Glass production process intelligent control system based on PLC

Abstract

The invention relates to the technical field of glass production process intelligent control, in particular to a glass production process intelligent control system based on a PLC. The system comprises a data acquisition module, a first processing module, a second processing module, a damage evaluation module, a state determination module and a parameter adjustment module. The system predicts a temperature field in the kiln and calculates thermal stress by collecting PLC and sensor data and using a reduced order model; the core is to track the historical fluctuation of thermal stress and determine the accumulated fatigue damage degree based on the fatigue accumulation theory; converting the temperature into the kiln health degree to determine a health perception regulation factor and a control mode; when the health degree is reduced, the system adaptively adjusts the control parameters of the PLC, so that the control response tends to be gentle and conservative; according to the invention, health perception adjustment of the control strategy is realized, and excessive active control behaviors causing physical damage can be actively avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent control of glass production process, and particularly relates to a PLC-based intelligent control system for glass production process. BACKGROUND

[0002] In the glass production process, PLC (e.g., PID) control is generally used to pursue high-precision temperature tracking; at present, this high-precision control strategy has a significant problem: the excessively active control behavior thereof can cause excessive thermal stress impact on the refractory material of the kiln; however, this instantaneous thermal stress impact can continuously accumulate to cause fatigue damage of the material, and finally cause the service life of the kiln to be shortened; therefore, how to actively avoid the physical damage caused by excessive control while ensuring the temperature precision required by the production process to prolong the service life of the kiln asset has become a technical problem to be solved in the field. SUMMARY

[0003] To solve the above technical problem, the present application provides a PLC-based intelligent control system for glass production process, and specifically, the technical scheme of the present application comprises: a data acquisition module, configured to acquire control output data of the PLC and measurement data of the kiln sensor in real time; a first processing module, configured to predict the internal temperature field distribution at the key section of the refractory material of the kiln in real time based on the control output data and the measurement data and by calling a pre-trained thermodynamic reduced-order model of the kiln; a second processing module, configured to calculate the instantaneous thermal stress at the key section based on the internal temperature field distribution and in combination with preset physical properties of the refractory material; a damage assessment module, configured to track the historical fluctuation data of the instantaneous thermal stress and determine the cumulative fatigue damage degree at the key section based on a preset fatigue analysis algorithm and a linear fatigue cumulative damage theory; a state determination module, configured to convert the cumulative fatigue damage degree into a kiln health degree and determine a health-aware adjustment factor and a control mode in response to the kiln health degree; a parameter adjustment module, configured to adjust the control parameters of the PLC based on the health-aware adjustment factor and the control mode.

[0004] Preferably, the data acquisition module is configured to acquire the control output data and the measurement data, wherein: the control output data comprises a fuel supply valve opening degree and an electric heating power; the measurement data comprises multi-point temperatures inside the kiln.

[0005] Preferably, the second processing module is configured to calculate the instantaneous thermal stress, comprising: Call internal temperature field distribution, and pre-calibrated elastic modulus, thermal expansion coefficient and stress-free reference temperature of refractory material; Determine instantaneous thermal stress based on the theory of thermoelasticity.

[0006] Preferably, the second processing module is further configured to: Calculate a thermal shock index based on a time rate of change of the internal temperature field distribution and a preset material safety limit; Determine an instantaneous damage risk classification at the critical section based on a comparison of the thermal shock index with a preset risk threshold.

[0007] Preferably, determining the instantaneous damage risk classification comprises: Determining safe when the thermal shock index is less than or equal to a first risk threshold; Determining a first-level risk when the thermal shock index is greater than the first risk threshold and less than or equal to a second risk threshold; Determining a second-level risk when the thermal shock index is greater than the second risk threshold.

[0008] Preferably, the damage assessment module is configured to determine a cumulative fatigue damage degree, comprising: Processing historical fluctuation data of the instantaneous thermal stress using a preset fatigue analysis algorithm to identify stress amplitude and occurrence frequency; Determining a total number of allowable failure cycles based on the stress amplitude and calling a preset material S-N curve; Determining the cumulative fatigue damage degree based on the occurrence frequency and the total number of allowable failure cycles and according to the Palmgren-Miner linear fatigue cumulative damage theory.

[0009] Preferably, the state determination module is further configured to: Predict a remaining service life of the critical section based on the kiln health degree and a preset total design life of the kiln.

[0010] Preferably, the state determination module is configured to determine a factor and a mode, comprising: Determining a global minimum health degree as the minimum kiln health degree among all critical sections; Calculating a health perception adjustment factor based on the global minimum health degree and a preset health degree sensitivity weight; Determining a control mode by comparing the global minimum health degree with preset maintenance mode and equilibrium mode thresholds.

[0011] Preferably, determining the control mode comprises: Determining a maintenance mode when the global minimum health degree is greater than the maintenance mode threshold; determining the balancing mode when the global minimum health degree is less than or equal to the balancing mode threshold value; determining the protection mode when the global minimum health degree is less than or equal to the balancing mode threshold value.

[0012] Preferably, the parameter adjustment module is configured to adjust the control parameter, comprising: determining a final adjustment factor according to the control mode; wherein, when the control mode is the maintenance mode or the balancing mode, the final adjustment factor is set to be equal to the health-aware adjustment factor; wherein, when the control mode is the protection mode, the final adjustment factor is set to be a product of the health-aware adjustment factor and a preset protection coefficient; correcting the proportional gain and the integral time of the PID control loop in the PLC based on the final adjustment factor.

[0013] Compared with the prior art, the present application has the following beneficial effects: 1. The system realizes adaptive adjustment of the control strategy; the system can dynamically adjust the control parameter of the programmable logic controller (PLC) according to the real-time health state of the kiln; when the health degree decreases, the control response tends to be gentle and conservative, and the overactive control behavior leading to physical damage is actively avoided; 2. The system realizes quantification from the control behavior to the physical consequences; the unmeasurable temperature field inside the kiln is predicted in real time by calling the pre-trained agent model, and the instantaneous thermal stress is calculated combined with the theory of thermal elasticity mechanics, so as to directly associate the output behavior of the controller with the physical impact consequences borne by the refractory material; 3. The system realizes evaluation from the instantaneous impact to the long-term damage; the fatigue analysis algorithm and the linear cumulative damage theory are adopted to scientifically accumulate the high-frequency and discrete instantaneous thermal stress fluctuations into the cumulative fatigue damage degree and the kiln health degree representing the long-term trend, thereby providing a core basis for predictive maintenance and asset life management; 4. The system establishes a closed-loop control dynamically balancing between production precision and asset life; the system takes the long-term health degree of the kiln as a high-level constraint condition to adjust the bottom-layer control parameter reversely, thereby constructing an intelligent closed loop dynamically balancing between guaranteeing the temperature precision required by the production process and prolonging the asset life of the kiln, and realizing maximization of the operation benefit. BRIEF DESCRIPTION OF DRAWINGS

[0014] The present application will be further explained in conjunction with the accompanying drawings and embodiments: Figure 1 is a structural diagram of the system of the present application. DETAILED DESCRIPTION

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0016] Example 1: Please see Figure 1 A PLC-based intelligent control system for glass production process, comprising: The data acquisition module is used to collect control output data from the PLC and measurement data from the kiln sensors in real time. The first processing module is used to predict the internal temperature field distribution at key sections of the kiln refractory material in real time based on control output data and measurement data, and by calling a pre-trained kiln thermodynamic reduced-order proxy model. The second processing module is used to calculate the instantaneous thermal stress at the key section based on the internal temperature field distribution and in combination with the preset physical properties of the refractory material. The damage assessment module is used to track historical fluctuation data of instantaneous thermal stress and determine the cumulative fatigue damage degree at key sections based on the preset fatigue analysis algorithm and linear fatigue cumulative damage theory. The status determination module is used to convert the cumulative fatigue damage degree into the kiln health degree, and in response to the kiln health degree, determine the health perception adjustment factor and control mode. The parameter adjustment module is used to adjust the control parameters of the PLC based on the health perception adjustment factor and control mode.

[0017] This embodiment provides a PLC-based intelligent control system for glass production processes. The system aims to solve the problem that traditional PLC control, such as PID control, may cause excessive thermal stress impact on the refractory materials of the kiln when pursuing high-precision temperature tracking, thereby leading to fatigue damage and shortening the service life of the kiln. This system actively extends the asset life of the kiln by constructing a health-aware adaptive closed loop while ensuring the production process. To achieve the above objectives, the system includes a data acquisition module, a first processing module, a second processing module, a damage assessment module, a status determination module, and a parameter adjustment module. The data acquisition module aims to provide a real-time data foundation for subsequent state perception and prediction. In this embodiment, this module is used to collect control output data from the programmable logic controller (PLC) and measurement data from the kiln sensors in real time, constructing an operating condition dataset. To further clarify, the control output data may include the opening degree of the fuel supply valve. Electric heating power etc.; the measurement data may include the temperature at multiple points inside the kiln. Smoke components To clearly illustrate the core technical path of this embodiment, the subsequent first processing module will mainly adopt... , and As a proxy model To predict the input; Other data can be used in extended applications to assist in calibration or model correction; The first processing module aims to perceive in real time the dynamic temperature field distributed within the refractory material, which cannot be directly measured by the PLC control system. This temperature field is the direct input for subsequent thermal stress analysis. In this embodiment, this module is based on the control output data acquired by the data acquisition module. , With measurement data And invoke a reduced-order proxy model of kiln thermodynamics. Real-time prediction of critical cross-sections of kiln refractory materials Internal temperature field distribution ; Here, the kiln thermodynamic reduced-order surrogate model This refers to a reduced-order model pre-trained for real-time computation; the model is built upon massive, offline, high-precision computational fluid dynamics (CFD) and thermo-mechanical coupled finite element analysis (FEA) simulation datasets; in this embodiment, the reduced-order surrogate model of the kiln thermodynamics... Specifically, it is a deep learning model based on a Long Short-Term Memory (LSTM) neural network, which is particularly suitable for processing... , and This type of input data has time-series characteristics; The training process of this model specifically includes the following steps: High-precision simulation data generation: A high-precision three-dimensional digital twin model of the kiln is constructed using computational fluid dynamics (CFD) and thermo-mechanical coupled finite element analysis (FEA) simulation software. Multiple sets of PLC control sequences covering common kiln operating conditions are set as simulation inputs, and transient simulations are run to obtain key internal sections of the kiln under different control inputs. Temperature field at Over time High-precision simulation data that is changing; Training dataset construction; using the simulation data obtained in step S1 , And the sensor measurements simulated by the simulation model The internal temperature field obtained from the simulation is used as the input feature of the model. As the output labels of the model; to reflect temporal characteristics, the input features can be constructed as a series of features from the past period. Data sequence; Model training and validation: The constructed dataset is divided into training and validation sets; using the training set, the network weights of the LSTM model are iteratively trained using the backpropagation algorithm and the Adam optimizer, with the training objective being to minimize the model's predictions. Compared with the simulated real value The root mean square error (RMSE) between the two sets is used to measure the accuracy of the model. When the RMSE of the model on the validation set is lower than the preset accuracy threshold, training is stopped, and a pre-trained reduced-order surrogate model of kiln thermodynamics is obtained. ; Its prediction logic follows a functional relationship. In this functional relationship, represent Always Predicted temperature at location; Represents a pre-trained surrogate model; and This represents the real-time control input of the PLC obtained by the data acquisition module; This represents the sensor measurement input acquired by the data acquisition module; Represents time; Represents spatial location; The second processing module aims to calculate the instantaneous physical stress borne by the material based on the predicted temperature field; in this embodiment, this module is based on the internal temperature field distribution output by the first processing module. And, in conjunction with the preset physical properties of refractory materials, calculate the critical section. Instantaneous thermal stress at the point ; The preset physical properties of refractory materials refer to the material physical constants that have been determined in advance through experiments, mainly including: elastic modulus. The coefficient of thermal expansion of the material and the stress-free reference temperature of the material. ; This calculation is based on standard thermoelasticity theory, and its formula is as follows: In this formula, represent Instantaneous thermal stress at the location; , and All of these are preset material constants as defined above; The predicted temperature for the first processing module; The damage assessment module aims to transform discrete, instantaneous thermal stress fluctuations into continuous, cumulative long-term damage metrics. In this embodiment, the module tracks the instantaneous thermal stress calculated by the second processing module in real time. Based on historical fluctuation data and a pre-defined fatigue analysis algorithm and linear fatigue cumulative damage theory, the key cross-section is determined. exist Cumulative fatigue damage over time ; The preset fatigue analysis algorithm refers to standard algorithms such as simplified rainflow counting, which are used for real-time processing. Time series data were used to identify complete stress cycle events. And statistically analyze the corresponding stress amplitude. and up to Number of times the moment occurs ; The linear fatigue cumulative damage theory refers to the Palmgren-Miner linear fatigue cumulative damage theory; Based on this theory, the formula for calculating cumulative fatigue damage is: ; in, It represents the cumulative fatigue damage degree, and its value ranges from 0 (no damage) to 1 (failure). The representative stress amplitude is The number of times the loop has occurred; Representative materials in Total number of allowable failure cycles under stress; Total number of allowed failure cycles The material's SN curve, also known as the Wöhler curve, is determined by fitting fatigue test data of refractory materials. This fitting process follows the principle of variable independence by establishing a calibration dataset containing multiple sets of data at different constant stress amplitudes. The number of cycle failures obtained by subscripting Through the and Data points are subjected to regression analysis, such as least squares regression, to obtain the fit. The functional relationship, where and These are independent variables in the calibration process, distinct from those calculated during model runtime. ; The state determination module aims to assess the overall health status of the kiln based on quantified cumulative damage and generate factor and mode signals to guide adjustments to the control strategy. In this embodiment, this module uses the cumulative fatigue damage determined by the damage assessment module. Convert to kiln health ; Kiln health This refers to a more intuitive evaluation indicator, with a value range of 1 (healthy) to 0 (failed). The calculation formula is as follows: ; This module responds to kiln health status. Identify the moderating factors of health perception With control mode; Health perception modulators The calculation requires determining the global minimum health level. It is defined as all critical sections. The lowest level of health in the system, i.e. ; Regulatory factors The calculation formula is: ; in, This is an adjustment factor with a value range of 0 to 1; The lowest overall health level; The health sensitivity weight is a preset adjustable parameter, for example... Used for regulation Follow The decreasing decay rate; the weight The value can be set based on the sensitivity to risk according to the operational strategy. For example, a system requiring a high sensitivity response can be set to a larger value. Value; for specific calibration A preferred method is to construct a kiln health value based on historical operating data or simulation data. Descent curve and control parameters , Adjustments to the remaining lifespan of the kiln An impact assessment model; different impacts were iteratively tested on this model. Values, evaluations in different Under this value, the system finds the Pareto optimal balance between control accuracy and asset lifespan extension, and selects the corresponding optimal balance point. The value is used as a preset parameter; The control mode is determined by... It is determined by comparing it with a preset mode threshold; The parameter adjustment module aims to execute the final control loop, feeding back the kiln's health status to the PLC's underlying control logic. In this embodiment, this module uses the health-sensing adjustment factor output by the status determination module. Based on the control mode, determine a final adjustment factor. and use Dynamically adjust the PLC control parameters; In this embodiment, the control parameter being adjusted is the proportional gain of the PID (Proportional-Integral-Derivative) control loop in the PLC. and points time ; , This refers to the original baseline value set by the engineer; Adjusted parameters and The calculation is as follows: ; ; This adjustment logic ensures that when the kiln's health declines... Reduced, leading to Decrease means the control response becomes weaker or slower. Increasing the integral action slows down and makes it smoother, thus making the response of the entire PLC control system more gradual, thereby reducing further impact on the kiln refractory material. This embodiment constructs an intelligent control system capable of sensing its own health status and adaptively adjusting by establishing a complete technical closed loop encompassing data acquisition, temperature prediction, stress calculation, damage assessment, state determination, and parameter adjustment. Its technical advantage lies in achieving the quantification from behavior to consequences, and translating the PLC's control output, for example... , Through proxy models and mechanical models, the physical damage of the kiln is analyzed. , It directly correlates and quantifies in real time the instantaneous impact and long-term cumulative consequences of control decisions on asset health; furthermore, it enables assessment from instantaneous to long-term perspectives, utilizing fatigue accumulation theory to analyze high-frequency instantaneous thermal stresses. This is transformed into a low-frequency change, representing a long-term trend in kiln health. This provides a core basis for predictive maintenance and asset management; at the same time, it realizes control from open-loop to closed-loop through health perception adjustment factors. And control modes, to ensure the long-term health of the kiln As a high-level constraint, it reverse-adjusts the PID parameters at the PLC's underlying level. , This system constructs an adaptive closed loop that dynamically balances control precision and asset lifespan, ultimately improving the overall efficiency of asset operation. While ensuring the temperature uniformity required for glass production processes, the system proactively avoids overly active control behaviors, sacrificing some unnecessary extreme control precision in exchange for a significant extension of the kiln's physical lifespan, thus maximizing operational efficiency.

[0018] Example 2: The data acquisition module is used to acquire control output data and measurement data, including: Control output data includes the fuel supply valve opening and electric heating power; The measurement data includes the temperature at multiple points inside the kiln.

[0019] According to the system of Embodiment 1, in this embodiment, the data content collected by the data acquisition module is specifically defined; The control output data specifically includes: the opening degree of the fuel supply valve. and electric heating power ; This refers to the percentage of valve opening that controls the supply of fuel, such as natural gas, to the kiln. This refers to the power applied to the electric auxiliary heating element; these two are the main means by which the PLC adjusts the energy input of the kiln. The measurement data specifically includes: multiple temperature points inside the kiln. ; This refers to sensors installed at different locations in the kiln, such as thermocouples, that report temperature readings in real time. By clarifying , and As the core data source of the system, it ensures that the first processing module can predict the temperature field. The data foundation is accurate and directly relevant; and This represents energy input. This minimized dataset represents the system's state response. , , This constitutes a thermodynamic surrogate model Predicting the internal temperature field This provides a high-precision input source for all subsequent thermal stress and damage calculations, serving as a necessary and sufficient condition.

[0020] Example 3: The second processing module is used to calculate instantaneous thermal stress, including: It calls upon the internal temperature field distribution, as well as the pre-calibrated elastic modulus, thermal expansion coefficient, and stress-free reference temperature of the refractory material; Instantaneous thermal stress is determined based on thermoelasticity theory.

[0021] According to the system of Embodiment 1, in this embodiment, the implementation path for the second processing module to calculate instantaneous thermal stress is specifically defined; This module calls the internal temperature field distribution output by the first processing module. ; At the same time, the module calls the pre-calibrated elastic modulus of the refractory material. Coefficient of thermal expansion and stress-free reference temperature As mentioned above, these parameters are all physical property constants determined in advance through materials science experiments and stored in the system's parameter library. This module is based on thermoelasticity theory and, through... The formula for determining instantaneous thermal stress It should be noted that here... It is an equivalent thermal stress characterizing the degree of constraint on local thermal expansion; this simplified formula is used in this embodiment, assuming a critical section. Temperature at the location It is the main contributor to thermal stress in this local area, while in the surrogate model During the training process, the spatial gradient effect has been implicitly included through CFD / FEA simulation; this simplified formula The equivalent calculation of one-dimensional thermal stress or biaxially confined thermal stress is used, and its accuracy depends on the approximation of the temperature gradient at the critical section to the one-dimensional or biaxial constraint. thermal stress was clarified The physical basis of the calculation is the theory of thermoelasticity, and the necessary parameter source is the predicted temperature. and preset material constants , , This ensures that the calculation of instantaneous thermal stress is a standard, reproducible, and physically meaningful process, avoiding the uncertainties brought about by black-box models or empirical formulas, and providing standardized and reliable stress input data for the rigorous mechanical analysis of subsequent fatigue damage accumulation.

[0022] Example 4: The second processing module is further used for: The thermal shock index is calculated based on the time-varying rate of change of the internal temperature field distribution and the preset material safety limits. The instantaneous damage risk level at the critical section is determined by comparing the thermal shock index with the preset risk threshold.

[0023] According to the system of Embodiment 1, in this embodiment, the function of the second processing module is further expanded, enabling it to not only calculate stress. It can also assess the risk of rate of temperature change; This module further considers the time-varying rate of change of the internal temperature field distribution. and preset material safety limits Calculate the thermal shock index ; Time-varying rate of change of internal temperature field distribution It refers to Regarding time The first derivative, i.e. It characterizes the heating and cooling rates within the material; to improve the robustness of the calculation, in the calculation... Previously, in response to predicted temperatures Time-series data can be smoothed using methods such as moving average filtering or Savitzky-Golay filtering to effectively suppress surrogate models. High-frequency noise introduced by the output or numerical differentiation process; Preset material safety limits This refers to a safe heating and cooling rate threshold preset based on the properties of refractory materials and process experience, for example, 5°C / minute. The determination of this threshold is based on statistical analysis of thermal shock stability test data of specific refractory materials, taking the critical temperature change rate at which no cracks are generated under multiple cycles. Thermal shock index It is a custom, normalized index designed to quantify localized thermal shocks caused by overactive PLC control. The calculation formula is as follows: During calculation, the system will perform verification. Is it greater than a safe non-zero value to prevent division by zero due to incorrect parameter configuration? If If the value is zero or close to zero, the system will use a preset default safety limit or issue a parameter error alarm; This module is based on the calculated thermal shock index. The critical section is determined by comparing it with a preset risk threshold. Risk classification of instantaneous damage at the site; By introducing the thermal shock index This innovative assessment dimension enables the system to surpass the capabilities of traditional stress analysis; This reflects the stress caused by uneven temperature distribution, and This reflects the impact caused by excessively rapid temperature changes; this enables the system to quantify the PLC control behavior in real time, especially the instantaneous and localized thermal shock damage caused by rapid adjustments to the physical entity of the kiln, and realizes a multi-dimensional assessment of the sources of damage risk, namely thermal stress and temperature change impact.

[0024] Example 5: Determining the instantaneous damage risk classification includes: When the thermal shock index is less than or equal to the first risk threshold, it is determined to be safe; When the thermal shock index is greater than the first risk threshold and less than or equal to the second risk threshold, it is determined to be a level 1 risk. When the thermal shock index is greater than the second risk threshold, it is determined to be a level two risk.

[0025] According to the system of Embodiment 4, the specific logic for determining the instantaneous damage risk classification is described in detail in this embodiment; This logic relies on two preset risk thresholds: a first risk threshold and a second risk threshold; The first risk threshold is set as follows in this embodiment. ; This means the current rate of temperature change Exactly equal to the material's safety limit ; The second risk threshold is set to [value] in this embodiment. ,For example This value It is a pre-set upper limit threshold for primary risk based on kiln operation experience and refractory material test data. Its setting is based on the following: When this value is exceeded, the probability of observing microcrack propagation increases significantly; The specific grading logic is defined as: when the thermal shock index When the temperature change rate is within the material's tolerance range, it is considered safe; when the thermal shock index... When the temperature change rate slightly exceeds the limit, it is identified as a Level 1 risk; at this point, the thermal shock index is considered to be at risk of thermal stress damage. At this point, the risk level was determined to be Level 2, indicating that the temperature change rate was severely exceeding the standard and the PLC's control behavior was causing significant thermal shock damage to the refractory material. By setting and Two clearly defined thresholds will provide a continuous, normalized thermal shock index. This is transformed into discrete risk levels with clear guiding significance, such as safety, level 1 risk, and level 2 risk. This enables the control system to intuitively understand the instantaneous degree of harm of the current control behavior and can serve as a clear basis for triggering immediate alarms or executing emergency control strategies, such as immediately leveling the control response.

[0026] Example 6: The damage assessment module is used to determine the cumulative fatigue damage degree, including: A preset fatigue analysis algorithm is used to process historical fluctuation data of instantaneous thermal stress and identify stress amplitude and occurrence frequency. Based on the stress amplitude and by calling the preset material SN curve, the total number of allowable failure cycles is determined; Based on the number of occurrences and the total number of allowable failure cycles, and according to the Palmgren-Miner linear fatigue cumulative damage theory, the cumulative fatigue damage degree is determined.

[0027] According to the system of Embodiment 1, in this embodiment, the cumulative fatigue damage degree is determined by the damage assessment module. The internal steps were broken down in detail; This process involves cycle identification, lifetime query, and damage accumulation; During the cyclic identification phase, this module employs a preset fatigue analysis algorithm, which in this embodiment is a simplified rainflow counting method, to continuously process the instantaneous thermal stress output by the second processing module. The historical fluctuation data is processed in real time; the purpose of this step is to identify closed fatigue cycle events from continuous, irregular stress fluctuations and to statistically determine the stress amplitude of each event. and the number of times it occurs ; During the lifespan query phase, based on the stress amplitude identified in the previous step... And call the preset material SN curve to determine the... The total number of allowable failure cycles for a material under stress level As mentioned earlier, the material SN curve, also known as the Wöhler curve, is obtained by fitting a large amount of material fatigue test experimental data in advance. It defines the relationship between stress amplitude and fatigue life, i.e., the number of cycles. During the damage accumulation phase, based on the number of occurrences Total number of allowable failure cycles And strictly based on the Palmgren-Miner linear fatigue cumulative damage theory, through The formula is used to calculate the cumulative fatigue damage. ; A standardized academic process for transforming stress history into cumulative damage is defined in detail; the stress history is scientifically decomposed using rainflow counting, and then damage is quantified using SN curves and Miner's rule; this process transforms discrete, instantaneous thermal stress... Therefore, scientifically and traceably accumulated, it forms a continuous, monotonically increasing cumulative damage level. In short, it has achieved a quantitative evolution from instantaneous impact to long-term irreversible damage, providing the most crucial input for subsequent prediction of kiln health and remaining life.

[0028] Example 7: The status determination module is further used for: Based on the kiln health status and the preset total design life of the kiln, the remaining service life of key sections is predicted.

[0029] According to the system of Embodiment 1, in this embodiment, the function of the status determination module is further extended so that it can predict the remaining lifespan based on the assessment of health status; This module is based on the calculated kiln health. ,in and the pre-set total lifespan of the kiln design. Predicting the remaining service life of critical sections ; Preset total design life of the kiln This refers to a parameter pre-set according to kiln design specifications and industry standards, such as 10 years or 12 years; Remaining service life It is a customized, more intuitive evaluation metric; In this embodiment, a linear mapping method is used to represent health status. Converted to remaining lifespan The calculation formula is as follows: ; This embodiment uses a linear mapping method to provide operators with an intuitive estimate of remaining useful life; in other embodiments, a nonlinear mapping based on a material damage evolution model can also be used, for example... ,in It is a nonlinear damage factor to more accurately reflect the accelerated cumulative effect of damage in the final stage; By using abstract, percentage-based health... For example, 0.6 represents 60% health, which is linearly mapped to a concrete remaining lifespan in units of time. For example, 10 years 0.6 = 6 years, providing more intuitive and actionable reference information for kiln operation and maintenance decision-makers; this lifespan prediction in years or months greatly improves the accuracy and operability of predictive maintenance plans, such as scheduling major overhauls.

[0030] Example 8: The state determination module is used to determine factors and patterns, including: Determine the lowest kiln health among all critical sections as the global minimum health; Based on the global minimum health level and the preset health sensitivity weight, calculate the health perception moderating factor; The control mode is determined by comparing the global minimum health level with the preset maintenance mode threshold and the balanced mode threshold.

[0031] According to the system of Embodiment 1, in this embodiment, the internal logic of the state determination module for determining factors and patterns is described in detail; This logic involves determining the global state, calculating the adjustment factor, and determining the control mode; To determine the global state, this module tracks all critical sections in real time. health And determine the minimum value among them, as the global minimum health level. ,Right now Global minimum health The introduction of this technology is based on the "weakest link" principle, ensuring that subsequent control decisions always respond to the weakest link in the kiln; and to prevent damage assessment caused by transient noise or localized instantaneous stress fluctuations from affecting long-term cumulative damage. Excessive influence occurs when calculating the moderating factor. hour, Smoothing based on time windows should be used, for example: using past data. Within the time window The exponentially weighted average or rolling average is used as the final value. enter; To calculate the adjustment factor, based on this global minimum health level and preset health sensitivity weights Calculate the moderating factors of health perception ; Health sensitivity weight It is a preset adjustable parameter, for example ; The value determines the adjustment factor. As health The decreasing rate of decay; for example, That is, the square root implies a more moderate decay. That is, a square means that the decline in health is very rapid as soon as it begins to decrease; The calculation formula is: ; To determine the control mode, the global minimum health level is set. With preset maintenance mode threshold For example, 0.8, and the equalization mode threshold. For example, 0.5 is used for comparison to determine the macroscopic control mode that the system should be in; and These are parameters pre-set based on operational strategies, such as risk tolerance. and The specific threshold can be determined based on the economic analysis of kiln overhauls and statistical analysis of historical damage data; specifically, It can be set to a statistical inflection point where the damage accumulation rate observed through refractory material fatigue testing begins to show a nonlinear acceleration when the health level falls below this value; and It can be set so that when the health level falls below this value, the system predicts the remaining lifespan without applying control intervention in the equilibrium mode. It will most likely reach the critical health point earlier than the scheduled major overhaul window. Defined from local health To the global control signal and the pattern generation path; through In other words, the weakest link effect ensures the safety of the control strategy, making it always accommodate the most dangerous point; through The nonlinear function that determines health That is, a state variable is transformed into a regulation factor. That is, a control coefficient between 0 and 1, which provides the mathematical basis for subsequent fine-tuning and non-linear adjustment of parameters; through... and The comparison will include continuous health scores. For example, 0.75 is classified as a discrete control mode, such as an equilibrium mode, which provides a decision-making basis for the subsequent implementation of hierarchical and goal-oriented control strategies.

[0032] Example 9: Determine the control mode, including: When the global minimum health level is greater than the maintenance mode threshold, it is determined to be in maintenance mode; When the global minimum health is less than or equal to the maintenance mode threshold and greater than the balance mode threshold, it is determined to be in balance mode; When the global minimum health level is less than or equal to the equilibrium mode threshold, it is determined to be in protection mode.

[0033] According to the system of Embodiment 8, the specific hierarchical logic for determining the control mode is described in detail in this embodiment; this logic is based on the global minimum health level. With two preset thresholds That is, the maintenance mode threshold, for example, 0.8, and That is, the comparison of the equalization mode threshold, for example, 0.5: When the global minimum health For example The kiln is now in good health and in its youth stage; When the global minimum health For example The kiln was determined to be in equilibrium mode; at this point, the kiln had already suffered some damage and entered its middle age. When the global minimum health For example The kiln is now in protection mode; at this point, the kiln's health is low, and it enters a period of decline. A clear, three-stage macro-control strategy based on the kiln's life cycle stages is provided. This enables the system to execute different, even conflicting, control objectives depending on the kiln's health stage: in maintenance mode, the system control objective prioritizes product uniformity, allowing for more aggressive control; in balancing mode, the system control objective strikes a balance between control accuracy and asset lifespan; and in protection mode, the system control objective will proactively sacrifice some temperature tracking accuracy to maximize the remaining service life and trigger predictive maintenance alarms.

[0034] Example 10: The parameter adjustment module is used to adjust control parameters, including: Determine the final adjustment factor based on the control mode; Among them, when the control mode is maintenance mode or balance mode, the final adjustment factor is set to be equal to the health perception adjustment factor; When the control mode is protection mode, the final adjustment factor is set to the product of the health perception adjustment factor and the preset protection coefficient. Based on the final adjustment factor, the proportional gain and integral time of the PID control loop in the PLC are corrected.

[0035] According to the system of Embodiment 9, in this embodiment, the complete logic of how the parameter adjustment module adjusts the control parameters based on the control mode is described in detail; This module determines the final adjustment factor used for PID parameter correction. This determination process is based on the control mode determined by the state determination module: When the control mode is maintenance mode or balancing mode, set the final adjustment factor. Equal to health perception modulators ;Right now: ; When the control mode is protection mode, set the final adjustment factor. Health perception regulators With preset protection coefficient The product of; that is: ; Preset protection factor It is Preset parameters, such as The purpose of this parameter is to ensure that when the kiln enters the protection mode, On top of the already reduced level, an additional penalty coefficient is applied to enforce a more conservative control strategy. The value can be determined through simulation testing; the specific method is as follows: in the simulation environment, put the system in protection mode, apply the strongest operating condition disturbance that the system design can withstand, and reduce it iteratively. The value is calculated up to the thermal shock index caused by the adjusted control system response. It remains below the Level 1 risk threshold. The value is the calibrated preset protection coefficient; The parameter adjustment module is further used to respond to the instantaneous damage risk classification determined in Example 5; when the classification is determined to be a level 2 risk, the system will execute emergency avoidance logic, and forcibly set the final adjustment factor regardless of the current control mode. Equal to a preset emergency protection coefficient This coefficient Independent of and The system will only continue to execute the above control-mode-based approach when the risk level is classified as safe or Level 1 risk. Computational logic; Based on this final regulating factor Real-time correction of the proportional gain of the PID control loop in the PLC and points time ; To ensure the robustness of the control system, especially in terms of health Under extreme operating conditions approaching 0, the final adjustment factor Perform lower limit clamping: ,in, It is a preset minimum key. Section factor, to prevent in calculation A division by zero error occurred; Corrected proportional gain: ; Corrected integration time: ; in, and These are the original baseline values ​​set by the engineers; A bridge was built to transform macro-level strategies (patterns) into micro-level execution parameters; through The design successfully transforms the macro-control mode defined in Example 9—namely, maintenance, balancing, and protection—into a strategy for controlling PID parameters. , Specific, quantitative adjustments; especially in protection mode, through The introduction of the coefficient achieves a dual attenuation of control intensity. ,in and All are less than 1, which leads to Significantly suppressed The temperature tracking accuracy is significantly increased; this design actively sacrifices temperature tracking accuracy, i.e., slower response, but maximizes the smoothness of control behavior to achieve the control objective of maximizing the remaining service life.

[0036] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A PLC-based intelligent control system for a glass production process, characterized in that, The method comprises the following steps: a data acquisition module is configured to acquire control output data and measurement data of the kiln in real time; a first processing module is configured to predict the internal temperature field distribution of the refractory material at the key section of the kiln in real time based on the control output data and the measurement data, and by calling a pre-trained kiln thermodynamic reduced-order surrogate model; a second processing module is configured to calculate the instantaneous thermal stress at the key section based on the internal temperature field distribution and in combination with the pre-set physical properties of the refractory material; a damage assessment module is configured to track the historical fluctuation data of the instantaneous thermal stress, and determine the cumulative fatigue damage degree at the key section based on a pre-set fatigue analysis algorithm and a linear fatigue cumulative damage theory; a state determination module is configured to convert the cumulative fatigue damage degree into a kiln health degree, and determine a health-aware adjustment factor and a control mode in response to the kiln health degree; a parameter adjustment module is configured to adjust the control parameters of the PLC based on the health-aware adjustment factor and the control mode.

2. The PLC-based intelligent control system for glass production processes according to claim 1, characterized in that, The data acquisition module is configured to acquire the control output data and the measurement data, wherein: the control output data includes the fuel supply valve opening degree and the electric heating power; the measurement data includes the multi-point temperature inside the kiln.

3. The PLC-based intelligent control system for glass production processes according to claim 1, characterized in that, The second processing module is configured to calculate the instantaneous thermal stress, comprising: calling the internal temperature field distribution, and the elastic modulus, the thermal expansion coefficient and the stress-free reference temperature of the refractory material pre-calibrated; determining the instantaneous thermal stress based on the theory of thermoelasticity.

4. The PLC-based intelligent control system for glass production processes according to claim 1, characterized in that, The second processing module is further configured to: calculate a thermal shock index based on the time variation rate of the internal temperature field distribution and the pre-set material safety limit value; determine the instantaneous damage risk classification at the key section by comparing the thermal shock index with the pre-set risk threshold value.

5. The PLC-based intelligent control system for glass production processes according to claim 4, characterized in that, Determining the instantaneous damage risk classification comprises: determining as safe when the thermal shock index is less than or equal to a first risk threshold value; determining as a first-level risk when the thermal shock index is greater than the first risk threshold value and less than or equal to a second risk threshold value; determining as a second-level risk when the thermal shock index is greater than the second risk threshold value.

6. The PLC-based intelligent control system for glass production processes according to claim 1, characterized in that, The damage assessment module is configured to determine the cumulative fatigue damage degree, comprising: processing the historical fluctuation data of the instantaneous thermal stress to identify the stress amplitude and the occurrence frequency by using a pre-set fatigue analysis algorithm; determining the total number of allowable failure cycles based on the stress amplitude and by calling a pre-set material S-N curve; determining the cumulative fatigue damage degree based on the occurrence frequency and the total number of allowable failure cycles, and in accordance with the Palmgren-Miner linear fatigue cumulative damage theory.

7. The PLC-based intelligent control system for glass production processes according to claim 1, characterized in that, The state determination module is further configured to: predict the remaining service life of the key section based on the kiln health degree and the pre-set total design life of the kiln.

8. The PLC-based intelligent control system for glass production processes according to claim 1, characterized in that, The state determination module is configured to determine the factor and the mode, comprising: determining the lowest kiln health degree in all key sections as a global minimum health degree; calculating the health-aware adjustment factor based on the global minimum health degree and a pre-set health degree sensitivity weight; determining the control mode by comparing the global minimum health degree with a pre-set maintenance mode threshold value and an equilibrium mode threshold value.

9. The PLC-based intelligent control system for glass production processes according to claim 8, characterized in that, Determining the control mode comprises: determining as the maintenance mode when the global minimum health degree is greater than the maintenance mode threshold value; When the global minimum health degree is less than or equal to the maintenance mode threshold and greater than the balance mode threshold, the balance mode is determined; When the global minimum health degree is less than or equal to the balance mode threshold, the protection mode is determined.

10. The PLC-based intelligent control system for glass production processes according to claim 9, characterized in that, The parameter adjustment module is configured to adjust the control parameter, comprising: determining a final adjustment factor according to the control mode; wherein, when the control mode is the maintenance mode or the balance mode, the final adjustment factor is set to be equal to the health-aware adjustment factor; wherein, when the control mode is the protection mode, the final adjustment factor is set to be a product of the health-aware adjustment factor and a preset protection coefficient; correcting the proportional gain and the integral time of the PID control loop in the PLC based on the final adjustment factor.

Images