PID-based allulose crystallization reactor temperature control method and system

Abstract

The application relates to the technical field of crystallizer temperature control, in particular to an allulose crystallization reactor temperature control method and system based on PID, which comprises the following steps: according to temperature data collected in real time at multiple positions in a crystallizer, extracting temperature fluctuation characteristics and temperature distribution characteristics from the time dimension and the space dimension, and analyzing the change trend and the mean level thereof in continuous multiple monitoring time periods; and combining the synchronous change relationship between the temperature data and the solution concentration data collected in real time at multiple positions in the crystallizer, constructing comprehensive abnormal characteristic values, using the comprehensive abnormal characteristic values to dynamically adjust the proportional term parameters of a PID controller, and using the adjusted PID controller to output an adjusting instruction to a temperature control execution mechanism, so as to control the temperature of the crystallization reactor. The application aims to effectively cope with the strong time-varying and lagging interference existing in the allulose crystallization process, and realizes accurate and timely control of the temperature of a horizontal cooling crystallizer.

Description

Technical Field

[0001] This application relates to the field of crystallizer temperature control technology, specifically to a PID-based temperature control method and system for an allulose crystallization reactor. Background Technology

[0002] Allulose, a rare ketose, is considered an ideal sucrose substitute due to its low calorie content, health benefits, high sweetness, and similar taste and volume characteristics to sucrose. Currently, the main method for converting high-purity fructose into allulose is through enzymatic processes.

[0003] During production, the high saturation concentration and viscosity of allulose solution lead to difficulties in crystallization and a lack of concentrated crystal distribution. Furthermore, heated allulose is prone to glass transition, and improper temperature control can easily affect crystallization efficiency. Crystallization is a complex process with strong time-varying and hysteresis characteristics. The reactor's heat capacity, heat transfer coefficient, and exothermic effect of crystallization all change as the process progresses. Traditional PID control methods struggle to achieve precise and timely temperature control in the face of such dynamic changes. Summary of the Invention

[0004] To address the aforementioned technical problems, this application provides a PID-based temperature control method and system for an allulose crystallization reactor, the specific technical solution of which is as follows: In a first aspect, one embodiment of this application provides a temperature control method for an allulose crystallization reactor based on PID control, the method comprising the following steps: Based on real-time temperature data collected from multiple locations within the crystallizer during the crystallization process of allulose solution, temperature fluctuation and temperature distribution features are extracted from both temporal and spatial dimensions and fused to obtain a first feature value; a second feature value is obtained based on the changing trend and average level of the first feature value over multiple consecutive monitoring periods. The third characteristic value was obtained based on the synchronous change relationship between temperature data and solution concentration data collected in real time at multiple locations within the crystallizer during the crystallization process of allulose solution; The second and third eigenvalues ​​are weighted and fused to obtain a comprehensive abnormal eigenvalue. Based on this comprehensive abnormal eigenvalue, the proportional term parameter of the PID controller is dynamically adjusted, and the adjusted PID controller outputs a regulation command to the temperature control actuator to regulate the temperature of the crystallization reactor.

[0005] Preferably, temperature fluctuation features are extracted from the time dimension, including a first feature parameter and a second feature parameter; By analyzing the dispersion of temperature data at any location within a single monitoring period, a first characteristic parameter is obtained. The first characteristic parameter is used to reflect the probability of a sudden local temperature change at that location within that monitoring period. The autocorrelation of the average temperature at any location over multiple consecutive monitoring periods is analyzed to obtain a second characteristic parameter. This second characteristic parameter is used to reflect the linear and smooth change characteristics of the overall average temperature of the crystallizer at that location.

[0006] Preferably, temperature distribution features are extracted from the spatial dimension, including a third feature parameter; Calculate the standard deviation of the temperature difference between any location and its surrounding neighborhood locations at each data acquisition time. The average of the standard deviations of all data collection times at any given location within a single monitoring period is used to obtain a third characteristic parameter, which reflects the degree of unevenness in temperature distribution at that location during the monitoring period.

[0007] Preferably, temperature fluctuation features and temperature distribution features are extracted from the time and spatial dimensions and fused to obtain the first feature value, including: The first feature value is positively correlated with the first feature parameter and the third feature parameter, and negatively correlated with the second feature parameter. The larger the first feature value, the more significant the nonlinear fluctuation and uneven local distribution of the temperature data at that location during the monitoring period.

[0008] Preferably, the second feature value is obtained based on the changing trend and average level of the first feature value over multiple consecutive monitoring periods, including: Obtain the slope estimate of the changing trend of the first feature value within multiple consecutive monitoring periods, and perform an exponential mapping on the ratio of the slope estimate to a preset benchmark value of the slope estimate; Calculate the ratio of the mean of the first characteristic value over multiple consecutive monitoring periods to a preset benchmark value of the mean of the first characteristic value; The second characteristic value is obtained by multiplying the calculated result of the mapped exponential function with the ratio result. The second characteristic value is used to reflect the continuous increase or overall abnormality of the temperature fluctuation at that location.

[0009] Preferably, the synchronous change relationship between temperature data and solution concentration data is analyzed, including a fourth characteristic parameter and a fifth characteristic parameter; The correlation coefficient between the average temperature and average concentration at all locations over multiple consecutive monitoring periods is calculated to obtain a fourth characteristic parameter, which is used to reflect the stable and synchronous change characteristics between temperature and concentration. The dispersion of concentration data within a single monitoring period is calculated to obtain a fifth characteristic parameter, which is used to reflect the relative dispersion of concentration data.

[0010] Preferably, the third characteristic value is obtained based on the synchronous change relationship between temperature data and solution concentration data collected in real time at multiple locations within the crystallizer during the allulose solution crystallization process, including: The third characteristic value is positively correlated with the fifth characteristic parameter and negatively correlated with the fourth characteristic parameter; the larger the third characteristic value, the more unstable the dynamic equilibrium between solution concentration and temperature change during crystallization.

[0011] Preferably, dynamically adjusting the proportional term parameter of the PID controller based on the comprehensive abnormal characteristic value includes: The maximum value of the comprehensive abnormal characteristic value calculated from all locations in the previous monitoring period is compared with the preset steady-state lower limit and warning upper limit to determine the proportional term parameter of the PID controller for the current monitoring period. When the maximum value of the comprehensive abnormal characteristic value calculated from all locations in the previous monitoring period is between the steady-state lower limit and the warning upper limit, the proportional term parameter of the PID controller for the current monitoring period is calculated based on the linear mapping relationship.

[0012] Preferably, adjusting the proportional term parameter of the PID controller further includes: When the maximum value is less than or equal to the steady-state lower limit, the proportional parameter of the PID controller for the current monitoring period is set to the steady-state lower limit. When the maximum value is greater than or equal to the warning upper limit value, the proportional parameter of the PID controller for the current monitoring period is set to the safe operating upper limit value.

[0013] Secondly, another embodiment of this application also provides a PID-based allulose crystallization reactor temperature control system, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the PID-based allulose crystallization reactor temperature control method described above.

[0014] This application has at least the following beneficial effects: 1. In response to the strong time-varying and hysteretic interference caused by factors such as uneven stirring, scaling on the inner wall, and decreased heat transfer efficiency during the crystallization process of allulose, this application does not simply report the current temperature deviation, but first deeply explores the intrinsic characteristics of temperature data from the time and space dimensions, and integrates the analysis results of the time and space dimensions, so as to be able to perceive the abnormal temperature fluctuation trend in advance from the data level, providing a forward-looking judgment basis for subsequent active control.

[0015] 2. This application goes beyond simple temperature data, introducing a synchronous relationship between temperature and solution concentration. Under ideal crystallization conditions, both temperature and concentration should exhibit a stable, synchronous decreasing trend. When this dynamic equilibrium is disrupted, it often indicates uneven heat transfer or an unexpected phase transition. By weightedly fusing features characterizing abnormal temperature deterioration with features characterizing the stability of the temperature-concentration dynamic equilibrium, a comprehensive abnormality characteristic value is obtained. This multi-dimensional fusion assessment method avoids the one-sidedness of single-data judgment and can more comprehensively and accurately reflect the overall risk level of the crystallization process deviating from ideal operating conditions.

[0016] 3. This application overcomes the shortcomings of traditional PID control, which has fixed parameters and is difficult to cope with strong time-varying and lag-related disturbances. By using the comprehensive abnormal characteristic value as the basis for dynamically adjusting the proportional term parameter of the PID controller, adaptive optimization of the control strategy is achieved. When the comprehensive abnormal characteristic value is high (high risk of crystallization state deterioration), the system automatically increases the proportional term parameter to speed up the response and suppress temperature fluctuations. When the comprehensive abnormal characteristic value is low (good crystallization state), the system automatically decreases the proportional term parameter to avoid excessive control action that could cause chattering. This dynamic and adaptive parameter adjustment mechanism enables the PID control to always match the actual state of the current crystallization process, thereby effectively coping with various complex disturbances, achieving precise and timely temperature control of the horizontal cooling crystallizer, and ensuring the stability of the crystallization process and product quality. Attached Figure Description

[0017] Figure 1 A flowchart of a PID-based temperature control method for an allulose crystallization reactor provided in one embodiment of this application. Detailed Implementation

[0018] This application provides a PID-based temperature control method for an allulose crystallization reactor, as detailed in one embodiment. Figure 1 The method includes the following steps: S1: Parameter collection within the crystallizer during the crystallization process of allulose solution.

[0019] This application utilizes the reaction of high-purity fructose syrup with D-allulose 3-epimerase to prepare an allulose solution, followed by ion exchange and concentration to obtain a pure allulose solution. Then, a horizontal cooling crystallizer is used to crystallize the pure allulose solution to obtain allulose crystals. The initial temperature during the cooling crystallization process is [temperature missing]. The final temperature is The cycle is In this embodiment, the starting temperature for cooling crystallization is set to... The final temperature is ,by The cooling rate was controlled by a PLC system, and the cooling was continuous for 20 hours. The stirring speed of the horizontal cooling crystallizer was set to 20 r / min.

[0020] Multiple temperature sensors are installed at equal intervals along the axial and circumferential directions on the inner wall of the horizontal cooling crystallizer to collect and monitor the temperature state inside the crystallizer in real time during the crystallization process. Simultaneously, the solution concentration is continuously measured using an online refractometer, thereby synchronously acquiring temperature and concentration data. In this embodiment, eight temperature sensors are used, and the time interval between temperature and concentration data acquisition is 0.5 seconds. To capture short-term, rapid changes in temperature data, a monitoring period of 30 seconds is set, i.e., every 30 seconds constitutes a monitoring period.

[0021] The crystallization process is carried out under saturated concentration conditions, and the concentration data obtained during the cooling crystallization process actually corresponds to the saturated concentration data.

[0022] S2: Crystallization state analysis.

[0023] Under optimal cooling and crystallization conditions, the temperature within a horizontal cooling crystallizer should exhibit a uniform, stable, and controllable decreasing trend. For example, the temperature difference between different locations within the crystallizer should be low, and the temperature changes at each location should be stable, maintaining consistency with the set linear cooling rate. However, interference from various factors such as changes in medium concentration, uneven stirring, and scaling on the inner wall can cause supersaturation changes in certain areas of the solution, leading to short-term sudden temperature fluctuations, resulting in localized hot spots or cold spots. This also disrupts the temperature uniformity of the crystallizer, affecting crystallization efficiency.

[0024] Therefore, this application extracts temperature fluctuation features and temperature distribution features from the time and space dimensions based on the temperature data collected in real time at multiple locations in the crystallizer during the crystallization process of allulose solution, and fuses them to obtain a first feature value. The larger the first feature value, the more significant the nonlinear fluctuation and local uneven distribution features of the temperature data.

[0025] Preferably, this application extracts temperature fluctuation features from the time dimension, including a first feature parameter and a second feature parameter; analyzes the dispersion of temperature data at any location within a single monitoring period to obtain the first feature parameter, which is used to reflect the possibility of a sudden local temperature change at that location within the monitoring period; analyzes the autocorrelation of the average temperature at any location within multiple consecutive monitoring periods to obtain the second feature parameter, which is used to reflect the linear and smooth change characteristics of the overall average temperature of the crystallizer at that location.

[0026] Specifically, this embodiment takes the temperature data collected at a certain collection location during the i-th monitoring period as an example. First, the standard deviation of all temperature data during this period is calculated and denoted as the first characteristic parameter. The standard deviation directly reflects the dispersion of temperature data over that period of time. The larger the value, the greater the likelihood that a sudden local temperature change will occur at that location within a given time period.

[0027] Furthermore, under optimal operating conditions, the temperature inside the crystallizer decreases linearly and smoothly over time, closely matching the cooling rate set in the process. Excessive cooling causes the solution to quickly reach a high degree of supersaturation, producing a large number of tiny crystals, resulting in finer product particles and a wider particle size distribution; while excessively slow cooling prolongs the crystallization cycle and reduces the crystallization efficiency of allulose.

[0028] Therefore, considering the long crystallization cycle and the insignificant temperature drop in the short term, the temperature change characteristics over multiple consecutive monitoring periods are analyzed. This embodiment takes the i-th monitoring period at a specific acquisition location as an example, calculating the mean of all temperature data within the i-th period and the preceding N (N is 12 in this embodiment) monitoring periods; that is, one temperature mean can be calculated for each monitoring period. Then, the autocorrelation coefficient of all temperature mean values ​​obtained from the i-th and preceding N monitoring periods at a specific acquisition location is calculated with a lag order of 1, and denoted as the second characteristic parameter. The result The larger the value, the more pronounced the linear smoothness of the overall average temperature of the crystallizer over multiple consecutive monitoring periods.

[0029] In particular, this embodiment uses edge copying to fill in the missing data in the first few monitoring periods.

[0030] The methods for calculating the standard deviation, autocorrelation coefficient, and marginal replication are all well-known techniques, and their specific processes will not be described in detail in this application.

[0031] In addition, from a spatial perspective, the temperature distribution within the crystallizer is relatively uniform under good operating conditions. However, uneven stirring or other interferences can disrupt the overall temperature uniformity of the crystallizer, easily leading to local hot spots or cold spots and generating a large number of fine crystals.

[0032] Preferably, this application extracts temperature distribution features from the spatial dimension, including a third feature parameter; calculates the standard deviation of the temperature difference between any location and its surrounding neighboring locations at each acquisition time; and takes the average of the standard deviations of any location at all acquisition times within a single monitoring period to obtain the third feature parameter, which is used to reflect the degree of non-uniformity of temperature distribution at that location within the monitoring period.

[0033] Specifically, in this embodiment, taking a certain sampling location as the center, the temperature difference between that location and its surrounding neighboring locations (defined as the four surrounding neighboring locations in this embodiment) is calculated at each sampling time. Then, the standard deviation of all calculated differences for that location at each sampling time is calculated, and the average of all standard deviations for all sampling times within the monitoring period is taken as a feature value reflecting the degree of temperature unevenness at that location within the monitoring period, denoted as the third feature parameter. The result The larger the value, the more significant the uneven distribution of temperature data at that location, and the more likely local hot spots or cold spots are to appear.

[0034] Specifically, for the edge sampling positions at both ends of the crystallizer, there are missing data in their 4-neighborhood. These missing neighbors can be excluded during the calculation process.

[0035] Preferably, this application extracts temperature fluctuation features and temperature distribution features from the time and spatial dimensions, and fuses them to obtain a first feature value, including: the first feature value is positively correlated with the first feature parameter and the third feature parameter, and negatively correlated with the second feature parameter; the larger the first feature value, the more significant the nonlinear fluctuation and local uneven distribution features of the temperature data at that location during the monitoring period.

[0036] It is understood that the positive and negative correlations mentioned in this application refer to the two having the same or opposite trend of change. The specific function form can be set according to the actual application scenario, and this application does not impose any special restrictions.

[0037] Specifically, in this embodiment, the first feature value of the temperature data at a certain acquisition location during the i-th monitoring period is calculated using the following formula: In the formula, To avoid adjustment coefficients with a denominator of 0, in this embodiment... The value is 1.01. The value range of is [-1, 1], and the result is The larger the value, the more significant the nonlinear fluctuations and uneven local distribution characteristics of the temperature data during that period.

[0038] It should be noted that the parameters , From the perspective of time variation, it reflects both the short-term abrupt changes in temperature data and the linear smooth fluctuations over a period of time. Parameters This reflects the uneven temperature distribution in the crystallizer from a spatial perspective, thus enabling a comprehensive assessment of the abnormal characteristics of temperature changes.

[0039] Furthermore, as the crystallization process continues and more crystals are produced, the viscosity of the solution gradually increases, leading to a decrease in the heat and mass transfer efficiency of the solution. This amplifies the negative impact of cooling crystallization. At this point, even minor temperature disturbances can trigger rapid crystal nucleation or growth. This manifests in the temperature data as an overall nonlinear cooling trend and increasingly severe uneven distribution. Therefore, if the obtained first characteristic value shows a continuous increase or remains at a high level, it indicates that the temperature anomaly at that location is constantly worsening and is subject to long-term external interference (such as uneven stirring, scaling on the inner wall of the crystallizer, or crystal adhesion).

[0040] In view of this, this application obtains a second characteristic value based on the changing trend and average level of the first characteristic value over multiple consecutive monitoring periods. The second characteristic value is used to reflect the continuous increase or overall abnormality of the temperature fluctuation at that location.

[0041] Preferably, this application obtains a second characteristic value based on the changing trend and mean level of the first characteristic value over multiple consecutive monitoring periods, including: obtaining an estimated slope value of the changing trend of the first characteristic value over multiple consecutive monitoring periods, and performing an exponential mapping on the ratio of the estimated slope value to a preset benchmark value of the estimated slope value; calculating the ratio of the mean of the first characteristic value over multiple consecutive monitoring periods to a preset benchmark value of the mean of the first characteristic value; multiplying the calculation result of the mapped exponential function with the ratio result to obtain the second characteristic value, wherein the second characteristic value is used to reflect the continuous increase or overall abnormal characteristics of the temperature fluctuation anomaly at the location.

[0042] Specifically, in this embodiment, for a certain data collection location, the Sen's Slope test algorithm is used to obtain the changing trend of all first feature values ​​within the i-th and the preceding N monitoring periods. The output of the algorithm is a slope estimate, denoted as the slope estimate of the first feature value in the i-th monitoring period. The result The larger the value, the faster the first characteristic value at that location increases over multiple consecutive monitoring periods, indicating that the abnormal temperature condition is deteriorating rapidly.

[0043] Then, calculate the mean of all first feature values ​​within the i-th monitoring period and the N preceding monitoring periods, and denot it as the mean of the first feature values ​​for the i-th monitoring period. The result The larger the value, the more likely the location is in an abnormal state due to long-term interference from external factors, resulting in poor temperature control stability.

[0044] Furthermore, in order to evaluate and To characterize the degree to which the abnormal temperature state at this location is deteriorating and the degree to which this location is subject to long-term interference from external factors, this embodiment obtains the average value of the corresponding monitoring parameters (slope estimate and the average value of the first characteristic value) from all previous production-compliant operation monitoring periods of similar crystallizers as the benchmark value. That is, the benchmark value of the slope estimate is denoted as K0, and the benchmark value of the average value of the first characteristic value is denoted as E0.

[0045] Specifically, if there is no previous monitoring period in which the same type of crystallizer has achieved production targets, the temperature of the crystallizer in the current monitoring period will not be adjusted. In this case, it is necessary for staff to manually determine whether the target has been met and then manually adjust the temperature accordingly.

[0046] The Sen's Slope test algorithm is a well-known technique, and its specific implementation process will not be described in detail in this application.

[0047] Specifically, in this embodiment, the second characteristic value of the temperature data change at a certain acquisition location during the i-th monitoring period is calculated using the following formula: In the formula, This represents an exponential function with the natural constant e as the base. To avoid negative results during calculation that could affect the accurate evaluation of subsequent features, an exponential function is used for mapping. The resulting... The larger the value, the more pronounced the abnormal temperature fluctuation at that location, or the more obvious the overall abnormal characteristics.

[0048] To avoid the denominator being zero in division operations due to the ideal history constant state, the following methods are used when substituting into the formula for calculation: , The value is taken as the denominator, where max is the maximum value function. , All are preset minimum positive numbers. In this embodiment, the value is 0.01. The specific value can be set by the implementer according to the actual situation.

[0049] Furthermore, during the cooling crystallization process, there exists a dynamic equilibrium between temperature and solution concentration. Specifically, as the temperature steadily decreases, the equilibrium solubility of the solution continuously declines, causing allulose to precipitate from the solution and attach to existing crystal nuclei or seeds for growth, resulting in a steady decrease in the actual concentration of the solution. Therefore, under ideal crystallization conditions, the solution concentration and temperature change in a stable and synchronous manner.

[0050] Therefore, this application obtains a third characteristic value based on the synchronous change relationship between temperature data and solution concentration data collected in real time at multiple locations within the crystallizer during the crystallization process of allulose solution. The larger the third characteristic value, the more unstable the dynamic balance relationship between solution concentration and temperature change during the crystallization process.

[0051] Preferably, this application analyzes the synchronous change relationship between temperature data and solution concentration data, including a fourth characteristic parameter and a fifth characteristic parameter; calculates the correlation coefficient between the average temperature and the average concentration at all locations within multiple consecutive monitoring periods to obtain the fourth characteristic parameter, which is used to reflect the stable synchronous change characteristics between temperature and concentration; calculates the dispersion of concentration data within a single monitoring period to obtain the fifth characteristic parameter, which is used to reflect the relative dispersion of concentration data.

[0052] Specifically, in this embodiment, the average concentration of the corresponding concentration data within each monitoring period and the average temperature of the corresponding temperature data at all locations within each monitoring period are calculated. The Spearman rank correlation coefficient between the average temperature and the average concentration within the i-th and the preceding N (values ​​are 12) monitoring periods is calculated and denoted as the fourth characteristic parameter. The result The larger the value, the more pronounced the stable and synchronous changes in temperature and concentration during the crystallization process.

[0053] Meanwhile, proper temperature control aims to generate and maintain a moderate and stable supersaturation, thereby producing allulose crystals with high quality and stability. Therefore, if the solution concentration exhibits abnormal changes, such as large fluctuations, during temperature decrease, it indicates possible uneven heat transfer or other unexpected phase transitions, which will directly affect the particle size distribution and purity of the crystal product.

[0054] Therefore, in this embodiment, the coefficient of variation of the concentration data during the i-th monitoring period is calculated and denoted as the fifth characteristic parameter. The result The larger the value, the higher the relative dispersion of the solution concentration data within that time period.

[0055] Preferably, this application obtains a third characteristic value based on the synchronous change relationship between temperature data and solution concentration data collected in real time at multiple locations within the crystallizer during the crystallization process of allulose solution, including: the third characteristic value is positively correlated with the fifth characteristic parameter and negatively correlated with the fourth characteristic parameter; the larger the third characteristic value, the more unstable the dynamic equilibrium relationship between solution concentration and temperature change during the crystallization process.

[0056] Specifically, in this embodiment, the third characteristic value reflecting the dynamic equilibrium relationship between solution concentration and temperature during the i-th monitoring period is calculated using the following formula: In the formula, The value range of is [-1, 1]. To avoid an adjustment factor with a denominator of 0, a value of 1.1 is used. The resulting... The larger the value, the less stable the dynamic equilibrium between solution concentration and temperature changes during crystallization.

[0057] In summary, in order to further comprehensively assess the health status of the crystallization process from multiple dimensions, this application weighted and fused the second and third characteristic values ​​to obtain a comprehensive abnormal characteristic value. The comprehensive abnormal characteristic value is used to characterize the comprehensive risk level of the crystallization state at this location deviating from good working conditions during the monitoring period.

[0058] Specifically, in this embodiment, the formula for calculating the comprehensive abnormal characteristic value reflecting the comprehensive abnormal state of the crystallization process at a certain acquisition location during the i-th monitoring period is: In the formula, w1 and w2 are preset weighting coefficients, where In this embodiment, both w1 and w2 are set to 0.5; in other embodiments, the implementer can set them according to the actual situation. The larger the value, the more significant and potentially worsening the temperature control anomaly is at that location during the monitoring period, and the more unstable the dynamic balance between temperature and concentration is. The higher the overall risk of the crystallization state deviating from optimal operating conditions.

[0059] S3: Optimized temperature control within the crystallizer.

[0060] This application conducts an in-depth analysis of the spatiotemporal variation characteristics and abnormal trends of temperature data during the crystallization process, and further considers the dynamic equilibrium relationship between temperature and solution concentration to accurately assess the overall health status of the crystallization process.

[0061] Since the crystallization process is relatively complex and easily affected by various factors, this application dynamically adjusts the proportional term parameter of the PID controller based on the comprehensive abnormal characteristic value, and uses the adjusted PID controller to output the adjustment command to the temperature control actuator to regulate the temperature of the crystallization reactor and improve the temperature control accuracy of the crystallizer.

[0062] Specifically, the proportional term in the PID controller determines the control system's response rate to temperature regulation. The larger the obtained comprehensive abnormal characteristic value, the higher the overall risk of the crystallization state deviating from the good operating condition. In this case, a larger proportional term parameter should be set. Conversely, if the obtained comprehensive abnormal characteristic value is smaller, it indicates a better crystallization state. A smaller proportional term parameter can be set to avoid chattering and improve the stability of the control system.

[0063] Preferably, the present application dynamically adjusts the proportional term parameter of the PID controller based on the comprehensive abnormal characteristic value, including: comparing the maximum value of the comprehensive abnormal characteristic value calculated from all locations in the previous monitoring period with the preset steady-state lower limit and warning upper limit to determine the proportional term parameter of the PID controller in the current monitoring period; when the maximum value of the comprehensive abnormal characteristic value calculated from all locations in the previous monitoring period is between the steady-state lower limit and the warning upper limit, calculating the proportional term parameter of the PID controller in the current monitoring period based on the linear mapping relationship.

[0064] In this embodiment, based on the preset steady-state lower limit and warning upper limit, the maximum and minimum values ​​of the corresponding monitoring parameters (comprehensive abnormal characteristic values) are obtained from all previous production-compliant operation monitoring periods of the same type of crystallizer, and are used as the warning upper limit values ​​of the comprehensive abnormal characteristic values, respectively. and steady-state lower limit .

[0065] Specifically, if If there is no previous monitoring period for achieving production targets for the same type of crystallizer, the temperature of the crystallizer in the current monitoring period will not be adjusted. In this case, it is necessary for staff to manually judge whether the target has been met and then manually adjust the temperature.

[0066] Furthermore, in this embodiment, the safe operating upper limit of the scaling term is obtained using the Ziegler-Nichols method. The lower limit for the smooth operation of the proportional term is , The Ziegler-Nichols method is a well-known technique, and its specific process will not be described in detail in this application.

[0067] Specifically, in this embodiment, the maximum value of the comprehensive abnormal feature value calculated from all locations within the previous monitoring period is obtained. ,when Between and In between, the proportional term parameter of the PID controller for the current monitoring period is calculated based on this. The calculation formula is: Calculated The proportional parameter will be sent as a configuration value to the PID controller for the current monitoring period.

[0068] Specifically, when Less than or equal to When the crystallizer temperature is confirmed to be within the envelope, a stable operating lower limit is issued to the controller. The proportional term parameter of the PID controller for the current monitoring period is assigned a value. When the crystallization state within the crystallizer faces deterioration, i.e. Greater than or equal to At this time, a safe operating limit is issued to the controller. The proportional term parameter of the PID controller for the current monitoring period is assigned a value. The proportional term parameter of the PID controller is locked at the safe operating limit during the current monitoring period. To ensure that the limits are not exceeded.

[0069] Finally, the continuously updated PID controller outputs adjustment commands to the temperature control actuator, thereby adjusting the temperature control actuator of the horizontal cooling crystallizer to regulate the temperature of the crystallization reactor and improve the temperature control accuracy.

[0070] The method of using an updated PID controller to output adjustment commands to control the temperature of the crystallization reactor is a known technique, and its specific process will not be described in detail in this application.

[0071] Another embodiment of this application provides a PID-based temperature control system for an allulose crystallization reactor. The system includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the PID-based temperature control method for the allulose crystallization reactor described above.

Claims

1. A temperature control method for an allulose crystallization reactor based on PID control, characterized in that, The method includes the following steps: Based on real-time temperature data collected from multiple locations within the crystallizer during the crystallization process of allulose solution, temperature fluctuation and temperature distribution features are extracted from both temporal and spatial dimensions and fused to obtain a first feature value. A second feature value is then obtained based on the changing trend and average level of this first feature value over multiple consecutive monitoring periods. The third characteristic value was obtained based on the synchronous change relationship between temperature data and solution concentration data collected in real time at multiple locations within the crystallizer during the crystallization process of allulose solution; The second and third eigenvalues ​​are weighted and fused to obtain a comprehensive abnormal eigenvalue. Based on this comprehensive abnormal eigenvalue, the proportional term parameter of the PID controller is dynamically adjusted, and the adjusted PID controller outputs a regulation command to the temperature control actuator to regulate the temperature of the crystallization reactor.

2. The temperature control method for an allulose crystallization reactor based on PID control as described in claim 1, characterized in that, Temperature fluctuation features are extracted from the time dimension, including the first feature parameter and the second feature parameter; By analyzing the dispersion of temperature data at any location within a single monitoring period, a first characteristic parameter is obtained. The first characteristic parameter is used to reflect the probability of a sudden local temperature change at that location within that monitoring period. The autocorrelation of the average temperature at any location over multiple consecutive monitoring periods is analyzed to obtain a second characteristic parameter. This second characteristic parameter is used to reflect the linear and smooth change characteristics of the overall average temperature of the crystallizer at that location.

3. The temperature control method for an allulose crystallization reactor based on PID control as described in claim 2, characterized in that, Temperature distribution features are extracted from the spatial dimension, including the third feature parameter; Calculate the standard deviation of the temperature difference between any location and its surrounding neighborhood locations at each data acquisition time. The average of the standard deviations of all data collection times at any location within a single monitoring period is used to obtain a third characteristic parameter, which reflects the degree of non-uniformity of temperature distribution at that location within the monitoring period.

4. The temperature control method for an allulose crystallization reactor based on PID control as described in claim 3, characterized in that, Temperature fluctuation and temperature distribution features are extracted from the time and space dimensions and fused to obtain the first feature value, including: The first feature value is positively correlated with the first feature parameter and the third feature parameter, and negatively correlated with the second feature parameter. The larger the first feature value, the more significant the nonlinear fluctuation and uneven local distribution of the temperature data at that location during the monitoring period.

5. The temperature control method for an allulose crystallization reactor based on PID control as described in claim 4, characterized in that, The second characteristic value is obtained based on the changing trend and mean level of the first characteristic value over multiple consecutive monitoring periods, including: Obtain the slope estimate of the changing trend of the first feature value within multiple consecutive monitoring periods, and perform an exponential mapping on the ratio of the slope estimate to a preset benchmark value of the slope estimate; Calculate the ratio of the mean of the first characteristic value over multiple consecutive monitoring periods to a preset benchmark value of the mean of the first characteristic value; The second characteristic value is obtained by multiplying the calculated result of the exponential function obtained by mapping with the ratio result. The second characteristic value is used to reflect the continuous increase or overall abnormality of the temperature fluctuation at this location.

6. The temperature control method for an allulose crystallization reactor based on PID control as described in claim 1, characterized in that, Analyze the synchronous changes between temperature data and solution concentration data, including the fourth and fifth characteristic parameters; The correlation coefficient between the average temperature and average concentration at all locations over multiple consecutive monitoring periods is calculated to obtain a fourth characteristic parameter, which is used to reflect the stable and synchronous change characteristics between temperature and concentration. The dispersion of concentration data within a single monitoring period is calculated to obtain a fifth characteristic parameter, which is used to reflect the relative dispersion of concentration data.

7. The temperature control method for an allulose crystallization reactor based on PID control as described in claim 6, characterized in that, The third characteristic value is obtained based on the synchronous change relationship between temperature data and solution concentration data collected in real time at multiple locations within the crystallizer during the allulose solution crystallization process. This includes: The third characteristic value is positively correlated with the fifth characteristic parameter and negatively correlated with the fourth characteristic parameter; the larger the third characteristic value, the more unstable the dynamic equilibrium between solution concentration and temperature change during crystallization.

8. The temperature control method for an allulose crystallization reactor based on PID control as described in claim 1, characterized in that, Based on this comprehensive abnormal characteristic value, the proportional term parameter of the PID controller is dynamically adjusted, including: The maximum value of the comprehensive abnormal characteristic value calculated from all locations in the previous monitoring period is compared with the preset steady-state lower limit and warning upper limit to determine the proportional term parameter of the PID controller for the current monitoring period. When the maximum value of the comprehensive abnormal characteristic value calculated from all locations in the previous monitoring period is between the steady-state lower limit and the warning upper limit, the proportional term parameter of the PID controller for the current monitoring period is calculated based on the linear mapping relationship.

9. The PID-based temperature control method for an allulose crystallization reactor as described in claim 8, characterized in that, The adjustment of the proportional term parameter of the PID controller also includes: When the maximum value is less than or equal to the steady-state lower limit, the proportional parameter of the PID controller for the current monitoring period is set to the steady-state lower limit. When the maximum value is greater than or equal to the warning upper limit value, the proportional parameter of the PID controller for the current monitoring period is set to the safe operating upper limit value.

10. A PID-based temperature control system for an allulose crystallization reactor, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the PID-based temperature control method for an allulose crystallization reactor as described in any one of claims 1-9.

Images