Perlite production process intelligent control system based on big data

Abstract

The invention relates to the technical field of production control, in particular to a perlite production process intelligent control system based on big data, which comprises a temperature monitoring module, a particle size rate fitting module, a correlation analysis module, a parameter adjustment optimization module and an instruction execution module. According to the method, the dynamic trend of the production process can be pre-judged by synchronously monitoring and analyzing the change rate and trend of the temperature of the inner core and the surface of the perlite particle instead of only performing lag compensation on the generated result, and on the basis, the correlation between the temperature trend and the change rate of the particle size is further established; indirect prediction and fine tracking of key attributes of products are achieved by fitting future particle size changes and calculating deviations with measured values, and the method achieves accurate tracking of the key attributes of the products by analyzing the trend consistency of control parameter changes and particle size change rate deviations for two key control parameters of heating power and airflow velocity. And dynamically judging the strength relationship of the influence of each parameter on the product quality under the current working condition.

Description

Technical Field

[0001] This invention relates to the field of production control technology, and in particular to an intelligent control system for perlite production process based on big data. Background Technology

[0002] The field of production control technology involves the monitoring, regulation, and optimization of production processes, aiming to improve production efficiency, reduce costs, ensure product quality, and achieve automated and intelligent control. This includes automatic control of production processes, data acquisition and processing, equipment fault diagnosis, process optimization, and production scheduling.

[0003] In particular, perlite production process control refers to the use of feedback control based on set parameters to adjust the heating and expansion processes of raw materials in perlite production in order to ensure the quality and stability of the final product.

[0004] Existing technologies primarily rely on feedback control based on set parameters for adjustment. This approach has significant limitations in addressing dynamic changes during perlite production. Its control logic is based on static, pre-set thresholds and cannot adaptively adjust to batch-to-batch differences in raw materials, fluctuations in ambient temperature and humidity, or gradual changes in equipment status. When deviations occur in the production process, control adjustments are only passively triggered after the deviation is detected, resulting in a delayed response. This can easily lead to parameter overshoot and oscillations, causing unnecessary fluctuations in product quality between batches. More importantly, this control method typically treats each control variable in isolation. For example, the adjustment of heating power and airflow velocity is separate, lacking a comprehensive consideration of the interaction between multiple parameters. In actual production scenarios, improper airflow may cause temperature changes, but the system may simply adjust the heating power without identifying the root cause, leading to poor control performance or even negative impacts. Ultimately, it is difficult to reliably guarantee product uniformity and stability. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing an intelligent control system for the perlite production process based on big data.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A big data-based intelligent control system for perlite production includes: The temperature monitoring module acquires the rate of temperature change of perlite in the expansion channel area at adjacent time points during each production stage, judges the temperature change trend, and outputs the temperature trend judgment result. The particle size rate fitting module obtains the particle size change rate of perlite in the expansion channel region at adjacent time points in each production stage, fits the future particle size change with reference to the temperature trend judgment result, calculates the deviation between the fitted value and the subsequent measured particle size change rate, and outputs the perlite particle size change rate deviation set. The correlation analysis module obtains the heater output power and airflow velocity in the expansion channel area of ​​each production stage, constructs a set of thermal energy output parameters, determines the correlation between the thermal energy output parameter set and the set of particle size change rate deviations, and obtains the correlation analysis results. The parameter adjustment and optimization module acquires big data records of the perlite expansion channel area operation process during a specified period, combines the correlation analysis results with the thermal energy output parameter set to optimize the control parameters, and outputs control commands. The instruction execution module adjusts the output power of the heater and the airflow velocity in the expansion channel area according to the control instructions to obtain the production process control results.

[0007] As a further aspect of the present invention, the temperature trend judgment result includes the temperature change direction, the temperature change rate range, and the temperature change continuity state; the perlite particle size change rate deviation set specifically includes a fitted value sequence, a measured value sequence, and a deviation value sequence; the correlation analysis result includes a particle size change item, a heater power change item, and an airflow velocity change item; the control command includes a power adjustment command, an airflow adjustment command, and a control execution time; and the production process control result specifically refers to heating status information, airflow status information, and adjusted parameter records.

[0008] As a further aspect of the present invention, the temperature monitoring module includes: The temperature acquisition submodule, by deploying thermocouples and infrared thermal imaging equipment, acquires the surface and core temperatures of the particles at the beginning of each production stage in the inlet, middle and outlet sections of the expansion channel area where the perlite enters after heating treatment, and merges them into a stage-specific temperature data group according to the time point. The temperature change rate calculation submodule constructs a time series data stream based on the staged temperature data group, performs differential and normalization processing on the time series data stream, calculates the ratio of temperature change value to time interval between adjacent time points within a specified time, and obtains the temperature change rate sequence. The temperature trend determination submodule calculates the slope of the temperature change trend based on the temperature change rate sequence, determines the continuity of the temperature change trend slope in each time period, marks continuous rise, continuous fall and fluctuation states, divides the trend type into segments, and obtains the temperature trend determination result.

[0009] As a further aspect of the present invention, the particle size rate fitting module includes: The particle size data acquisition submodule acquires the average particle size value of perlite particles at each location in the expansion channel region at the time point synchronized with temperature acquisition, and establishes a phased average particle size data sequence. The rate of change calculation submodule calculates the ratio of the difference in particle size values ​​at adjacent time points to the time interval in the phased average particle size data sequence, and generates a particle size change rate data sequence. The trend fitting submodule uses the slope change continuity of the temperature change rate in the temperature trend judgment result as a reference standard to perform extrapolation of the trend of the particle size change rate data sequence and predict the particle size change rate in the future time period. It obtains the fitted value and compares it point by point with the subsequent measured particle size change rate data to obtain the perlite particle size change rate deviation set.

[0010] As a further aspect of the present invention, the correlation analysis module includes: The thermal energy parameter acquisition submodule acquires the heater output power and airflow velocity of the expansion channel area in each production stage based on the time point synchronized with temperature and particle size monitoring, organizes and pairs them in chronological order, and generates heating and airflow time series parameter groups. The parameter set construction submodule aligns the heating and airflow time series parameter set with the particle size change rate data in time, constructs a set of pairing relationships between output power and airflow velocity within a continuous time period, and generates a thermal energy output parameter set. The trend correlation determination submodule calls the set of thermal energy output parameters and the set of deviations in the perlite particle size change rate to compare and analyze the consistency and fluctuation direction of particle size change with heating and airflow direction, and obtains the correlation analysis results.

[0011] As a further aspect of the present invention, the process of comparing and analyzing the relationship between particle size change and the consistency and fluctuation direction of heating and airflow direction specifically involves extracting the change direction sequence of heater output power and airflow velocity from the set of thermal energy output parameters, aligning it with the deviation direction sequence from the set of deviations in perlite particle size change rate, pairing and comparing them to generate a matching ratio, and comparing the matching ratio with a preset consistency threshold to determine the correlation relationship.

[0012] As a further aspect of the present invention, the parameter tuning and optimization module includes: The big data sample extraction submodule calls the dataset stored by the production management platform during the operation of the perlite expansion channel area within a specified period, filters the temperature change rate, heater output power adjustment records and airflow velocity control trajectory that are consistent with the current temperature trend, and extracts data windows with consistent time to form a big data sample set. The parameter association and integration submodule calls the big data sample set and combines it with the current thermal energy output parameter set and the association analysis results, and merges them uniformly by time to form a control parameter association input group; The optimization calculation execution submodule inputs the control parameter associated input group into the Bayesian optimization model to optimize the control parameters, selects the optimal control parameters based on the maximum a posteriori probability criterion, and outputs control commands.

[0013] As a further aspect of the present invention, the process of optimizing the control parameters specifically involves inputting the set of control parameters as an input sample set into a Bayesian optimization model, constructing a posterior distribution based on the prior distribution and the likelihood function in the Bayesian optimization model, and using the maximum a posteriori probability criterion to screen the control parameter set and extract the combination of control parameters with the highest posterior probability value.

[0014] As a further aspect of the present invention, the instruction execution module includes: The control command parsing submodule identifies the control items that need to be adjusted in the expansion channel area of ​​the target production stage according to the control command, parses the heater output power adjustment amount and the airflow velocity proportional factor modification value, and generates a control execution parameter group. The parameter adjustment execution submodule calls the control execution parameter group to adjust the output power of the heater in the target area and the airflow velocity proportional factor of the air supply equipment, thus forming an execution control parameter set; The feedback data acquisition submodule, based on the execution control parameter set, collects the heater output power and airflow velocity at each time point in the adjusted expansion channel region to obtain the production process control results.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by simultaneously monitoring and analyzing the rate and trend of temperature change in the core and surface of perlite particles, the dynamic direction of the production process can be predicted, rather than simply compensating for already occurred results with lag. Furthermore, a correlation between temperature trends and the rate of change in particle size is established. By fitting future particle size changes and calculating the deviation from measured values, a core indicator accurately reflecting process stability is obtained. This achieves indirect prediction and refined tracking of key product attributes. For the two key control parameters, heating power and airflow velocity, adjustments are no longer made in isolation, but rather by analyzing the trend of changes in control parameters and the deviation in particle size change rate. By maintaining consistency with the current operating conditions, the system dynamically assesses the relative strengths of each parameter's impact on product quality, providing a clear basis for subsequent optimization decisions. Combining historical big data samples similar to the current operating conditions with real-time correlation analysis results, the system employs Bayesian optimization to find the combination of heating power and airflow velocity that maximizes the posterior probability, ensuring the global optimality and foresight of the control commands. Finally, the adjusted actual production data is used as feedback input, forming a closed-loop control process that continuously learns and optimizes. This significantly improves the adaptability and control accuracy to complex operating conditions, effectively guaranteeing the high consistency and stability of perlite product quality. Attached Figure Description

[0016] Figure 1 This is a system flowchart of the present invention; Figure 2 This is a flowchart of the temperature monitoring module of the present invention; Figure 3 This is a flowchart of the particle size and rate fitting module of the present invention; Figure 4 This is a flowchart of the correlation analysis module of the present invention; Figure 5 This is a flowchart of the parameter tuning and optimization module of the present invention; Figure 6 This is a flowchart of the instruction execution module of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0018] Please see Figure 1 The intelligent control system for perlite production based on big data includes: The temperature monitoring module acquires the rate of temperature change of perlite in the expansion channel area at adjacent time points during each production stage, judges the temperature change trend, and outputs the temperature trend judgment result. The particle size rate fitting module obtains the particle size change rate of perlite at adjacent time points in the expansion channel area at each production stage, and fits the future particle size change based on the temperature trend judgment result. It calculates the deviation between the fitted value and the subsequent measured particle size change rate and outputs the perlite particle size change rate deviation set. The correlation analysis module obtains the heater output power and airflow velocity in the expansion channel area of ​​each production stage, constructs a set of thermal energy output parameters, judges the correlation between the thermal energy output parameter set and the particle size change rate deviation set, and obtains the correlation analysis results. The parameter tuning and optimization module acquires big data records of the perlite expansion channel area's operation process during a specified period, combines them with correlation analysis results and thermal energy output parameter sets to optimize control parameters, and outputs control commands. The instruction execution module adjusts the output power and airflow velocity of the heater in the expansion channel area according to the control instructions to obtain the production process control results. The temperature trend judgment results include the direction of temperature change, the range of temperature change rate, and the continuity of temperature change. The perlite particle size change rate deviation set specifically includes the fitted value sequence, the measured value sequence, and the deviation value sequence. The correlation analysis results include particle size change items, heater power change items, and airflow velocity change items. The control commands include power adjustment commands, airflow adjustment commands, and control execution time. The production process control results specifically refer to heating status information, airflow status information, and adjusted parameter records.

[0019] Please see Figure 2 The temperature monitoring module includes: The temperature acquisition submodule, by deploying thermocouples and infrared thermal imaging equipment, acquires the surface and core temperatures of the particles at the beginning of each production stage in the inlet, middle and outlet sections of the expansion channel area where the perlite enters after heating treatment, and merges them into a stage-specific temperature data group according to the time point. Thermocouples and infrared thermal imaging devices are deployed at the entrance, middle, and exit sections of the perlite expansion channel area. These devices are triggered at multiple preset time points during each production stage to collect data. Thermocouples acquire the core temperature of perlite particles through direct contact measurement, while infrared thermal imaging devices acquire the surface temperature of perlite particles through non-contact scanning. At each preset time point, a pair of core temperature and surface temperature readings are simultaneously acquired at each monitoring location. The core temperature and surface temperature values ​​acquired at all collection points at the same time are paired and marked with the corresponding time and location information, forming a data entry containing time, location, core temperature, and surface temperature. After all production stages, all locations, and all time points are collected, the data entries are merged and arranged according to the chronological order of the time points recorded in the data entries, forming a stage-specific temperature data group.

[0020] The temperature change rate calculation submodule constructs a time series data stream based on a set of phased temperature data, performs differential and normalization processing on the time series data stream, calculates the ratio of the temperature change value to the time interval between adjacent time points within a specified time period, and obtains the temperature change rate sequence. The particle core temperature and particle surface temperature are extracted separately from the phased temperature data set to construct independent time series data streams for particle core temperature and particle surface temperature. Differential processing is performed on each time series data stream. The specific action of differential processing is to calculate the difference between the temperature value at each time point in the sequence and the temperature value at the previous adjacent time point in the sequence to obtain a series of temperature change values. Then, all calculated temperature change values ​​are normalized. The specific operation of normalization is to divide each actual calculated temperature change value by a maximum theoretical temperature change value determined jointly based on the upper limit of safe operation of the equipment and the physical properties of perlite material, thereby transforming all temperature change values ​​into a unified interval. Next, the ratio of the temperature change value between adjacent time points within a specified time is calculated. The calculation process is to divide the unnormalized temperature change value after differential processing by the corresponding time interval length. The ratios calculated for all time points are arranged in chronological order to obtain a temperature change rate sequence.

[0021] The temperature trend determination submodule calculates the slope of the temperature change trend based on the temperature change rate sequence, judges the continuity of the temperature change trend slope in each time period, marks the continuous rise, continuous fall and fluctuation states, divides the trend type segments, and obtains the temperature trend determination results. The input temperature change rate sequence is processed by first calculating the slope of the line segment formed by multiple consecutive data points in the sequence, and using this calculated slope as the temperature change trend slope. Then, the continuity of the temperature change trend slope within each time period is judged based on a pre-set standard. This standard is determined by statistical analysis of temperature data from a large number of stable production batches to identify the slope fluctuation range during normal production. A condition is defined as a trend slope that is positive for three or more consecutive time periods and consistently exceeds the upward trend judgment benchmark of 0.05 determined based on historical data statistical analysis. If the trend slope of three or more consecutive time periods is negative and the absolute value of the trend slope is consistently greater than the downward trend judgment benchmark of 0.05 determined based on historical data statistical analysis, then the current time period is marked as a continuous upward trend. All other situations besides continuous upward and continuous downward trends, such as alternating positive and negative trend slopes or trend slope values ​​fluctuating within the range of -0.05 to +0.05, are marked as fluctuating trends. Finally, time periods marked with the same continuity are merged to divide different trend type segments and obtain the temperature trend judgment result.

[0022] Please see Figure 3 The particle size and rate fitting module includes: The particle size data acquisition submodule acquires the average particle size value of perlite particles at each location in the expansion channel region at the time point synchronized with temperature acquisition, and establishes a phased average particle size data sequence. At a time point completely synchronized with temperature acquisition, laser particle size analyzers deployed at the entrance, middle, and exit sections of the expansion channel area are activated to obtain the average particle size value of perlite particles at each location in the expansion channel area. The laser particle size analyzer performs instantaneous measurements on all perlite particles within a specified field of view at each monitoring location and outputs the volume average particle size value of all measured perlite particles within the current field of view. The process of obtaining the average particle size value is repeated at multiple preset time points in each production stage. The average particle size values ​​obtained at all synchronous time points from all locations are arranged in chronological order to establish a staged average particle size data sequence.

[0023] The rate of change calculation submodule calculates the ratio of the difference in particle size values ​​at adjacent time points to the time interval in the phased average particle size data sequence, and generates a particle size change rate data sequence. The particle size values ​​at adjacent time points in the phased average particle size data sequence are used to calculate the particle size difference. The calculation process involves subtracting the average particle size value of the previous time point from the average particle size value of the later time point. Then, the calculated particle size difference is compared with the corresponding time interval by dividing the particle size difference by the time interval between the two time points. This process of calculating the difference and ratio is repeated for all adjacent time points in the entire phased average particle size data sequence, generating a particle size change rate data sequence corresponding to the time points of the original data sequence.

[0024] The trend fitting submodule uses the slope change continuity of the temperature change rate in the temperature trend judgment result as a reference standard to perform extrapolation of the trend of the particle size change rate data sequence and predict the particle size change rate in the future time period. It obtains the fitted value and compares it point by point with the subsequent measured particle size change rate data to obtain the perlite particle size change rate deviation set. Referring to the segmentation of continuous rise, continuous fall, and fluctuation states in the temperature trend judgment results, the data of the corresponding time period in the particle size change rate data sequence are extrapolated to predict the particle size change rate in future time periods. Specifically, if the temperature trend judgment result shows a continuous rise in a certain time period, a linear regression method is used to fit and extrapolate the particle size change rate data points in the time period. If the judgment result shows a fluctuation state, a moving average method is used to smooth the particle size change rate of the time period and the calculated average value is used as the predicted value. This yields the fitted value of the particle size change rate for future time periods. After the production process enters the future time period, the obtained measured particle size change rate data is compared with the previous fitted value point by point. The comparison process is to subtract the fitted rate from the measured rate at the same time point to obtain the deviation value at each time point. All deviation values ​​are collected to obtain the perlite particle size change rate deviation set.

[0025] Please see Figure 4 The correlation analysis module includes: The thermal energy parameter acquisition submodule acquires the heater output power and airflow velocity of the expansion channel area in each production stage based on the time point synchronized with temperature and particle size monitoring, organizes and pairs them in chronological order, and generates heating and airflow time series parameter groups. At each time point synchronized with temperature and particle size monitoring, the real-time output power data of the heater in the expansion channel area and the airflow velocity data of the corresponding air supply equipment are read from the production control equipment. The heater output power value and airflow velocity value collected at each time point are paired as a data pair and the corresponding timestamp information is added. All the collected data pairs are organized in chronological order to generate a heating and airflow time series parameter group.

[0026] The parameter set construction submodule aligns the heating and airflow time series parameter set with the particle size change rate data in time, constructs a set of pairing relationships between output power and airflow velocity over a continuous time period, and generates a thermal energy output parameter set. The heating and airflow time series parameter set and the particle size change rate data series are aligned based on timestamps. The process uses timestamps as unique identifiers to match and merge data with the same timestamp in the heating and airflow time series parameter set and the particle size change rate data series. Specifically, it iterates through each time point in the particle size change rate data series, searches for data pairs with the same time point marker in the heating and airflow time series parameter set, extracts the output power value and airflow velocity value from the data pair, and places them in the same data structure along with the particle size change rate value at the current time point. Finally, a thermal energy output parameter set that is completely aligned in time and includes the three indicators of heater output power, airflow velocity, and particle size change rate is formed.

[0027] The trend correlation determination submodule calls the thermal energy output parameter set and the perlite particle size change rate deviation set to conduct a comparative analysis on the consistency and fluctuation direction of particle size change with heating and airflow direction, and obtains the correlation analysis results. Specifically, the process of comparing and analyzing the consistency and fluctuation direction of particle size change with heating and airflow direction involves extracting the change direction sequence of heater output power and airflow velocity from the thermal energy output parameter set, aligning it with the deviation direction sequence from the perlite particle size change rate deviation set in time, pairing and comparing them to generate a matching ratio, and comparing the matching ratio with a preset consistency threshold to determine the correlation relationship. The thermal output parameter set and the perlite particle size change rate deviation set are retrieved, and the change direction sequences of heater output power and airflow velocity in the thermal output parameter set are extracted. The extraction process involves calculating the difference between each parameter at adjacent time points. If the difference is positive, the direction is recorded as "rising"; if the difference is negative, it is recorded as "falling"; and if the difference is zero, it is recorded as "stable". The same operation is performed on the perlite particle size change rate deviation set to extract the change direction sequence of the deviation values. Subsequently, the change direction sequences of heater output power, airflow velocity, and deviation values ​​are strictly aligned according to time points and then paired for comparison. The comparison process involves statistically analyzing the change direction of heater output power and the change direction of deviation values ​​at different points in time. The number of times the direction is consistent at the same time point, and the number of times the direction of change of airflow velocity and the direction of change of deviation are consistent at the same time point, are used to divide these two numbers of consistency by the total number of time points to generate the heater power direction matching ratio and the airflow velocity direction matching ratio. Finally, the heater power direction matching ratio and the airflow velocity direction matching ratio are compared with a preset consistency threshold. The preset consistency threshold is set based on the analysis of a large number of high-quality product batches in historical production data to identify the statistical characteristics of parameter synergy. The value is 0.75. If the matching ratio is greater than 0.75, it is determined that there is a strong correlation between the corresponding thermal energy parameter and the particle size change; otherwise, it is a weak correlation. The final correlation analysis result is obtained.

[0028] Please see Figure 5 The parameter tuning and optimization module includes: The big data sample extraction submodule calls the dataset stored by the production management platform during the operation of the perlite expansion channel area within a specified period, filters the temperature change rate, heater output power adjustment records and airflow velocity control trajectory that are consistent with the current temperature trend, and extracts data windows with consistent time to form a big data sample set. The system retrieves a dataset from the production management platform showing the operation of the perlite expansion channel area within a specified period. Based on the current temperature trend, it filters the historical dataset to identify all production period data with the same temperature trend marker. Furthermore, it extracts the corresponding temperature change rate sequence, heater output power adjustment record, and airflow velocity control trajectory from these selected period data. The extracted temperature change rate sequence, heater output power adjustment record, and airflow velocity control trajectory are then aligned with time windows to form a series of time-consistent data windows. All these extracted, time-consistent data windows together constitute a large data sample set.

[0029] The parameter association and integration submodule calls up a large data sample set and combines it with the current thermal energy output parameter set and association analysis results, and merges them uniformly by time to form a control parameter association input group; The big data sample set, the current thermal energy output parameter set, and the correlation analysis results are uniformly merged based on timestamps. Specifically, a new time series data structure is created, with the current production time as the axis. The heater output power and airflow velocity from the current thermal energy output parameter set are filled in as baseline data. Then, the correlation relationship judgment results obtained from the correlation analysis are attached as a label to the corresponding time point. Finally, the big data sample set is traversed, and the control parameter sequence from the sample data that is closest to the current operating condition is integrated into the new data structure as a reference sequence to form a control parameter correlation input group.

[0030] The optimization calculation execution submodule inputs the control parameter associated input group into the Bayesian optimization model for control parameter optimization. It selects the optimal control parameters based on the maximum a posteriori probability criterion and outputs control commands. Specifically, the control parameter optimization process involves inputting the control parameter set as an input sample set into the Bayesian optimization model. In the Bayesian optimization model, a posterior distribution is constructed based on the prior distribution and the likelihood function. The maximum a posteriori probability criterion is used to filter the control parameter set and extract the combination of control parameters with the highest posterior probability value. The control parameters are associated with the input set and fed into the Bayesian optimization model to initiate the control parameter optimization process. Specifically, the optimization process involves inputting the control parameter set as an input sample set into the Bayesian optimization model. Based on the prior distribution and the likelihood function, the Bayesian optimization model constructs a posterior distribution. The maximum a posteriori probability criterion is used to filter the control parameter set, extracting the combination of control parameters with the highest posterior probability value. The calculation of the posterior probability follows Bayes' theorem, and its mathematical expression is: In the current formula, Represents the posterior probability distribution, that is, given the observed data Under these conditions, control parameters The probability distribution, The control parameter association input group, represented by the observed data, is a data set that integrates current operating conditions and historical references, formed by previous steps. This represents a combination of control parameters to be optimized, specifically a particular combination of values ​​for the heater output power and the airflow velocity. It is the likelihood function, representing the likelihood under the assumption of using a certain set of control parameters. During production, this set of observation data appeared. The probability, the likelihood function, is established based on fitting historical production data and describes the physical response characteristics of the production process. It is a prior probability distribution, representing the probability distribution obtained when acquiring the current observation data. Previously, the analysis of large datasets led to the determination of optimal control parameters. Initial probability judgment, symbol It indicates a direct proportional relationship; the logic of the entire formula lies in utilizing current observation data. Update the optimal control parameters The initial judgment is to find the posterior probability that makes the optimization goal. Maximize the combination of parameters Mathematically represented as ,in This represents the combination of control parameters that has the maximum posterior probability. This represents finding the parameter that maximizes the subsequent function value. Value operations, for example, the model selects two candidate combinations of control parameters for comparison, candidate combination one Given {heater output power = 155 kW, airflow velocity = 2.6 m / s}, candidate combination two. Given {heater output power = 160 kW, airflow velocity = 2.4 m / s}, based on the analysis of a large dataset, the prior probabilities of the two candidate combinations are as follows: and This indicates that historical data favors candidate combination one, and subsequently, based on current observational data... Calculate the likelihood function and obtain and This indicates that the actual response under the current operating condition is more consistent with the physical model of candidate combination two. Next, the posterior probabilities of the two candidate combinations are calculated, and the posterior probability of candidate combination one is compared with... The posterior probability of candidate combination two is directly proportional to... They are directly proportional. Comparing the two results, because... Therefore, candidate combination two It has a higher posterior probability, therefore the model selects based on the maximum a posteriori probability criterion. The optimal control parameters are {heater output power = 160 kW, airflow velocity = 2.4 m / s}, and the control command is output.

[0031] Please see Figure 6 The instruction execution module includes: The control command parsing submodule identifies the control items that need to be adjusted in the expansion channel area of ​​the target production stage based on the control command, parses the heater output power adjustment amount and the airflow velocity proportional factor modification value, and generates a control execution parameter group. The received control commands are parsed to identify the target production stage and the expansion channel area requiring adjustment, and the heater output power adjustment and airflow velocity proportional factor modification values ​​are extracted. The final setpoint for the heater output power is determined using the formula... Calculations show that, in this formula, This represents the target output power that will be set. This represents the current heater output power value read in real time from the production control equipment. This represents the power adjustment amount parsed from the control command, where The subscript "target" represents the target value, and the subscript "current" represents the current value. The symbols represent the changes, the subscript cmd indicates that it comes from an instruction, and the final proportionality factor of the airflow velocity is determined by the formula. Calculations show that, in this formula, This represents the target airflow velocity scaling factor to be set. This represents the current scaling factor setting value. This represents the percentage-based scaling factor modification value parsed from the control command, where... Represents the flow rate proportionality factor. The symbol represents a proportional change. For example, the optimal parameters output by the previous optimization calculation were {heater output power = 160 kW, airflow velocity = 2.4 m / s}, while the parameters currently read from the device are {heater output power}. =150 kW, airflow velocity scaling factor =2.5}, then the adjustment amount parsed from the control command is Kilowatts, airflow velocity scaling factor modification value Then, calculations are performed to determine the target output power. Target airflow velocity scaling factor The calculated kilowatts and Combined together, they generate a set of control execution parameters.

[0032] The parameter adjustment execution submodule calls the control execution parameter group to adjust the output power of the heater in the target area and the airflow velocity proportional factor of the air supply equipment, thus forming the execution control parameter set; The target output power value of the heater in the target area of ​​the control execution parameter group is sent to the corresponding address of the programmable logic controller (PLC) responsible for controlling the heater via the industrial bus protocol. After receiving the new power value, the PLC immediately adjusts the output signal to change the actual output power of the heater. At the same time, the target airflow velocity proportional factor value in the control execution parameter group is sent to the frequency converter controlling the air supply equipment. The frequency converter adjusts its output frequency according to the received new proportional factor value, thereby changing the speed of the fan motor and ultimately regulating the airflow velocity of the air supply equipment. All adjustment commands are issued and executed and are recorded, forming the execution control parameter set.

[0033] The feedback data acquisition submodule, based on the set of execution control parameters, collects the heater output power and airflow velocity at each time point in the expansion channel region after adjustment, and obtains the production process control results. After the execution of instructions from the control parameter set, and following a pre-calibrated delay time based on the equipment's response characteristics, the data acquisition process is initiated. Adjusted production process data is collected from sensors and controllers in the expansion channel area. At each time point, the actual output power reading of the heater in the corresponding area and the actual airflow velocity measured by the anemometer are synchronously collected and recorded. The collected actual output power reading and actual airflow velocity are paired and timestamped to form a data stream reflecting the actual response of the equipment after the control instructions are executed. This data stream is the production process control result. This production process control result is used as the initial state data for the next round of control optimization, thus forming a closed-loop feedback. In this closed-loop feedback, the control effect is verified, and the optimization model is dynamically corrected based on the actual response, ultimately achieving continuous adaptive intelligent control of the perlite production process to ensure product quality stability and maximize production efficiency.

[0034] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A big data-based intelligent control system for perlite production process, characterized in that: The system includes: The temperature monitoring module acquires the rate of temperature change of perlite in the expansion channel area at adjacent time points during each production stage, judges the temperature change trend, and outputs the temperature trend judgment result. The particle size rate fitting module obtains the particle size change rate of perlite in the expansion channel region at adjacent time points in each production stage, fits the future particle size change with reference to the temperature trend judgment result, calculates the deviation between the fitted value and the subsequent measured particle size change rate, and outputs the perlite particle size change rate deviation set. The correlation analysis module obtains the heater output power and airflow velocity in the expansion channel area of ​​each production stage, constructs a set of thermal energy output parameters, determines the correlation between the thermal energy output parameter set and the set of particle size change rate deviations, and obtains the correlation analysis results. The parameter adjustment and optimization module acquires big data records of the perlite expansion channel area operation process during a specified period, combines the correlation analysis results with the thermal energy output parameter set to optimize the control parameters, and outputs control commands. The instruction execution module adjusts the output power of the heater and the airflow velocity in the expansion channel area according to the control instructions to obtain the production process control results.

2. The intelligent control system for perlite production process based on big data as described in claim 1, characterized in that, The temperature trend judgment results include the direction of temperature change, the range of temperature change rate, and the continuity of temperature change. The perlite particle size change rate deviation set specifically includes the fitted value sequence, the measured value sequence, and the deviation value sequence. The correlation analysis results include particle size change items, heater power change items, and airflow velocity change items. The control commands include power adjustment commands, airflow adjustment commands, and control execution time. The production process control results specifically refer to heating status information, airflow status information, and adjusted parameter records.

3. The intelligent control system for perlite production process based on big data as described in claim 1, characterized in that, The temperature monitoring module includes: The temperature acquisition submodule, by deploying thermocouples and infrared thermal imaging equipment, acquires the surface and core temperatures of the particles at the beginning of each production stage in the inlet, middle and outlet sections of the expansion channel area where the perlite enters after heating treatment, and merges them into a stage-specific temperature data group according to the time point. The temperature change rate calculation submodule constructs a time series data stream based on the staged temperature data group, performs differential and normalization processing on the time series data stream, calculates the ratio of temperature change value to time interval between adjacent time points within a specified time, and obtains the temperature change rate sequence. The temperature trend determination submodule calculates the slope of the temperature change trend based on the temperature change rate sequence, determines the continuity of the temperature change trend slope in each time period, marks continuous rise, continuous fall and fluctuation states, divides the trend type into segments, and obtains the temperature trend determination result.

4. The intelligent control system for perlite production process based on big data according to claim 3, characterized in that, The particle size rate fitting module includes: The particle size data acquisition submodule acquires the average particle size value of perlite particles at each location in the expansion channel region at the time point synchronized with temperature acquisition, and establishes a phased average particle size data sequence. The rate of change calculation submodule calculates the ratio of the difference in particle size values ​​at adjacent time points to the time interval in the phased average particle size data sequence, and generates a particle size change rate data sequence. The trend fitting submodule uses the slope change continuity of the temperature change rate in the temperature trend judgment result as a reference standard to perform extrapolation of the trend of the particle size change rate data sequence and predict the particle size change rate in the future time period. It obtains the fitted value and compares it point by point with the subsequent measured particle size change rate data to obtain the perlite particle size change rate deviation set.

5. The intelligent control system for perlite production process based on big data according to claim 4, characterized in that, The correlation analysis module includes: The thermal energy parameter acquisition submodule acquires the heater output power and airflow velocity of the expansion channel area in each production stage based on the time point synchronized with temperature and particle size monitoring, organizes and pairs them in chronological order, and generates heating and airflow time series parameter groups. The parameter set construction submodule aligns the heating and airflow time series parameter set with the particle size change rate data in time, constructs a set of pairing relationships between output power and airflow velocity within a continuous time period, and generates a thermal energy output parameter set. The trend correlation determination submodule calls the set of thermal energy output parameters and the set of deviations in the perlite particle size change rate to compare and analyze the consistency and fluctuation direction of particle size change with heating and airflow direction, and obtains the correlation analysis results.

6. The intelligent control system for perlite production process based on big data according to claim 5, characterized in that, The process of comparing and analyzing the consistency and fluctuation direction of particle size change with heating and airflow direction specifically involves extracting the change direction sequence of heater output power and airflow velocity from the set of thermal energy output parameters, aligning it with the deviation direction sequence from the set of deviations in perlite particle size change rate, pairing and comparing them to generate a matching ratio, and comparing the matching ratio with a preset consistency threshold to determine the correlation.

7. The intelligent control system for perlite production process based on big data according to claim 5, characterized in that, The parameter tuning and optimization module includes: The big data sample extraction submodule calls the dataset stored by the production management platform during the operation of the perlite expansion channel area within a specified period, filters the temperature change rate, heater output power adjustment records and airflow velocity control trajectory that are consistent with the current temperature trend, and extracts data windows with consistent time to form a big data sample set. The parameter association and integration submodule calls the big data sample set and combines it with the current thermal energy output parameter set and the association analysis results, and merges them uniformly by time to form a control parameter association input group; The optimization calculation execution submodule inputs the control parameter associated input group into the Bayesian optimization model to optimize the control parameters, selects the optimal control parameters based on the maximum a posteriori probability criterion, and outputs control commands.

8. The intelligent control system for perlite production process based on big data according to claim 7, characterized in that, The process of optimizing the control parameters specifically involves inputting the set of control parameters as an input sample set into a Bayesian optimization model. In the Bayesian optimization model, a posterior distribution is constructed based on the prior distribution and the likelihood function. The maximum a posteriori probability criterion is used to filter the control parameter set and extract the combination of control parameters with the highest posterior probability value.

9. The intelligent control system for perlite production process based on big data according to claim 7, characterized in that, The instruction execution module includes: The control command parsing submodule identifies the control items that need to be adjusted in the expansion channel area of ​​the target production stage according to the control command, parses the heater output power adjustment amount and the airflow velocity proportional factor modification value, and generates a control execution parameter group. The parameter adjustment execution submodule calls the control execution parameter group to adjust the output power of the heater in the target area and the airflow velocity proportional factor of the air supply equipment, thus forming an execution control parameter set; The feedback data acquisition submodule, based on the execution control parameter set, collects the heater output power and airflow velocity at each time point in the adjusted expansion channel region to obtain the production process control results.

Images