Centrifuge control method and system for soybean protein production

Abstract

The present invention relates to the field of centrifuge PID control technology, and specifically to a centrifuge control method and system for soy protein production. The method obtains time series data of the centrifuge speed and torque, analyzes data differences and fluctuation trends, calculates abnormal indicators of state parameters at each moment, accurately identifies abnormal data points and abnormal moments, and locates key nodes of speed mutation; determines abnormal time periods based on the time series characteristics of abnormal moments, analyzes the connection between speed and torque, the similarity of abnormal indicators, and speed fluctuation characteristics within the period, quantifies the dynamic correlation between the two, and formulates adjustment indicators; in response to the control lag problem caused by fixed parameters of traditional PID, the present invention combines the abnormal indicator differences of speed and torque at the current moment and the adjustment indicators determined in the early stage, dynamically adjusts PID parameters, and makes the control strategy adaptively optimized as the working conditions change, ultimately achieving stable, efficient, and precise control of the centrifuge speed.

Description

Technical Field

[0001] The present invention relates to the technical field of centrifuge PID control, and in particular to a centrifuge control method and system for soy protein production. Background Art

[0002] In the soy protein production process, centrifuges are core separation equipment, primarily used to efficiently separate the solid phase (soy dregs) and liquid phase (protein solution) of soybean meal slurry. The centrifuge's high-speed rotation generates centrifugal force, rapidly separating the refined mixture into separate layers, achieving a thorough separation of the solid and liquid phases. This allows extraction of the protein-rich liquid phase and improves protein yield. Therefore, the centrifuge's operational stability directly impacts final product quality (such as protein purity and yield) as well as the equipment's lifespan.

[0003] Existing technology typically uses a PID controller to adjust the centrifuge speed to accommodate slurries of varying viscosities or particle sizes, ensuring stable separation and minimizing protein loss. Traditional PID control methods often employ fixed PID parameters to adjust the centrifuge speed. However, during soy protein production, the concentration, viscosity, and solids content of the soy protein slurry fluctuate with raw material batches and process stages, causing the centrifuge's operating conditions to dynamically change. Therefore, fixed PID parameters are unable to adapt to these dynamically changing operating conditions, resulting in reduced centrifuge control accuracy and even equipment damage. Summary of the Invention

[0004] In order to solve the technical problem that the concentration, viscosity, solid content, etc. of the soy protein liquid fluctuate with the raw material batch and process stage during the soy protein production process, thereby causing the working conditions of the centrifuge to change dynamically, and the fixed PID parameters cannot adapt to the dynamically changing working conditions, resulting in reduced centrifuge control accuracy and even damage to the equipment, the purpose of the present invention is to provide a centrifuge control method and system for soy protein production, and the technical solutions adopted are as follows:

[0005] A centrifuge control method for soybean protein production, comprising:

[0006] Obtaining time series data of status parameters of a centrifuge during the soy protein production process, wherein the status parameters include speed and torque;

[0007] In each state parameter time series data, the difference characteristics between the state parameter values ​​are analyzed and combined with the fluctuation trend of the state parameter values ​​to determine the abnormal indicators of the state parameter values ​​at each moment; in the speed time series data, the abnormal time points are screened out based on the abnormal indicators of the speed values;

[0008] The abnormal time period is determined based on the time series characteristics of the abnormal time point; in the abnormal time period, the connection characteristics between the torque value and the speed value, the similar characteristics of the abnormal indicators are analyzed, and combined with the numerical fluctuation characteristics of the speed value to determine the speed adjustment index of the centrifuge;

[0009] The difference characteristics of abnormal indicators between the speed value and torque value of the centrifuge at the current moment are analyzed, and combined with the speed adjustment index of the centrifuge, the preset PID parameters are adjusted to perform PID control on the speed value of the centrifuge at the next moment.

[0010] Furthermore, the method for obtaining the abnormality indicator includes:

[0011] In each state parameter time series data, the state parameter value at each moment is taken as the center to determine the corresponding preset neighborhood;

[0012] In the preset neighborhood corresponding to each state parameter value in each state parameter time series data, the absolute value of the difference between the central state parameter value and each neighboring state parameter value is calculated as the prominence factor of the central state parameter value, and the average level of the prominence factor between the central state parameter value and all corresponding neighboring state parameter values ​​is analyzed to obtain the first abnormal factor of the central state parameter value;

[0013] In each state time series data, within the preset neighborhood corresponding to each state parameter value, the fluctuation trend between the state parameter values ​​is analyzed to obtain the second abnormal factor of each state parameter value;

[0014] The first anomaly factor and the second anomaly factor of each state parameter value are integrated to obtain the anomaly index of the state parameter value at each moment.

[0015] Furthermore, the method for obtaining the second abnormal factor includes:

[0016] In each state parameter time series data, within a preset neighborhood corresponding to each state parameter value, curve segments of all state parameter values ​​are obtained based on the least squares method;

[0017] On the curve segment, the slope value at each state parameter value is obtained, and the variance of all slope values ​​is calculated as the second abnormal factor of the central state parameter value.

[0018] Furthermore, the method for obtaining the speed regulation index includes:

[0019] In each state parameter time series data, the state parameter value in each abnormal time period is combined into an abnormal data segment;

[0020] In the two state parameter time series data, the two abnormal data segments corresponding to the same abnormal time period are combined as one data segment;

[0021] In each data segment combination, the connection characteristics between the state parameter values ​​in the two abnormal data segments and the similar characteristics of the abnormal indicators are analyzed to obtain the adjustment factor corresponding to each data segment combination;

[0022] In each data segment combination, the variance of all speed values ​​in the abnormal data segment corresponding to the speed is used as the fluctuation factor corresponding to each data segment combination;

[0023] The fluctuation factors are weighted and summed using the adjustment factors of the data segment combination, and the value obtained by normalizing the weighted result is used as the speed adjustment index of the centrifuge.

[0024] Furthermore, the method for obtaining the adjustment factor includes:

[0025] In each data segment combination, the Pearson correlation coefficient between the state parameter values ​​in the two abnormal data segments is calculated, and the Pearson correlation coefficient is normalized to serve as the influencing factor between the torque and the speed in each data segment combination;

[0026] In each data segment combination, the absolute value of the difference between the abnormal indicators of the torque value and the speed value at the same time is calculated as the abnormal deviation factor. The sum of all abnormal deviation factors corresponding to the data segment combination is negatively correlated and mapped to the value as the abnormal similarity factor between the torque value and the speed value in each data segment combination;

[0027] The product of the impact factor and the abnormal similarity factor corresponding to each data segment combination is normalized and used as the adjustment factor corresponding to each data segment combination.

[0028] Furthermore, the analysis of the difference characteristics of the abnormal indicators between the speed value and the torque value of the centrifuge at the current moment, and the adjustment of the preset PID parameters in combination with the speed adjustment index of the centrifuge, includes:

[0029] The preset PID parameters include a preset proportional coefficient, a preset integral coefficient, and a preset differential coefficient;

[0030] At the current moment, the absolute value of the difference between the centrifuge's speed value and torque value is used as the state difference factor;

[0031] The sum of the normalized value of the product of the state difference factor and the speed regulation index and a preset constant is used as the regulation degree value;

[0032] The product of the adjustment degree value and the preset proportional coefficient is used as the adjustment proportional coefficient.

[0033] Furthermore, the PID control of the rotation speed value of the centrifuge at the next moment includes:

[0034] Get the system error at the current moment;

[0035] The product of the current centrifuge speed value and the adjustment proportional coefficient is used as the first adjustment factor, and the product of the system error and the preset integral coefficient is used as the second adjustment factor;

[0036] The sum of the first adjustment factor, the second adjustment factor and the preset differential coefficient is used as the input of the PID controller to obtain a control output value, which is used to perform PID control on the rotation speed value of the centrifuge at the next moment.

[0037] Furthermore, the method for obtaining the abnormal time point includes:

[0038] In the speed time series data, the moment corresponding to the speed value where the abnormality index is greater than the preset abnormality threshold is taken as the abnormal moment point.

[0039] Furthermore, the method for obtaining the abnormal time period includes:

[0040] In terms of time series, the abnormal time points that are continuous in time series are grouped into abnormal time periods, thereby obtaining all abnormal time periods, wherein each abnormal time period should contain at least two moments.

[0041] A centrifuge control system for soy protein production includes a processor and a memory. The memory stores at least one instruction, at least one program, code set, or instruction set. When the at least one instruction, at least one program, code set, or instruction set is loaded and executed by the processor, the steps of a centrifuge control method for soy protein production are implemented.

[0042] The present invention has the following beneficial effects:

[0043] First, the system acquires time-series data of centrifuge state parameters during the soy protein production process. These state parameters include speed and torque. This provides a digital representation of the centrifuge's operating status and continuous information generated by operating condition changes. By analyzing the differential characteristics and fluctuation trends of the speed and torque time-series data, it dynamically calculates anomaly indicators for each state parameter value at each moment. This design significantly improves the accuracy of anomaly identification, enabling early detection of precursors to operating condition changes and reducing response lag. Then, based on the speed anomaly indicators, it identifies anomaly moments, quickly locating key time points where the centrifuge speed suddenly changes, providing a precise basis for subsequent control adjustments. Furthermore, the anomaly time period is determined based on the time-series characteristics of the anomaly moments. Within the anomaly time period, this solution quantifies the dynamic relationship between speed and torque by analyzing the connection characteristics between speed and torque (such as the lag in speed response to torque abrupt changes), similarity characteristics (such as the correlation in the intensity of anomaly indicators), and numerical fluctuation characteristics of speed values. This analysis clarifies the specific requirements for speed regulation, enabling the development of speed regulation indicators that better align with actual operating conditions. Traditional PID control, due to its fixed parameters, is prone to control lag when operating conditions change (e.g., regulation is initiated only after an anomaly occurs). This solution dynamically adjusts PID parameters by combining the current abnormal speed and torque indicator differences with the previously determined speed regulation indicators. This design enables PID control to "change with operating conditions," ultimately achieving stable, efficient, and precise control of the centrifuge speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 A flow chart of a centrifuge control method for soy protein production provided by one embodiment of the present invention;

[0046] Figure 2 A flowchart of a method for obtaining abnormal indicators provided by one embodiment of the present invention;

[0047] Figure 3 A flow chart of a method for obtaining a speed regulation index provided by one embodiment of the present invention;

[0048] Figure 4 A system block diagram of a centrifuge control system for soy protein production provided by one embodiment of the present invention;

[0049] Figure 5 A schematic diagram of the system structure of a centrifuge control system for soy protein production provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0050] To further illustrate the technical means and effectiveness of the present invention in achieving its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail a centrifuge control method and system for soy protein production according to the present invention, including its specific implementation, structure, features, and effectiveness. In the following description, references to different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0051] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0052] The following describes in detail a centrifuge control method and system for soybean protein production provided by the present invention with reference to the accompanying drawings.

[0053] See also Figure 1 , which shows a method flow chart of a centrifuge control method for soybean protein production provided by one embodiment of the present invention, the method comprising the following steps:

[0054] Step S1: obtaining time series data of status parameters of a centrifuge during the soy protein production process, wherein the status parameters include rotation speed and torque.

[0055] In the soy protein production process, centrifuges are key separation equipment, and their operational stability directly affects product quality (such as protein purity and yield) and equipment life. The centrifuge generates centrifugal force through high-speed rotation, quickly separating the mixed liquor after refining, achieving a thorough separation of the solid and liquid phases, thereby extracting the protein-rich liquid phase and increasing protein yield. By adjusting the speed, it can adapt to slurries of different viscosities or particle sizes, ensuring a stable separation effect and reducing protein loss. In addition, centrifugal separation is more efficient than traditional filtration methods and can be operated continuously, significantly improving the processing capacity of the production line. At the same time, the separated protein liquid has a higher purity, which is beneficial for subsequent acid precipitation, neutralization and other processes, and the removed soybean dregs can be further used for feed or fiber extraction, maximizing resource utilization.

[0056] Therefore, precise centrifuge control is crucial for ensuring product quality, production efficiency, and process safety during soy protein production. Soy protein extraction requires centrifugal separation to remove impurities such as fiber and oil, and the centrifuge speed directly impacts separation purity. For example, controlling the speed within the 3000-5000 RPM range can ensure a 5%-8% increase in protein yield while reducing solid phase carryover. The application of PID (proportional-integral-differential) control systems demonstrates significant technical advantages in the centrifugal separation process of soy protein production. However, traditional PID parameter fixation struggles to adapt to the dynamic, nonlinear characteristics of the centrifuge during acceleration, constant speed, and deceleration phases (such as the random fluctuations in material viscosity within the 8%-15% range during soy protein separation), leading to erratic speed adjustments. Furthermore, there is a lag in the response to sudden load changes, causing the separation factor to deviate beyond a reasonable threshold, which can also severely impact protein separation purity.

[0057] The operating status of a centrifuge is characterized by a variety of parameters, but the speed directly determines the separation efficiency and product quality, and the torque reflects the stress on the centrifuge drive shaft, which is directly related to the equipment load and mechanical state. Therefore, the speed and torque are considered to be the two most core parameters. Therefore, in the soy protein production process, the two state parameter time series data of the centrifuge are obtained, namely the speed time series data and the torque time series data. Specifically, a laser velocimeter or encoder can be installed on the main shaft of the centrifuge to measure the speed of the drum, and a strain torque sensor can be installed on the drive shaft of the centrifuge to measure the torque. The acquisition frequency of the speed time series data and the torque time series data is set to once per second, and the length is set to 30 minutes of history starting from the current moment.

[0058] It should be noted that the frequency and length of data collection can be adjusted according to the implementation scenario and are not limited here.

[0059] Step S2: In each state parameter time series data, analyze the difference characteristics between the state parameter values ​​and combine them with the fluctuation trend of the state parameter values ​​to determine the abnormal indicators of the state parameter values ​​at each moment; in the speed time series data, filter out abnormal time points based on the abnormal indicators of the speed values.

[0060] During the soy protein production process, changes in the state of the mixed liquid (such as fluctuations in the concentration, viscosity, and solid content of the liquid) will cause dynamic changes in the working conditions of the centrifuge, which can be specifically characterized by the state time series data of the centrifuge. Therefore, we can first analyze the difference characteristics and fluctuation trends between the state parameter values ​​in each state parameter time series data to reflect the changes in the state parameters in the time dimension, thereby identifying mutation-type anomalies and obtaining the anomaly indicators of the state parameter values ​​at each moment to determine whether there is an anomaly in the state parameter values ​​at each moment.

[0061] Preferably, in one embodiment of the present invention, the method for obtaining abnormal indicators includes:

[0062] See also Figure 2 , which shows a flow chart of a method for obtaining abnormal indicators in one embodiment of the present invention, the method includes the following steps:

[0063] Step S201: In each state parameter time series data, determine a preset neighborhood corresponding to the state parameter value at each moment.

[0064] By setting the neighborhood range, we can perform local analysis on the state parameter value at each moment, thereby analyzing the short-term changes in the working conditions and avoiding the interference of noise or long-term trends in the abnormal judgment in the global analysis. Therefore, in each state parameter time series data, the state parameter value at each moment is centered, and the central state parameter value and the preset number of state parameter values ​​closest to the central state parameter value are combined to form the neighborhood range corresponding to the central state parameter value. In this way, the preset neighborhood corresponding to the state parameter value at each moment is obtained.

[0065] It should be noted that, in this embodiment of the present invention, the preset number is 5, and the specific value can be adjusted according to the implementation scenario and is not limited here.

[0066] Step S202: In each state parameter time series data, within a preset neighborhood corresponding to each state parameter value, analyze the difference characteristics between the state parameter values ​​to obtain a first abnormal factor of each state parameter value.

[0067] The normal fluctuation range of state parameters under different working conditions is different, so through relative difference analysis within a preset neighborhood, the fluctuation characteristics of different working conditions can be adaptively adjusted to obtain more accurate abnormal measurement parameters.

[0068] In the preset neighborhood corresponding to each state parameter value in each state parameter time series data, the absolute value of the difference between the central state parameter value and each neighborhood state parameter value (other state parameter values ​​except the central state parameter value) is calculated as the prominence factor of the central state parameter value. The larger the prominence factor, the greater the degree of deviation between the central state parameter and the neighborhood state parameter, the higher the local prominence, and the more likely an abnormal situation has occurred.

[0069] At this time, there is a prominence factor between the central state parameter value and each neighborhood state parameter value. Then, the average value of the prominence factors between the central state parameter value and all corresponding neighborhood state parameter values ​​is normalized to obtain the first anomaly factor of the central state parameter value. Based on the above analysis, it can be seen that the larger the first anomaly factor, the higher the local prominence of the central state parameter value, and the more likely an anomaly has occurred. Normalization is a technical means well known to those skilled in the art. The normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0070] Step S203: In each state time series data, within a preset neighborhood corresponding to each state parameter value, analyze the fluctuation trend between the state parameter values ​​to obtain a second abnormal factor for each state parameter value.

[0071] The changing trend of the state parameter value can be used to analyze the stability of the data. Under normal circumstances, the state parameter value should show a relatively smooth growth or decline trend. Therefore, if the data change trend is more chaotic, it means that the data change stability is poor, which can be used to distinguish normal fluctuations from abnormal fluctuations.

[0072] Therefore, in each state parameter time series data, within the preset neighborhood corresponding to each state parameter value, the curve segments of all state parameter values ​​are obtained based on the least squares method.

[0073] Then, on the curve segment, the slope value at each state parameter value is obtained, and the variance of all slope values ​​is calculated as the second anomaly factor of the central state parameter value. The larger the second anomaly factor, the more drastic the fluctuation of the preset neighborhood state parameter value corresponding to the central state parameter value, and the more likely it is an abnormal fluctuation.

[0074] Step S204: In each state time series data, the first abnormality factor and the second abnormality factor of each state parameter value are integrated to obtain the abnormality index of the state parameter value at each moment.

[0075] Based on the analysis in the above steps, it can be seen that the first abnormal factor and the second abnormal factor corresponding to each state parameter value are positively correlated with the probability of the state parameter value being abnormal. Therefore, in this embodiment of the present invention, in each state time series data, the product of the first abnormal factor and the second abnormal factor of each state parameter value is normalized to the value after being used as the abnormality index of the state parameter value at each moment, and the larger the abnormality index, the greater the possibility that the state parameter value is an abnormal data point. Normalization is a technical means well known to those skilled in the art, and the normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0076] At this point, the abnormal indicators of the state parameter values ​​at each moment can be obtained in each state parameter time series data. Then, the abnormal time points can be screened out based on the abnormal indicators of the state parameter values, thereby providing clear time points and data basis for subsequent control adjustments.

[0077] Preferably, in one embodiment of the present invention, the method for obtaining the abnormal time point includes:

[0078] The core control goal of the centrifuge is to maintain a stable speed, because the speed directly affects the separation effect and is the final output of the control. Therefore, in this embodiment of the present invention, more focus is placed on speed stability control, so the abnormal time point can be determined based on the speed time series data.

[0079] Based on the above analysis, it can be seen that the larger the abnormal index is, the greater the possibility of abnormality in the data. Therefore, in the speed time series data, the moment corresponding to the speed value where the abnormal index is greater than the preset abnormal threshold is taken as the abnormal moment point.

[0080] It should be noted that the preset abnormality threshold is 0.7, and the specific value can be adjusted according to the implementation scenario and is not limited here.

[0081] Step S3: Determine the abnormal time period based on the timing characteristics of the abnormal time point; in the abnormal time period, analyze the connection characteristics between the torque value and the speed value, the similarity characteristics of the abnormal indicators, and combine them with the numerical fluctuation characteristics of the speed value to determine the speed adjustment index of the centrifuge.

[0082] In step S2, discrete abnormal time points can be screened. In this step, the discrete abnormal time points can be integrated into time segments based on their temporal characteristics to obtain abnormal time periods, which facilitates centralized analysis of abnormal situations.

[0083] Preferably, in one embodiment of the present invention, a method for obtaining an abnormal time period includes:

[0084] In terms of time sequence, the abnormal time points that are continuous in time sequence are combined into abnormal time periods, thereby obtaining all abnormal time periods, wherein each abnormal time period should contain at least two moments, that is, in this embodiment of the present invention, only two or more consecutive abnormal time points can constitute an abnormal data segment, and a single abnormal time point is not within the scope of consideration.

[0085] Furthermore, during soy protein production, the centrifuge's torque fluctuates due to changes in process conditions and material properties. When the solids content (soy dregs ratio) or protein concentration of the soy slurry fluctuates, high-viscosity slurry increases frictional resistance within the drum, leading to increased torque. Conversely, thin slurry reduces torque. Excessive feed can lead to material accumulation within the drum, increasing the load on the screw conveyor and causing a sudden increase in torque. Insufficient flow can cause idling and a decrease in torque. Therefore, the abnormal data points obtained above may be caused by changes in torque during the operation of the centrifuge. The greater the load, the corresponding decrease in speed under the same power; conversely, the smaller the load, the corresponding increase in speed under the same power. Therefore, under normal circumstances, the changes between the torque value and the speed value should show an approximately negative correlation. Therefore, if in actual scenarios, the changes between the torque value and the speed value of the centrifuge show an opposite change relationship, it can reflect the influence of the torque value on the speed value during the operation of the centrifuge, that is, the degree of deviation between the working state and the expected state. This feature can be used to adjust the speed of the centrifuge. At the same time, in the abnormal time period, analyzing the similar characteristics between the abnormal indicators of the torque value and the speed value can analyze the synchronization of their patterns. It can also quantify the influence of the torque value on the speed value. Combined with the numerical fluctuation characteristics of the speed value, it is used to determine the specific fluctuation of the speed value, thereby obtaining a specific indicator that is more consistent with the degree of adjustment of the centrifuge speed state, namely the speed regulation index.

[0086] Preferably, in one embodiment of the present invention, the method for obtaining the speed regulation index includes:

[0087] See also Figure 3 , which shows a flow chart of a method for obtaining a speed regulation index in one embodiment of the present invention, the method comprising the following steps:

[0088] Step S301: In the two-state time series data, determine a data segment combination based on an abnormal time period.

[0089] In each state parameter time series data, the state parameter value in each abnormal time period is combined into an abnormal data segment, and in two state parameter time series data, the two abnormal data segments corresponding to the same abnormal time period are combined as one data segment.

[0090] Step S302: In each data segment combination, the connection characteristics between the state parameter values ​​in two abnormal data segments and the similarity characteristics of the abnormal indicators are analyzed to obtain the adjustment factor corresponding to each data segment combination.

[0091] In each data segment combination, the Pearson correlation coefficient between the state parameter values ​​in the two abnormal data segments is calculated. Based on the above analysis, it can be seen that under normal circumstances, the torque value and the speed value should show an approximately negative correlation. If an approximately positive correlation is shown, it means that the speed value is abnormally affected and deviates from the expected pattern. Therefore, the Pearson correlation coefficient is normalized as the influencing factor between the torque and speed in each data segment combination. The larger the influencing factor, the greater the impact on the speed value. There may be a large lag, so the degree of adjustment should be increased. Given that the Pearson correlation coefficient may be positive or negative, the normalization here can be used. function.

[0092] In each data segment combination, the absolute value of the difference between the abnormal indicators of the torque value and the speed value at the same time is calculated as the abnormal deviation factor. The smaller the abnormal deviation factor, the higher the abnormal similarity of the speed value and the torque value in the same time interval. It can be regarded as the greater the influence of the torque value on the speed value. Therefore, the sum of all the abnormal deviation factors corresponding to the data segment combination is negatively correlated and mapped to achieve logical relationship correction, thereby obtaining the abnormal similarity factor between the torque value and the speed value in each data segment combination. The larger the abnormal similarity factor, the greater the speed value is affected. The negative correlation mapping process here can be used ,in, It represents the exponential function with the natural constant e as the base, and x represents the independent variable.

[0093] Based on the above analysis, it can be seen that both the impact factor and the anomaly similarity factor are positively correlated with the degree of influence on the speed value. Therefore, the product of the impact factor and the anomaly similarity factor corresponding to each data segment combination is normalized and used as the adjustment factor corresponding to each data segment combination. A larger adjustment factor indicates a greater deviation between the actual speed operation in the current system and the expected speed, and a greater adjustment lag. Normalization is a technical means well known to those skilled in the art. The normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0094] Step S303: In each data segment combination, based on the fluctuation characteristics of the rotation speed value, determine the fluctuation factor corresponding to each data segment combination.

[0095] In each data segment combination, if the fluctuation of the centrifuge speed value itself is more obvious, it means that the speed itself is more unstable, and the necessity of adjustment is greater. Therefore, in each data segment combination, the variance of all speed values ​​in the abnormal data segment corresponding to the speed is used as the fluctuation factor corresponding to each data segment combination. The larger the fluctuation factor, the greater the necessity of adjustment.

[0096] Step S304: The adjustment factor and the fluctuation factor of the data segment combination are combined to obtain the speed adjustment index of the centrifuge.

[0097] Based on the above analysis, it can be seen that when the adjustment factor corresponding to the data segment combination is larger, it means that there is a large deviation between the actual operation of the speed in the current system and the expected speed and there is a large adjustment hysteresis, and the more adjustment is needed; when the fluctuation factor corresponding to the data segment combination is larger, the necessity of adjustment is greater. Therefore, in this step, the fluctuation factor is weighted and summed using the adjustment factor of the data segment combination, and the value after the obtained weighted result is normalized is used as the speed adjustment index of the centrifuge. The larger the speed adjustment index is, the more obvious the speed adjustment hysteresis of the centrifuge in the current system is, and the greater the influence of torque is, the more synchronous adjustment should be performed to reduce the error caused by hysteresis. Wherein normalization is a technical means well known to those skilled in the art, and the choice of normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0098] Step S4: Analyze the difference characteristics of the abnormal indicators between the speed value and the torque value of the centrifuge at the current moment, and adjust the preset PID parameters in combination with the speed adjustment index of the centrifuge to perform PID control on the speed value of the centrifuge at the next moment.

[0099] The speed regulation index calculated in the above steps reflects the changes in the entire system. However, the centrifuge has different operating states at different stages. For example, when the raw materials are first added for separation, because there is a lot of raw materials and the viscosity is high, the load is large when running at a constant power. After a period of separation, the concentration of the raw materials decreases, the load decreases, and the speed changes. Therefore, in order to ensure the quality of the separated soy protein, real-time adjustment is required. In addition, the PID model can be adjusted according to the real-time speed and torque conditions. Therefore, if the speed data at the current moment is highly abnormal, and its relationship with the torque data is analyzed, combined with the speed regulation index, the preset PID parameters can be adjusted to have better robustness.

[0100] When using a PID system to regulate speed, the proportional coefficient is the undisputed primary control component. The proportional coefficient generates a control signal based on the current error. The larger the error, the stronger the output, resulting in a faster response to the error and reduced adjustment lag. Therefore, in this embodiment of the present invention, when adjusting the preset PID parameters, the proportional coefficient is primarily adjusted.

[0101] Preferably, in one embodiment of the present invention, the difference characteristics of the abnormal indicators between the speed value and the torque value of the centrifuge at the current moment are analyzed, and the preset PID parameters are adjusted in combination with the speed adjustment index of the centrifuge, including:

[0102] The preset PID parameters include a preset proportional coefficient, a preset integral coefficient, and a preset differential coefficient, which can be calculated by the Ziegler-Nichols law in the PID data tuning method. This is a well-known technology and the specific process will not be repeated here.

[0103] At the current moment, the absolute value of the difference between the speed value and the torque value of the centrifuge is used as a state difference factor. The larger the state difference factor is, the lower the state consistency of the speed value and the torque value when the centrifuge is working is, so the hysteresis adjustment demand is higher. Based on the analysis in step S3, when the speed regulation index is larger, it is shown that the speed adjustment hysteresis of the centrifuge in the current system is more obvious. Therefore, the speed regulation index and the state difference factor are positively correlated with the adjustment response demand degree. Therefore, the product of the state difference factor and the speed regulation index is normalized. The sum of the values ​​of the preset constants is used as the adjustment degree value. At this time, the larger the adjustment degree value is, the larger the demand for improving the response speed is. Finally, the product of the adjustment degree value and the preset proportional coefficient is used as the adjustment proportional coefficient. The adjustment proportional coefficient at this time can better adapt to the problem of the response hysteresis existing in the current system, thereby ensuring the separation quality and efficiency of soy protein production. Wherein normalization is a technical means well known to those skilled in the art. The selection of normalization function can be linear normalization or standard normalization, etc., and specific normalization method is not limited here.

[0104] It should be noted that, in order to prevent overshoot, the preset constant in the embodiment of the present invention is 1.

[0105] After the adjustment proportional coefficient is obtained, it can be combined with the preset integral coefficient and the preset differential coefficient to jointly perform PID control on the rotation speed value of the centrifuge at the next moment.

[0106] Preferably, in one embodiment of the present invention, the PID control of the rotation speed value of the centrifuge at the next moment includes:

[0107] Obtain the system error at the current moment (i.e., the difference between the set value and the actual measured value); then use the product of the centrifuge speed value at the current moment and the adjustment proportional coefficient as the first adjustment factor, and use the product of the system error and the preset integral coefficient as the second adjustment factor.

[0108] Finally, the sum of the first adjustment factor, the second adjustment factor and the preset differential coefficient is used as the input of the PID controller to obtain a control output value, which is used to perform PID control on the rotation speed value of the centrifuge at the next moment.

[0109] By dynamically adjusting the PID parameters, the final control output value of the PID controller can be made more consistent with the current working conditions of the centrifuge. This design enables the PID control to "change with the working conditions" and ultimately achieve stable, efficient and precise control of the centrifuge speed.

[0110] In summary, the approach first acquires time-series data of centrifuge state parameters during the soy protein production process. These state parameters include speed and torque, providing a digital representation of the centrifuge's operating status and continuous information generated by operating condition changes. By analyzing the differential characteristics and fluctuation trends of the speed and torque time-series data, anomaly indicators for each state parameter value at each moment can be dynamically calculated. This design significantly improves the accuracy of anomaly identification, enabling early detection of precursors to operating condition changes and reducing response lag. Then, based on the speed anomaly indicators, abnormal moments can be identified, quickly locating the critical time points where the centrifuge speed suddenly changes, providing a precise basis for subsequent control adjustments. Furthermore, the abnormal time period is determined based on the time-series characteristics of the abnormal moments. Within the abnormal time period, this approach quantifies the dynamic relationship between speed and torque by analyzing the connection characteristics between speed and torque (such as the lag response time of the speed when the torque suddenly changes), similarity characteristics (such as the correlation of the abnormal indicator intensity), and numerical fluctuation characteristics of the speed values. This analysis clarifies the specific requirements for speed regulation, enabling the development of speed regulation indicators that better align with actual operating conditions. Traditional PID control, due to its fixed parameters, is prone to control lag when operating conditions change (e.g., regulation is initiated only after an anomaly occurs). In this embodiment, the PID parameters are dynamically adjusted by combining the current abnormal speed and torque indicators with the previously determined speed regulation indicators. This design enables PID control to adapt to operating conditions, ultimately achieving stable, efficient, and precise control of the centrifuge speed.

[0111] The present invention also provides a centrifuge control system for soybean protein production, see Figure 4 , which shows a system block diagram, including a data acquisition module 401, used to implement step S1 in the above method embodiment; an abnormality analysis module 402, used to implement step S2 in the above method embodiment; a speed regulation analysis module 403, used to implement step S3 in the above method embodiment; and a PID control module 404, used to implement step S4 in the above method embodiment.

[0112] It should be noted that the system provided in the above embodiment is merely illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the centrifuge control system for soy protein production provided in the above embodiment and the centrifuge control method for soy protein production are based on the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0113] See also Figure 5 , which shows a system structure diagram of a centrifuge control system for soy protein production provided by one embodiment of the present invention, including a processor 500, a memory 501, a bus 502 and a communication interface 503, wherein the processor 500, the communication interface 503 and the memory 501 are connected via the bus 502; wherein the memory 501 may include a high-speed random access memory, the bus 502 may be an ISA bus, a PCI bus or an EISA bus, etc., and the processor 500 may be an integrated circuit chip with signal processing capabilities; the memory 501 stores at least one instruction, at least one program, a code set or an instruction set, and when the at least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor, the steps of a centrifuge control method for soy protein production are implemented.

[0114] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0115] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A centrifuge control method for soybean protein production, characterized in that: The method comprises: Obtaining time series data of status parameters of a centrifuge during the soy protein production process, wherein the status parameters include speed and torque; In each state parameter time series data, the difference characteristics between the state parameter values ​​are analyzed and combined with the fluctuation trend of the state parameter values ​​to determine the abnormal indicators of the state parameter values ​​at each moment; in the speed time series data, the abnormal time points are screened out based on the abnormal indicators of the speed values; The abnormal time period is determined based on the time series characteristics of the abnormal time point; in the abnormal time period, the connection characteristics between the torque value and the speed value, the similar characteristics of the abnormal indicators are analyzed, and combined with the numerical fluctuation characteristics of the speed value to determine the speed adjustment index of the centrifuge; Analyze the difference characteristics of abnormal indicators between the speed value and torque value of the centrifuge at the current moment, and adjust the preset PID parameters in combination with the speed adjustment index of the centrifuge to perform PID control on the speed value of the centrifuge at the next moment; The method for obtaining the abnormal indicator includes: In each state parameter time series data, the state parameter value at each moment is taken as the center to determine the corresponding preset neighborhood; In the preset neighborhood corresponding to each state parameter value in each state parameter time series data, the absolute value of the difference between the central state parameter value and each neighboring state parameter value is calculated as the prominence factor of the central state parameter value, and the average level of the prominence factor between the central state parameter value and all corresponding neighboring state parameter values ​​is analyzed to obtain the first abnormal factor of the central state parameter value; In each state time series data, within the preset neighborhood corresponding to each state parameter value, the fluctuation trend between the state parameter values ​​is analyzed to obtain the second abnormal factor of each state parameter value; The first anomaly factor and the second anomaly factor of each state parameter value are integrated to obtain the anomaly index of the state parameter value at each moment; The method for obtaining the speed regulation index includes: In each state parameter time series data, the state parameter value in each abnormal time period is combined into an abnormal data segment; In the two state parameter time series data, the two abnormal data segments corresponding to the same abnormal time period are combined as one data segment; In each data segment combination, the connection characteristics between the state parameter values ​​in the two abnormal data segments and the similar characteristics of the abnormal indicators are analyzed to obtain the adjustment factor corresponding to each data segment combination; In each data segment combination, the variance of all speed values ​​in the abnormal data segment corresponding to the speed is used as the fluctuation factor corresponding to each data segment combination; The fluctuation factors are weighted and summed using the adjustment factors of the data segment combination, and the value obtained by normalizing the weighted result is used as the speed adjustment index of the centrifuge.

2. A centrifuge control method for soybean protein production according to claim 1, characterized in that: The method for obtaining the second abnormal factor includes: In each state parameter time series data, within a preset neighborhood corresponding to each state parameter value, curve segments of all state parameter values ​​are obtained based on the least squares method; On the curve segment, the slope value at each state parameter value is obtained, and the variance of all slope values ​​is calculated as the second abnormal factor of the central state parameter value.

3. The centrifuge control method for soybean protein production according to claim 1, characterized in that: The method for obtaining the regulating factor includes: In each data segment combination, the Pearson correlation coefficient between the state parameter values ​​in the two abnormal data segments is calculated, and the Pearson correlation coefficient is normalized to serve as the influencing factor between the torque and the speed in each data segment combination; In each data segment combination, the absolute value of the difference between the abnormal indicators of the torque value and the speed value at the same time is calculated as the abnormal deviation factor. The sum of all abnormal deviation factors corresponding to the data segment combination is negatively correlated and mapped to the value as the abnormal similarity factor between the torque value and the speed value in each data segment combination; The product of the impact factor and the abnormal similarity factor corresponding to each data segment combination is normalized and used as the adjustment factor corresponding to each data segment combination.

4. The centrifuge control method for soybean protein production according to claim 1, characterized in that: The analysis of the difference characteristics of the abnormal indicators between the speed value and the torque value of the centrifuge at the current moment, and the adjustment of the preset PID parameters in combination with the speed adjustment index of the centrifuge, includes: The preset PID parameters include a preset proportional coefficient, a preset integral coefficient, and a preset differential coefficient; At the current moment, the absolute value of the difference between the centrifuge's speed value and torque value is used as the state difference factor; The sum of the normalized value of the product of the state difference factor and the speed regulation index and a preset constant is used as the regulation degree value; The product of the adjustment degree value and the preset proportional coefficient is used as the adjustment proportional coefficient.

5. A centrifuge control method for soybean protein production according to claim 4, characterized in that: The PID control of the rotation speed value of the centrifuge at the next moment includes: Get the system error at the current moment; The product of the current centrifuge speed value and the adjustment proportional coefficient is used as the first adjustment factor, and the product of the system error and the preset integral coefficient is used as the second adjustment factor; The sum of the first adjustment factor, the second adjustment factor and the preset differential coefficient is used as the input of the PID controller to obtain a control output value, which is used to perform PID control on the rotation speed value of the centrifuge at the next moment.

6. The centrifuge control method for soybean protein production according to claim 1, characterized in that: The method for obtaining the abnormal time point includes: In the speed time series data, the moment corresponding to the speed value where the abnormality index is greater than the preset abnormality threshold is taken as the abnormal moment point.

7. The centrifuge control method for soybean protein production according to claim 1, characterized in that: The method for obtaining the abnormal time period includes: In terms of time series, the abnormal time points that are continuous in time series are grouped into abnormal time periods, thereby obtaining all abnormal time periods, wherein each abnormal time period should contain at least two moments.

8. A centrifuge control system for soy protein production, characterized in that: The invention comprises a processor and a memory, wherein the memory stores at least one instruction, at least one program, code set or instruction set, and when the at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor, the steps of a centrifuge control method for soy protein production as described in any one of claims 1 to 7 are implemented.

Images