite weight matrices, by adjusting their relative sizes, the temperature tracking accuracy (represented by the first term) and the smoothness and economy of the control input (represented by the second term) can be balanced, so as to achieve comprehensive optimization of safety and efficiency; This function is minimized at each control cycle ; ​​The decision function of the dynamic optimization control unit is solved according to a well-designed preset multi-objective optimization function; the function ingeniously combines two seemingly conflicting objectives: "quality and safety" represented by the temperature tracking objective term, and "efficiency and speed" represented by the control input objective term; in each control cycle, the task of the optimization algorithm is to find an optimal control instruction that minimizes the comprehensive cost of the two objective terms within the feasible region defined by the core safety constraints; the deep effect of this mechanism is that it converts the complex production decision-making process into a well-defined mathematical optimization problem; the system no longer needs human intervention, but autonomously makes the best trade-off between safety and efficiency, and the result is to produce a perfect process path that closely follows the dynamic safety boundary and constantly seeks maximum energy input; this is a truly intelligent manufacturing.

[0031] Embodiment 7: The optimal control instruction includes a microwave power instruction and a vacuum degree instruction; the process parameter execution unit includes a microwave power adjustment unit and a vacuum degree control unit; the microwave power adjustment unit adjusts the output of the microwave generator according to the microwave power instruction; and the vacuum degree control unit adjusts the opening degree of the vacuum valve according to the vacuum degree instruction.

[0032] Finally, the optimal control instruction calculated by the dynamic optimization control unit is converted into accurate manipulation of the physical world through the process parameter execution unit; the optimal control instruction is not an abstract signal, but a numerical sequence containing specific microwave power instructions and vacuum degree instructions; the microwave power adjustment unit and the vacuum degree control unit faithfully execute these instructions to adjust the output of the microwave generator and the opening degree of the vacuum valve, respectively; this execution process constitutes the last link of the entire closed-loop control loop, which applies the optimization decision of the information world to the production equipment of the physical world without loss; then, the product state changes caused thereby are captured by the online state perception unit, thereby starting the next "perception-prediction-decision-execution" cycle; through such millisecond-level, uninterrupted rolling optimization and feedback regulation, the system ultimately realizes unmanned, adaptive and optimal control of the wide powder drying process, ensuring that every military material has extreme performance.

[0033] The above is only a preferred embodiment of the present application, and does not limit the present application in other forms. Any skilled person in the art can modify or change the above disclosed technical content to equivalent embodiments applied to other fields, but any simple modification, equivalent change and modification made to the above embodiments without departing from the technical solution content of the present application, according to the technical essence of the present application, still belongs to the protection scope of the technical solution of the present application.

Claims

1. A wide powder production control system with safety monitoring, characterized by, The method comprises the following steps: an online state sensing unit is used to collect a wide powder's diffuse reflectance spectrum vector and a core temperature in real time during a drying process; a state estimation and prediction unit is used to receive the diffuse reflectance spectrum vector collected by the online state sensing unit, and calculate a predicted glass transition temperature based on the diffuse reflectance spectrum vector and a preset glass transition temperature prediction model; the state estimation and prediction unit is further used to calculate a safety margin, and generate a dynamic temperature set point and a core safety constraint based on the predicted glass transition temperature and the safety margin; a dynamic optimization control unit is used to receive the core temperature collected by the online state sensing unit, and the dynamic temperature set point and the core safety constraint generated by the state estimation and prediction unit; based on the core temperature, the dynamic temperature set point and the core safety constraint, the dynamic optimization control unit performs a rolling optimization calculation according to a preset multi-objective optimization function to generate an optimal control instruction; a process parameter execution unit is used to receive the optimal control instruction generated by the dynamic optimization control unit, and adjust a microwave power and a vacuum degree according to the optimal control instruction.

2. A wide powder production control system with safety monitoring according to claim 1, characterized in that, The glass transition temperature prediction model comprises a preset regression coefficient vector and a preset intercept constant; the state estimation and prediction unit calculates the predicted glass transition temperature by multiplying the diffuse reflectance spectrum vector by the regression coefficient vector, and adding the result of the multiplication to the intercept constant.

3. A wide powder production control system with safety monitoring according to claim 2, characterized in that, The preset regression coefficient vector and the preset intercept constant are obtained by solving a partial least squares regression algorithm based on actual glass transition temperature reference values of a plurality of wide powder standard samples and near-infrared spectra corresponding to the actual glass transition temperature reference values through offline calibration experiments.

4. The wide powder production control system with safety monitoring according to claim 1, characterized in that, The state estimation and prediction unit calculates the safety margin by linearly superimposing a prediction root mean square error of the glass transition temperature prediction model, a measurement noise standard deviation of the core temperature sensor and a preset fixed compensation margin.

5. A wide powder production control system with safety monitoring according to claim 4, characterized in that, The prediction root mean square error is obtained by statistically evaluating prediction results of a verification set in offline calibration experiments.

6. A wide powder production control system with safety monitoring according to claim 1, characterized in that, The state estimation and prediction unit generates the dynamic temperature set point by subtracting the safety margin from the predicted glass transition temperature.

7. A wide powder production control system with safety monitoring according to claim 1, characterized in that, The core safety constraint requires that a future core temperature predicted by a preset process model built in the dynamic optimization control unit at any time within a prediction time domain is less than or equal to a value obtained by subtracting the safety margin from the predicted glass transition temperature.

8. A wide powder production control system with safety monitoring according to claim 1, characterized in that, The preset multi-objective optimization function comprises a temperature tracking objective term and a control input objective term; the temperature tracking objective term is used to minimize a deviation between a future core temperature and the dynamic temperature set point; the control input objective term is used to drive a control input to tend to a rated maximum value of equipment.

9. A wide powder production control system with safety monitoring according to claim 1, characterized in that, The optimal control instruction comprises a microwave power instruction and a vacuum degree instruction; the process parameter execution unit comprises a microwave power adjusting unit and a vacuum degree control unit; the microwave power adjusting unit adjusts the output of the microwave generator according to the microwave power instruction; and the vacuum degree control unit adjusts the opening degree of the vacuum valve according to the vacuum degree instruction.

Images