An ice cream production parameter self-optimization control system

Abstract

This invention discloses a self-optimizing control system for ice cream production parameters, relating to the field of production control technology. It includes: a working condition construction module, which collects process operating parameters of a continuous freezing section, divides the process into time periods according to different operating stages, and maps the process operating parameters of each time period to the current position of each actuator, forming a working condition page; a state classification module, which extracts freezing state features from the working condition page, classifies the freezing sections, and matches them with historical operation segments, forming a freezing state table and a list of reachable paths; and a sequence optimization module, which uses a working condition transition sequence optimization algorithm to progressively expand the adjustment parameters corresponding to the freezing state table and the list of reachable paths, and uses a sequence forward convergence pruning algorithm to filter out sequences that deviate from the target production zone and disrupt continuous output, forming candidate operation chains. This invention significantly reduces the effective search space and ensures the safety of the adjustment process through the sequence forward convergence pruning algorithm.

Description

Technical Field

[0001] This invention relates to the field of production control technology, and in particular to a self-optimizing control system for ice cream production parameters. Background Technology

[0002] Controlling the process parameters in the continuous freezing section of ice cream is crucial for determining the product's overrun, texture, and continuous output. In a typical industrial continuous freezing process using a freezer drum, the slurry feed flow rate, slurry feed temperature, freezer drum inlet and outlet pressures, freezer drum outlet temperature, agitator speed and current, air flow rate, and the opening positions of actuators such as back pressure valves and refrigeration valves constitute a multivariable, strongly coupled, time-varying operating space. Currently, the operation and control of the continuous freezing section mainly rely on two conventional methods: one is single-loop or cascade loop regulation based on proportional-integral-derivative (PID) controllers, which maintains core parameters such as freezer drum outlet temperature and agitator current near preset values, and continuously fine-tunes each actuator through decentralized independent control loops; the other is open-loop control based on a process formula library, where operators or batch management computers pre-set combinations of parameters such as slurry pump frequency, refrigeration valve opening, and air valve opening based on historical experience for different product formulas, and maintain these combinations constant after production starts or allow only manual correction.

[0003] However, the aforementioned conventional control methods have significant limitations when facing scenarios involving alternating operating conditions, such as the start-up freezing phase, the filling cycle change phase, and the phases before and after formula switching. On the one hand, the decoupling of single-loop regulation is insufficient, and the coordinated action sequence between the refrigeration valve, back pressure valve, slurry pump, and agitator lacks a systematic basis. This can easily lead to oscillations in the outlet temperature of the freezing cylinder, discontinuous discharge flow, or sudden pressure changes during the transition between operating conditions, resulting in fluctuations in product expansion rate and filling line shutdowns. On the other hand, open-loop control based on a fixed formula library lacks the ability to dynamically plan sequences based on the current actual freezing state and available actuator paths. It cannot converge the outlet state to the target production zone with the shortest action sequence while ensuring continuous discharge, nor can it structurally feed back the parameter migration relationships of each successful adjustment into the control knowledge base for reuse. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a self-optimizing control system for ice cream production parameters to solve the problems of slow convergence of the outlet state and easy interruption of continuous discharge caused by the lack of coordination sequence of the actuators and blind search of the sequence space during the multi-condition migration of the continuous freezing section.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a self-optimizing control system for ice cream production parameters, comprising: The working condition construction module collects the process operation parameters of the continuously frozen section, divides the time period according to different operation stages, and associates the process operation parameters of each time period with the current position of each actuator to form a working condition page; The status classification module extracts the frozen status features from the working status page, completes the classification of frozen segments and matches historical operation fragments, and forms a frozen status table and a list of reachable paths. The sequence optimization module uses the working condition migration sequence optimization algorithm to expand the adjustment parameters corresponding to the frozen state table and reachable path list step by step, and uses the sequence forward convergence pruning algorithm to screen out sequences that deviate from the target production zone and disrupt continuous material output, forming candidate operation chains. The continuation execution module issues control actions segment by segment according to the candidate operation chain and checks the running position at the end of each sampling window. It performs continuation rearrangement for incomplete chain segments to form an updated working condition page and continuation operation chain. The rule generation module determines the stable production segment that returns to the target production zone based on the updated work status page and the continuous operation chain, extracts the corresponding parameter migration relationship, and forms a production rule table.

[0007] As a preferred embodiment of the self-optimization control system for ice cream production parameters described in this invention, the step of forming the operating condition page specifically includes: Synchronously read the slurry feed flow rate, slurry feed temperature, pressure before and after the freezing cylinder, freezing cylinder outlet temperature, agitator current, agitator speed, air flow rate, back pressure valve position, refrigeration valve position, and discharge flow rate to obtain a set of timing parameters; The timing parameter group is encapsulated according to the unified sampling time to obtain the running record frame; The operation is divided into stages based on the continuous changes of the operation record frames on the time axis, resulting in a stage sequence; The operation record frames in the stage sequence are matched with the current positions of the refrigeration valve, slurry pump, air valve, back pressure valve and agitator for each time period to form an operation status page.

[0008] As a preferred embodiment of the self-optimization control system for ice cream production parameters described in this invention, the step of dividing the operation into stages to obtain a stage sequence specifically includes: The slurry feeding status, freezing cylinder outlet status, filling cycle status, and formula switching status in the operation record frame are continuously identified to obtain the stage discrimination results. Based on the stage discrimination results, the stages of start-up freezing establishment, stable material output, filling cycle change, and before and after formula switching are distinguished to obtain a stage label set; Arrange the stage label set in chronological order to obtain the stage sequence.

[0009] As a preferred embodiment of the self-optimization control system for ice cream production parameters described in this invention, the formation of the frozen state table and the reachable path list specifically includes: Extract the characteristics of outlet temperature change, freezing cylinder pressure difference, agitator current fluctuation, continuous discharge, and air flow rate and slurry flow rate matching from the operating condition page to obtain the freezing state feature set. Based on the feature set of the frozen state, each time period is classified to obtain the frozen segment category results; The frozen status table is obtained by arranging the time periods in order according to the frozen segment category results. Based on the frozen status table, retrieve historical operation segments with the same frozen segment category and similar actuator location status to obtain a set of historical segments; Extract the action sequence and location migration relationships of the actuators from the historical fragment set to form a list of reachable paths.

[0010] As a preferred embodiment of the self-optimization control system for ice cream production parameters described in this invention, the extraction of the action sequence and position migration relationship of the actuator specifically includes: The actions of the actuators in the historical fragment set are segmented and identified to obtain a set of action segments; The action sequence results are obtained by statistically analyzing the action segments and the sequential action order of the refrigeration valve, slurry pump, air valve, back pressure valve, and agitator in different freezing stages. Based on the action sequence results, the next position that each actuator can enter from its current position is extracted to obtain the position migration result; Based on the location migration results, the movable gears and adjacent gear relationships of each actuator are organized to obtain a path unit set, and the path unit set is arranged in order to form a list of reachable paths.

[0011] As a preferred embodiment of the self-optimization control system for ice cream production parameters described in this invention, the step of using a working condition migration sequence optimization algorithm to progressively expand the adjustment parameters corresponding to the frozen state table and the reachable path list specifically includes: Based on the freeze status table, determine the export status corresponding to the current freeze segment and the preset target production zone to obtain the optimization target; Extract the actuators and corresponding adjustment parameters of the allowed actions of the current frozen segment from the list of reachable paths to obtain the first-level adjustment parameter set; The primary adjustment parameter set is arranged step-by-step according to the optimization objective to obtain the initial migration sequence set; Following the order of refrigeration valve, slurry pump, air valve, back pressure valve and agitator, subsequent adjustment parameters are added step by step to the initial migration sequence set to obtain the extended migration sequence set; The sequence evaluation results are obtained by statistically analyzing the action length, execution order, and rhythm changes of each migration sequence in the extended migration sequence set. Based on the sequence evaluation results, the migration sequences that satisfy the location migration relationship of the implementing agency are retained to form a set of sequences to be screened.

[0012] As a preferred embodiment of the self-optimization control system for ice cream production parameters described in this invention, the step of using a sequence forward convergence pruning algorithm to filter out sequences that deviate from the target production zone and disrupt continuous output specifically includes: For each migration sequence in the sequence set to be screened, a next sampling window prediction segment is established to obtain a set of prediction segments; By examining the direction of change in the outlet status, the maintenance of continuous material discharge, and the maintenance of the filling cycle time for each predicted segment, the sequence discrimination results are obtained. Based on the sequence discrimination results, the migration sequences that are far from the target production zone are deleted, and the migration sequences that cause material discharge interruption, pressure change and air entrainment instability are also deleted, resulting in a converged sequence set; Based on the total motion length, cycle time perturbation degree, and retention status after reaching the target production zone, candidate ranking results are obtained by statistically analyzing the convergent sequence set; Based on the candidate ranking results, the migration sequence with the highest ranking is selected to form a candidate operation chain.

[0013] As a preferred embodiment of the self-optimization control system for ice cream production parameters described in this invention, the step of issuing control actions segment by segment according to the candidate operation chain and verifying the running position at the end of each sampling window specifically includes: The candidate operation chain is decomposed into action to obtain a set of sequential control segments, and control actions are sent to the field controller segment by segment according to the set of sequential control segments to obtain the control segments in execution; A sampling window is maintained for the control segment in execution. The current operation record at the end of the sampling window is obtained. The current operation record is compared with the position status of the actuator in the corresponding time period in the working condition page to obtain the position verification result. Based on the location verification results, the completed chain segment is confirmed, and the chain segment execution result is generated.

[0014] As a preferred embodiment of the self-optimization control system for ice cream production parameters described in this invention, the step of performing continuation rearrangement on incomplete chain segments specifically includes: Based on the execution results of the chain segments, complete and incomplete chain segments are extracted to obtain the chain segment separation results. The completed chain segments in the chain segment separation results are then matched with the current running record to obtain the current position record. Correct the time period position record in the working condition page according to the current position record, and at the same time extract the remaining control actions in the incomplete chain segment to form the updated working condition page and the set of the continuation control segment respectively. The execution order of the remaining control actions is rearranged based on the updated operating condition page and the set of successive control segments to form a successive operation chain.

[0015] As a preferred embodiment of the self-optimization control system for ice cream production parameters described in this invention, the step of forming a production rule table specifically includes: Extract the consecutive running segments corresponding to the continuous operation chain from the updated work status page to obtain a segment set; Based on the segment set, check the continuity of the export status after returning to the target production zone to obtain the stable production segment; Extract the parameter migration sequence, execution duration, filling cycle time correspondence, and discharge holding time from the stable production segment to obtain rule records. Then, classify and organize the rule records to form a production rule table.

[0016] The beneficial effects of this invention are as follows: By forming a working condition page by corresponding the time sequence process parameters of the continuous freezing section with the current position of each actuator according to the operating stage, and extracting freezing state characteristics to form a freezing state table and a list of reachable paths, the adjustment process is limited to the range of reachable paths consistent with the current freezing section category, the migration direction of the actuator, and the order of historical successful actions. In the sequence optimization process, the adjustment parameters are expanded step by step according to the actuator order, and a next sampling window prediction segment is immediately established after each level of parameter is added. Forward checks are performed on the direction of outlet state change, the continuous discharge status, and the filling cycle time status. Migration sequences that are far from the target production zone, disrupt continuous discharge, or do not meet the actuator position migration relationship are continuously deleted, so that sequence convergence is transformed from unified screening after full expansion to expansion. The gradual contraction during the process reduces the accumulation of invalid searches under multi-condition migration and improves the consistency between candidate operation chains and the executable state of field equipment. At the same time, during the execution of candidate operation chains, the running position is checked at the end of each sampling window and incomplete chain segments are reordered for continuation. This allows completed actions to be retained and incomplete parts to be continued locally by combining with the updated condition page. By organizing the stable production segments that successfully return to the target production zone, the parameter migration order, execution time, and filling cycle corresponding relationships into a production rule table, the system forms a closed-loop self-optimization mechanism of current condition identification, restricted sequence generation, gradual forward convergence, execution feedback correction, and stable production rule writing. This makes it more suitable for multi-condition migration control scenarios in the continuous freezing process of ice cream where the establishment segment, stable production segment, switching segment, and disturbance segment alternate. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A schematic diagram of a self-optimizing control system for ice cream production parameters.

[0019] Figure 2 Flowchart for building and categorizing the status of the work status page.

[0020] Figure 3 The flowchart shows the sequence optimization and forward convergence pruning process.

[0021] Figure 4 A flowchart is generated to continue the execution and production rules. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 This is one embodiment of the present invention, which provides a self-optimizing control system for ice cream production parameters, including the following steps: The working condition construction module collects the process operation parameters of the continuously frozen section, divides the time period according to different operation stages, and associates the process operation parameters of each time period with the current position of each actuator to form a working condition page.

[0026] After the production line enters the continuous freezing section, the central control computer first synchronously receives process operating parameters from the field controller, including slurry feed flow rate, slurry feed temperature, pressure before and after the freezing cylinder, freezing cylinder outlet temperature, agitator current, agitator speed, air flow rate, back pressure valve position, refrigeration valve position, and discharge flow rate. All parameters collected at the same sampling time are combined into a single operating record frame.

[0027] In this embodiment, the sampling time is provided by the real-time clock of the central control computer, and the sampling interval is 1 second.

[0028] It should be noted that the process operating parameters are obtained from the communication interface between the central control computer and the controllers of each actuator via the fieldbus. Each operation record frame is accompanied by a timestamp accurate to milliseconds to ensure the timing accuracy of subsequent stage divisions. The operation record frame is composed as follows: all process operating parameters are sequentially stored in a data frame structure. The data frame structure includes a frame number, a timestamp, eleven parameter value fields, and a check field. The frame number increments sequentially from the initial value at the start of the current batch. The check field is used to verify data integrity during transmission. The data frame structure is continuously stored in the central control computer's memory in the form of a circular buffer to ensure that it does not overflow due to the increase in data volume during the continuous operation of the continuous freeze section.

[0029] In this embodiment, the slurry feed flow rate refers to the volumetric flow rate of the ice cream mixture pumped into the freezing cylinder per unit time; the slurry feed temperature refers to the temperature of the slurry before entering the freezing cylinder, typically ranging from +2°C to +5°C; the pressures before and after the freezing cylinder refer to the slurry pressures at the inlet and outlet ends of the freezing cylinder, respectively; the freezing cylinder outlet temperature refers to the temperature of the ice cream leaving the freezing cylinder, typically ranging from -3°C to -6°C; the agitator current refers to the operating current of the motor driving the agitator inside the freezing cylinder; the agitator speed refers to the rotational speed of the agitator shaft; the air flow rate refers to the volumetric flow rate of air entering the freezing cylinder simultaneously with the slurry; the back pressure valve position refers to the opening position of the back pressure valve installed at the outlet end of the freezing cylinder, expressed as a percentage; the refrigeration valve position refers to the opening position of the refrigeration valve controlling the flow of refrigerant into the freezing cylinder jacket, expressed as a percentage; and the discharge flow rate refers to the volumetric flow rate of ice cream discharged from the freezing cylinder outlet per unit time.

[0030] As the central control computer continuously receives and stores operation record frames, it divides the production process of the current batch into stages based on the continuous changes of the operation record frames on the time axis, thus obtaining a stage sequence.

[0031] The specific method for dividing the stages is as follows: the central control computer analyzes the characteristics of the changes of each parameter in the stored operation record frame over time and identifies the switching points between different operation stages.

[0032] In this embodiment, the operation phase includes four production periods: the start-up and freezing phase, the stable discharge phase, the filling cycle change phase, and the phases before and after formula switching.

[0033] The start-up freezing stage refers to the time interval from when the freezing cylinder begins to cool until the first batch of ice cream reaches the discharge conditions. During this stage, the freezing cylinder outlet temperature gradually decreases from the slurry feed temperature to the target discharge temperature range, the agitator current gradually increases to the stable operating current range, and the back pressure valve is adjusted from the initial position to the operating opening.

[0034] Entering the stable discharge stage refers to the time interval during which the operating parameters of each process remain stable within the target range. During this stage, the temperature fluctuation of the refrigeration cylinder outlet does not exceed ±0.5℃, the fluctuation of the agitator current does not exceed ±2%, and the back pressure valve position and the refrigeration valve position remain relatively stable.

[0035] The filling cycle change stage refers to the time interval during which the discharge flow rate is adjusted accordingly and the back pressure valve position changes compensatorily when the production cycle of the downstream filling equipment changes.

[0036] The period before and after formula switching refers to the time interval during which the ice cream formula changes, and the flow rate, temperature, or air flow rate of the syrup are switched.

[0037] The identification of the above four types of production periods is achieved by analyzing the numerical variation of each parameter in the operation record frame. Taking the refrigeration cylinder outlet temperature as an example, the central control computer calculates the standard deviation of the refrigeration cylinder outlet temperature within the sliding window. When the standard deviation is greater than the set standard deviation change threshold, the parameter is determined to be in an unstable state; when the standard deviation is less than the set standard deviation change threshold, the parameter is determined to have entered a stable state.

[0038] It should be noted that the standard deviation change threshold is determined based on the maximum allowable fluctuation range of the freezer outlet temperature when entering the stable discharge stage. When entering the stable discharge stage, the allowable fluctuation range of the freezer outlet temperature is ±0.5℃, assuming the sliding window includes 10 sampling times. Considering the influence of refrigerant supply fluctuations, minor changes in slurry composition, and measurement noise on temperature fluctuations in actual production, a standard deviation change threshold of 0.5 to 0.8 times the theoretical upper limit is used. In this embodiment, the standard deviation change threshold is set to 0.10℃.

[0039] The transition time from an unsteady to a steady state at the freezer outlet temperature is calculated by comparing the difference between the standard deviation of the current sliding window and the standard deviation of the adjacent previous sliding window. The expression is as follows: ; in, This indicates the time coordinates corresponding to the stage switching point; Indicates the first The standard deviation of the outlet temperature of the freezer cylinder within each sliding window characterizes the degree of temperature fluctuation within that time interval. Indicates the duration of the sliding window; Indicates the first The start time of each sliding window is determined by the timestamp of the first running record frame within the sliding window; This represents the preset standard deviation change threshold, which is the threshold value taken at the switching time. The window and the first The average value at the center of each window.

[0040] In this embodiment, the time length of the sliding window is set to 10 sampling times, and the step length between two adjacent sliding windows is set to 5 sampling times.

[0041] When the difference in standard deviation between two adjacent sliding windows exceeds a preset standard deviation change threshold, it is determined that the first... The window and the first The boundary area between the windows underwent a phase transition.

[0042] Similarly, the sliding window characteristic values ​​of parameters such as stirrer current and freezer outlet pressure are calculated, and the location of the switching point for each stage is determined by combining the change characteristics of multiple parameters. When one or more key parameters change significantly at the same time, the switching point is marked on the time axis. All operation record frames between two adjacent switching points on the time axis are classified into the same stage. After the above processing, the production process of the current batch is divided into several consecutive time periods, each time period corresponding to a certain operation stage, which are arranged in chronological order to obtain the stage sequence.

[0043] The operation record frames in the stage sequence are matched with the current positions of the refrigeration valve, slurry pump, air valve, back pressure valve, and agitator on a time-by-time basis. Specifically, for each stage in the stage sequence, the actual position feedback values ​​of each actuator corresponding to each sampling moment within the time interval of that stage are retrieved from the storage unit of the central control computer. The position feedback value of the refrigeration valve comes from the position sensor installed on the refrigeration valve; the position feedback value of the slurry pump comes from the operating frequency of the frequency converter driving the slurry pump, and the operating frequency directly corresponds to the discharge output of the slurry pump; the position feedback value of the air valve comes from the opening sensor of the air valve; the position feedback value of the back pressure valve comes from the valve position transmitter of the back pressure valve; and the position feedback value of the agitator comes from the speed feedback signal of the agitator motor. Each operation record frame is paired one-to-one with the position feedback value of each actuator at its sampling moment, and is appended to the operation record frame in the form of an extended field, or a mapping relationship between the operation record frame and the actuator position is established through an index.

[0044] After the above processing, the current batch's operating status page is formed. The operating status page contains all operation record frames organized according to the stage sequence. Each frame contains the values ​​of all process operation parameters and the current position value of the actuator, along with a timestamp, stage identifier, and frame sequence number. The operating status page does not simply store the values ​​of process operation parameters, but rather registers the relationship between each operation record frame and the current equipment position.

[0045] The status classification module extracts the frozen status features from the operating status page, classifies the frozen segments, matches them with historical operation fragments, and forms a frozen status table and a list of reachable paths.

[0046] The central control computer extracts the characteristics of outlet temperature change, continuous pressure difference of freezing cylinder, agitator current fluctuation, continuous discharge, and matching characteristics of air flow and slurry flow from the operating status page to obtain the frozen state feature set.

[0047] The extraction method for outlet temperature variation characteristics is as follows: In the operating condition page, for each stage of the current batch, read the freezer outlet temperature values ​​from all operation record frames within that stage, and calculate the change in freezer outlet temperature between adjacent frames in chronological order. Statistical analysis is then performed on the absolute value sequence of the changes to determine whether a stable temperature plateau has been formed.

[0048] In this embodiment, when the absolute value of the change in the outlet temperature of the freezer cylinder does not exceed 0.05℃ within 30 consecutive sampling times, the outlet temperature change characteristic is determined to be that a stable plateau has been formed; when the absolute value of the change in any two adjacent sampling times exceeds 0.05℃ within 30 consecutive sampling times, the outlet temperature change characteristic is determined to be that a stable plateau has not been formed or that it is in a decreasing process.

[0049] The temperature change trend is quantified by calculating the first-order difference mean of the refrigeration cylinder outlet temperature over time during the current period. The expression is as follows: ; in, This represents the average absolute change in the outlet temperature of the freezer cylinder during the current time period; This indicates the total number of running record frames contained in the current time period; Indicates the first The temperature value of the freezer outlet in the frame.

[0050] when When the temperature is less than or equal to 0.05℃, the central control computer determines that the outlet temperature change characteristic is a stable platform state; when When the temperature exceeds 0.05℃, the central control computer determines that the outlet temperature change is in an unstable state.

[0051] The extraction method for the continuous feature of the freezer cylinder pressure difference is as follows: In the operating condition page, for each stage period of the current batch, read the pressure before and after the freezer cylinder in all running record frames within that period, and calculate the freezer cylinder pressure difference corresponding to each frame, that is, the pressure before the freezer cylinder minus the pressure after the freezer cylinder.

[0052] In this embodiment, when the pressure difference of the cryogenic cylinder is greater than 0.2 bar for 20 consecutive sampling times and the pressure difference change between two adjacent frames does not exceed 0.05 bar, the cryogenic cylinder pressure difference is determined to be a continuous pressure difference; otherwise, it is determined that a continuous pressure difference has not been formed or the pressure difference has not been established.

[0053] The method for extracting the agitator current fluctuation characteristics is as follows: In the operating condition page, for each stage period of the current batch, read the agitator current values ​​from all running record frames within that period, and calculate the coefficient of variation of the agitator current within that period. The coefficient of variation is the ratio of the standard deviation to the mean, expressed as a percentage. When the coefficient of variation is less than or equal to 3%, the agitator current fluctuation characteristic is determined to have moved out of the start-up fluctuation range; when the coefficient of variation is greater than 3%, the agitator current fluctuation characteristic is determined to still be within the start-up fluctuation range.

[0054] Among them, the start-up fluctuation stage is a typical characteristic of the start-up freezing stage. In this stage, the agitator current increases rapidly as the viscosity of the slurry in the freezing cylinder increases, and the fluctuation amplitude is significantly higher than that in the stable discharge stage.

[0055] The extraction method for continuous discharge characteristics is as follows: In the operating status page, for each stage of the current batch, read the discharge flow rate values ​​in all operation record frames within that stage, and determine whether the discharge flow rate is consistently greater than zero. When the discharge flow rate is greater than 0.5 liters per hour for 20 consecutive sampling times within that stage and there are no breakpoint frames with zero flow, the continuous discharge characteristic is determined to be continuous; otherwise, it is determined that the discharge is not yet continuous. Zero-value breakpoints in the discharge flow rate usually occur in the early stage of the start-up freeze phase or during the pause of the filling equipment.

[0056] The method for extracting the airflow and slurry flow matching characteristics is as follows: In the operating condition page, for each stage of the current batch, read the airflow and slurry feed flow in all operation record frames within that stage, and calculate the ratio of airflow to slurry feed flow in each frame, i.e., the instantaneous expansion ratio. When the deviation of the instantaneous expansion ratio from its mean value within 20 consecutive sampling times does not exceed 5%, the airflow and slurry flow matching characteristics are determined to have entered a stable matching segment; otherwise, it is determined that it has not yet entered a stable matching segment.

[0057] Based on the frozen state feature set, the time periods in the working condition page are classified to obtain the frozen segment category results. The specific classification rules are as follows: When the outlet temperature change characteristic is that a stable platform has not yet been formed, the cryogenic cylinder pressure difference characteristic is that a continuous pressure difference has not yet been formed, the agitator current fluctuation characteristic is that it is still in the start-up fluctuation stage, the discharge continuity characteristic is that the discharge is not yet continuous, and the air flow and slurry flow matching characteristic is that it has not yet entered the stable matching stage, the period is classified as the establishment stage; when the outlet temperature change characteristic is that a stable platform has been formed, the cryogenic cylinder pressure difference characteristic is that a continuous pressure difference has been formed, the agitator current fluctuation characteristic is that it has left the start-up fluctuation stage, the discharge continuity characteristic is that the discharge is continuous, and the air flow and slurry flow matching characteristic has entered the stable matching stage, the period is classified as the stable production stage; when the outlet temperature change characteristic is that a stable platform has been formed but the cryogenic cylinder pressure difference characteristic shows stage fluctuations, and the discharge continuity characteristic is related to the filling cycle change period, the period is classified as the switching stage; when any of the five characteristics shows an irregular deviation, and the deviation exceeds the corresponding normal fluctuation range but recovers within a short time window, the period is classified as the disturbance stage.

[0058] In this embodiment, the switching section corresponds to the filling cycle change stage, and the disturbance section corresponds to short-term parameter deviations caused by external disturbances such as short-term fluctuations in incoming material temperature and short-term changes in cooling water temperature. The setup section corresponds to the start-up freeze setup stage, and the stable production section corresponds to the stable material discharge stage.

[0059] The time periods are sequentially arranged according to the frozen segment category results to obtain the frozen status table. The frozen status table is a time sequence table, listing all time periods from the start time to the current time of the current batch from left to right. Each time period is accompanied by the corresponding frozen segment category and indicates the average position value of the actuators within each time period. The average position value of the actuators includes the average opening degree of the refrigeration valve, the average operating frequency of the slurry pump, the average opening degree of the air valve, the average opening degree of the back pressure valve, and the average rotational speed of the agitator.

[0060] Based on the frozen status table, historical operation segments with the same frozen segment category and similar actuator position status are retrieved to obtain a historical segment set. These historical operation segments are stored in the central control computer's process database. Each historical operation segment contains a working status page of a complete or partial production process and the corresponding final product quality indicators. During the retrieval, the central control computer first filters the process database for all historical operation segments with the same frozen segment category as the current batch and current time period.

[0061] The central control computer reads the preset number of operation record frames corresponding to the starting position of the current batch and the current time period, extracts the slurry feed temperature, freezing cylinder outlet temperature, freezing cylinder pressure difference, discharge flow rate and air flow rate to form the initial condition group for the current time period; and reads the preset number of operation record frames corresponding to the starting position of the time period for each selected historical operation segment, extracts the same parameters to form the historical initial condition group.

[0062] The central control computer compares the historical initial condition set with the initial condition set for the current period, retaining historical operation segments where the deviations of each parameter fall within their respective allowable ranges and the direction of change is consistent, thus obtaining a set of initial condition matching segments. Next, for each historical operation segment in the initial condition matching segment set, the Euclidean distance between its actuator position state and the average position value of the actuator in the current period is calculated. The Euclidean distance is calculated based on five dimensions: average opening degree of the refrigeration valve, average operating frequency of the slurry pump, average opening degree of the air valve, average opening degree of the back pressure valve, and average rotational speed of the agitator. The values ​​of each dimension are pre-normalized to the range of 0 to 1. Several historical operation segments with the smallest Euclidean distances in the initial condition matching segment set are selected; in this embodiment, the five historical operation segments with the smallest distances are selected to form a historical segment set.

[0063] In this embodiment, the preset number of running record frames is 10 sampling times after the start of the current time period; the allowable deviation of the slurry feed temperature is ±0.5℃, the allowable deviation of the freezing cylinder outlet temperature is ±0.5℃, the allowable deviation of the freezing cylinder pressure difference is ±0.05 bar, the allowable deviation of the discharge flow rate is ±5% of the average value corresponding to the current time period, and the allowable deviation of the air flow rate is ±5% of the average value corresponding to the current time period.

[0064] The central control computer extracts the action sequence and position migration relationship of the actuators based on the historical fragment set, and forms a list of reachable paths. The extraction process consists of the following four steps: The first step involves segmenting and identifying the actuator actions in the historical data set to obtain an action segment set. For each historical operation segment in the set, the central control computer reads the change in the actuator position feedback value over time recorded in its operating status page. Intervals where the actuator position changes continuously are identified as an action segment, while intervals where the actuator position remains unchanged are considered a steady-state segment. Each action segment records the start time, end time, the specific actuator involved, the direction of the action, and the position values ​​before and after the action. After segmenting and identifying all historical operation segments in the historical data set, all identified action segments are aggregated to obtain the action segment set.

[0065] The second step involves statistically analyzing the sequence of actions of the refrigeration valve, slurry pump, air valve, back pressure valve, and agitator in different freezing stages, according to the action segment set. The central control computer categorizes and summarizes the action segments in the action segment set based on the freezing stage category. Within each freezing stage category, the frequency of each actuator's action segment on the time axis is statistically analyzed. For example, during the transition from the stable production stage to the switching stage, the number of times the action sequence of adjusting the back pressure valve before adjusting the refrigeration valve, and the number of times the action sequence of adjusting the refrigeration valve before adjusting the back pressure valve, occurs is statistically analyzed. The action sequence with the highest frequency is determined as the typical action sequence under that freezing stage category. For the transition from the establishment stage to the stable production stage, the sequence of actions such as starting the slurry pump, opening the refrigeration valve, opening the air valve, accelerating the agitator, and adjusting the back pressure valve is statistically analyzed. These statistical results are recorded as the action sequence results.

[0066] The third step involves extracting the next position that each actuator can enter from its current position based on the action sequence results, thus obtaining the position migration results. The central control computer determines the position range that each actuator is allowed to migrate to from its current position based on the direction and amplitude of each actuator's action recorded in the action sequence results, combined with the actual distribution of position values ​​before and after that action in the historical data set. For continuously regulating actuators such as refrigeration valves, back pressure valves, and slurry pump frequency converters, the position range is represented by a discrete interval of opening degree or frequency.

[0067] In this embodiment, the opening of the refrigeration valve is in 5% increments, the frequency of the slurry pump is in 2Hz increments, the opening of the back pressure valve is in 5% increments, the opening of the air valve is in 10% increments, and the speed of the agitator is in 20 revolutions per minute increments.

[0068] The central control computer does not retain all adjacent gear migrations that have appeared in the historical segment set. Instead, it first reads the operation record frames before and after each historical action segment corresponding to the adjacent gear migration, checks whether the migration still maintains continuous material output, continuous pressure difference of the freezing cylinder, and stable matching relationship between air flow and slurry flow after actual execution, and judges whether the migration is a valid operation by combining the final product quality indicators of the corresponding historical operation segment.

[0069] For historical action segments that experience interruptions in material discharge, sudden pressure changes, air entrainment instability, correction of the same actuator by subsequent reverse actions, failure to continue to a stable production segment, or abnormal quality indicators of the corresponding final product after the action is executed, the central control computer identifies the migration of the adjacent gear corresponding to the historical action segment as a historical abnormal operation and removes it from the statistical results.

[0070] After removing historical abnormal operations, the central control computer then organizes the historical action segments of each actuator under the same frozen segment category that migrate from a certain current position to an adjacent position, retaining the migration paths in the historical segments that can form a continuous and effective working condition evolution relationship, as the possible migration directions of the actuator in the current position.

[0071] Among them, the continuous effective working condition evolution relationship refers to the following: after a certain actuator completes the migration of the adjacent gear, the subsequent operation records in the corresponding historical segment do not show material discharge interruption, pressure change and air entrainment instability, and the migration can be consistent with the typical action sequence under the current frozen segment category, and maintain stable material discharge or connect to the stable production segment in the subsequent period.

[0072] The fourth step involves organizing the movable positions and adjacent positions of each actuator according to the position migration results, obtaining a set of path units, and then sequentially arranging these path units to form a reachable path list. A path unit is defined as a single operation that starts from the current position of an actuator and reaches the next position via an adjustment action. The central control computer converts all movable directions of each actuator into path units, each containing the actuator name, starting position, target position, and direction of action. All path units under the same frozen segment category are arranged according to the order of actions recorded in the action sequence results, forming a sequence of optional operations starting from the current equipment state. The resulting arrangement is the reachable path list, which clearly indicates which actuator (refrigeration valve, slurry pump, air valve, back pressure valve, agitator) can be moved first, which can be moved last, and which position can be reached in each step at the current position.

[0073] The sequence optimization module uses the working condition migration sequence optimization algorithm to expand the adjustment parameters corresponding to the frozen state table and reachable path list step by step, and uses the sequence forward convergence pruning algorithm to screen out sequences that deviate from the target production zone and disrupt continuous material output, forming candidate operation chains.

[0074] The central control computer determines the outlet status and target production zone corresponding to the current frozen segment based on the freeze status table, thus obtaining the optimization target. The freeze status table records the category of the frozen segment to which the current time period belongs, namely, the establishment segment, stable production segment, switching segment, or disturbance segment, as well as the average position value of the actuators within each time period. The outlet status is jointly characterized by three core parameters: the outlet temperature of the freezer cylinder, the discharge flow rate, and the agitator current.

[0075] The central control computer reads the product formula identifier corresponding to the current batch, and extracts the target production zone of the freezer outlet temperature, the target production zone of the discharge flow rate, and the target production zone of the agitator current corresponding to the product formula from the formula target mapping table based on the product formula identifier. The formula target mapping table is calibrated offline by the process engineer according to the product standard and stored in the central control computer. The system does not automatically adjust the range of the target production zones that have been read during operation.

[0076] When the current time period corresponding to the frozen status table belongs to the stage before and after formula switching, the central control computer confirms whether the product formula identifier has changed by combining the formula switching status in the operation record frame; after confirming that the product formula identifier has changed, it rereads the target production zone corresponding to the changed product formula from the formula target mapping table, and uses the read target production zone as the basis for distinguishing the current time period and subsequent time periods.

[0077] In this embodiment, the target production range for the outlet temperature of the freezer cylinder is defined as the range of -4.5°C to -5.5°C, the target production range for the discharge flow rate is set according to the cycle time of the filling equipment, for example, 180 liters per hour to 220 liters per hour under normal cycle time, and the target production range for the agitator current is 28 amperes to 32 amperes.

[0078] The specific optimization objective is to adjust the current export status to within the target production zone, while ensuring that the filling cycle remains undisrupted during the adjustment process; that is, to prevent the discharge flow from being interrupted or exceeding the allowable receiving range of the filling equipment due to parameter adjustments.

[0079] Extract the actuators and corresponding adjustment parameters allowed for the current frozen segment from the reachable path list to obtain the primary adjustment parameter set. The reachable path list shows the next gear that each actuator can move to from the current position, as well as the recommended order of which actuator to move first and which to move later.

[0080] In this embodiment, following the typical action sequence recorded in the reachable path list, the adjustment parameters corresponding to the refrigeration valve are extracted first as the primary adjustment parameter set. Specifically, the primary adjustment parameter set includes the various opening values ​​that the refrigeration valve can migrate to from its current opening. For example, if the current refrigeration valve opening is 60%, the primary adjustment parameter set includes three selectable positions: 55%, 60%, and 65%. If the refrigeration valve has no available positions at its current location, the refrigeration valve is skipped, and the primary adjustment parameter set is constructed starting from the adjustment parameters corresponding to the slurry pump.

[0081] The initial migration sequence set is obtained by single-step arrangement of the primary adjustment parameter set according to the optimization objective. The single-step arrangement method is as follows: each selectable position in the primary adjustment parameter set is treated as a separate action step, combined with the current actuator position to form an initial migration sequence. Taking a refrigeration valve as an example, the initial migration sequence set includes three initial migration sequences: refrigeration valve adjusted from 60% to 55%, refrigeration valve remaining at 60%, and refrigeration valve adjusted from 60% to 65%. Each initial migration sequence records the actuator name, starting position, target position, and direction of action.

[0082] Following the sequence of refrigeration valve, slurry pump, air valve, back pressure valve, and agitator, subsequent adjustment parameters are added step-by-step to the primary retention sequence set to form the current-level migration sequence set. Immediately after each addition, the next sampling window prediction and sequence check are performed. Specifically, based on the initial migration sequence set, the central control computer first establishes a primary prediction segment for each migration sequence in the initial set, obtaining a primary prediction segment set. Then, it checks the outlet state change direction, continuous discharge maintenance, and filling cycle time maintenance for each sequence, deleting migration sequences that are far from the target production zone and disrupt continuous discharge, thus obtaining the primary retention sequence set. The adjustment parameters corresponding to the slurry pump are extracted from the reachable path list, and each selectable gear of the slurry pump is appended to the end of each primary retention sequence, forming the current-level migration sequence set containing both refrigeration valve and slurry pump actions.

[0083] After forming the current level migration sequence set, the central control computer establishes the next sampling window prediction segment and checks the direction of outlet status change, continuous material discharge maintenance, and filling cycle time maintenance one by one. Migration sequences that are far from the target production zone and disrupt continuous material discharge are deleted, while those that satisfy the actuator position migration relationship are retained. This process is repeated sequentially, adding adjustment parameters for the air valve, back pressure valve, and agitator until the adjustment parameters for all five actuators are added and checked at each level, resulting in an extended migration sequence set. Each sequence in the extended migration sequence set is a migration sequence continuously retained during the addition process at each level, containing one or more actuator action steps arranged in chronological order. The number of action steps is determined by the actual number of levels added.

[0084] The central control computer statistically analyzes the action length, execution sequence, and cycle time variation of each migration sequence in the extended migration sequence set to obtain sequence evaluation results. Action length refers to the total number of action steps in a migration sequence; more action steps indicate more frequent intervention in the production process. Execution sequence refers to the order in which the actions of each actuator are performed, compared with the typical action sequence given in the reachable path list, and the degree of consistency between the execution sequence and the typical action sequence is recorded. Cycle time variation refers to the magnitude and duration of possible changes in discharge flow rate during the execution of the migration sequence, used to assess the impact on downstream filling equipment. Combining fault prediction and health management analysis, the central control computer also considers the cumulative operating time and current health status indicators of each actuator when statistically analyzing cycle time variation, prioritizing the elimination of migration sequences that may cause actuators to exceed the recommended operating range or accelerate wear, ensuring that candidate operation chains not only meet current adjustment needs but also match the long-term healthy operating status of the production equipment.

[0085] Based on the sequence evaluation results, migration sequences that satisfy the actuator position migration relationship are retained to form a set of sequences to be screened. Satisfying the actuator position migration relationship means that the target position of each step in the sequence is within the allowed migration direction given in the reachable path list, and the change in actuator position between two adjacent steps conforms to the physical continuity constraint, without any jump gear changes.

[0086] The forward convergence pruning algorithm is used to process the set of sequences to be screened. Specifically, a next sampling window prediction segment is established for each migration sequence in the set of sequences to be screened, resulting in a set of prediction segments.

[0087] The specific method for establishing the next sampling window prediction segment is as follows: the central control computer does not linearly superimpose all action steps in the migration sequence to form a single prediction result, but instead establishes a corresponding current level prediction segment for each level of the migration sequence formed by each action step in the migration sequence; each current level prediction segment corresponds to the parameter response of only one newly added action step within a sampling window.

[0088] For the outlet temperature of the refrigeration cylinder, the central control computer calculates the current predicted value using a step-by-step recursive method, the expression of which is: ; in, Indicates the first The level of action steps corresponds to the predicted value of the freezer outlet temperature within the sampling window. This indicates the predicted temperature of the freezer outlet within the sampling window corresponding to the previous action step. This represents the measured value of the freezer outlet temperature at the current sampling time. Indicates the first The change in position of the actuator corresponding to each action step. Indicates the relationship with the first Temperature response sensitivity corresponding to each action step.

[0089] The prediction of discharge flow rate and agitator current adopts the same step-by-step recursive method, and is calculated based on the prediction value of the previous stage and the response sensitivity corresponding to the current action step, respectively.

[0090] The central control computer checks the direction of change in the outlet status, the maintenance of continuous material discharge, and the maintenance of filling cycle time step by step according to the predicted segment set from the previous level action step to the next level action step, and obtains the sequence discrimination result. The check of the direction of change in the outlet status is based on the target production belt corresponding to the current product formula identifier as the comparison benchmark.

[0091] The inspection of the direction of change in the outlet status is as follows: determine whether the predicted value of the freezer outlet temperature is moving towards the target production zone relative to the current value. If the current freezer outlet temperature is below the lower limit of the target production zone and the predicted value is higher than the current value, or if the current freezer outlet temperature is above the upper limit of the target production zone and the predicted value is lower than the current value, then the direction of change in the outlet status is determined to be towards the target production zone; otherwise, it is determined to be moving away from the target production zone.

[0092] The continuous discharge status is checked against the target production zone corresponding to the current product formula identifier, and the judgment is made in conjunction with the receiving range corresponding to the current filling cycle. The filling cycle maintenance status is checked against the receiving relationship between the target production zone corresponding to the current product formula identifier and the current cycle of the downstream filling equipment. If the current time period is before or after a formula change, the central control computer will reread the target production zone after the formula change, and then perform checks on the outlet status change direction, continuous discharge status, and filling cycle maintenance status for the subsequent migration sequence.

[0093] The central control computer deletes migration sequences that are far from the target production zone based on the sequence discrimination results, and also deletes migration sequences that cause material discharge interruption, pressure change and air entrainment instability, thus obtaining a converged sequence set.

[0094] When a migration sequence deviates from the target production zone, experiences material discharge interruption, pressure surge, or air entrainment instability in a certain prediction stage, the central control computer deletes the migration sequence and stops further inspection of subsequent levels of the migration sequence; only migration sequences that maintain convergence toward the target production zone in each prediction stage without disrupting the continuous material discharge and filling cycle relationship are retained, resulting in a convergence sequence set.

[0095] Migration sequences causing discharge interruption are those where the predicted discharge flow rate is zero or below 0.5 liters per hour in the next sampling window; migration sequences causing pressure abrupt changes are those where the predicted differential pressure change in the freezer exceeds 0.5 bar after a single-step action; migration sequences causing air entrainment instability are those where the predicted air flow rate to slurry feed flow rate ratio deviates from its historical average by more than 10%. After removing the above three types of sequences, as well as sequences far from the target production zone, from the set of sequences to be screened, the remaining sequences form the convergent sequence set.

[0096] The central control computer calculates the total action length, cycle time disturbance, and retention status after reaching the target production zone based on the converged sequence set, obtaining the candidate ranking results. The total action length is the total number of action steps contained in each sequence in the converged sequence set. The cycle time disturbance is quantified by the cumulative deviation of the predicted discharge flow rate from the expected flow rate of the filling equipment during sequence execution. Retention status after reaching the target production zone refers to whether the freezer outlet temperature has entered the target production zone at the end of the prediction window and whether the temperature fluctuation over the last five sampling times within the prediction window is less than 0.1℃. During ranking, the central control computer further considers the historical fault records and remaining service life assessment results of the high-frequency actuators involved in each migration sequence. Sequences involving frequent recent faults or actuators with remaining service life below the warning threshold are ranked at a lower priority position in the candidate ranking results to reduce the risk of adjustment failure due to potential actuator failures.

[0097] Based on the candidate ranking results, the migration sequence with the highest ranking is selected to form a candidate operation chain. The migration sequence with the highest ranking is the one with the shortest total action length, the smallest cycle time disturbance, and the best maintenance performance after entering the target production zone. The candidate operation chain is a complete sequence of actions with a pre-arranged order, clearly listing the actuators, action sequence, and target gear level of each step in the refrigeration valve, slurry pump, air valve, back pressure valve, and agitator that need to perform the actions.

[0098] The continuation execution module issues control actions segment by segment according to the candidate operation chain and checks the running position at the end of each sampling window. It performs continuation rearrangement for incomplete chain segments to form an updated working status page and continuation operation chain.

[0099] The candidate operation chain is decomposed into a set of sequential control segments. The specific method of action decomposition is as follows: adjacent steps in the candidate operation chain that involve the same actuator and have continuous action direction are merged into one control segment, and action steps involving different actuators are split into independent control segments.

[0100] Taking a candidate operation chain containing three steps—adjusting the cooling valve from 60% to 65%, adjusting the slurry pump from 40 Hz to 42 Hz, and adjusting the back pressure valve from 50% to 45%—as an example, the actions are decomposed into three sequential control segments: the first control segment is the cooling valve adjustment action, the second control segment is the slurry pump adjustment action, and the third control segment is the back pressure valve adjustment action. Each control segment records the actuator name, starting position, target position, direction of action, and estimated execution time.

[0101] In this embodiment, the estimated execution time of the control segment is set according to the response speed of the actuator. The valve position adjustment of the refrigeration valve and the back pressure valve is expected to be completed within 2 seconds, the frequency conversion adjustment of the slurry pump is expected to be completed within 1 second, the air valve adjustment is expected to be completed within 1 second, and the agitator speed adjustment is expected to be completed within 3 seconds. The estimated execution time of the control segment is used to characterize the action completion time required for the actuator to move from the starting position to the target position. One control segment corresponds to one sampling window. The sampling window is used to cover the complete process of issuing the command, the actuator reaching the position, and observing the response of the process operation parameters in the control segment.

[0102] The central control computer sends control actions to the field controller segment by segment according to the sequence control segment set. The sending method is as follows: according to the arrangement order of the control segments in the sequence control segment set, the adjustment command corresponding to the first control segment is sent to the corresponding actuator controller. After receiving the command, the field controller drives the actuator to move to the target position.

[0103] When the first control segment command is issued, the central control computer synchronously starts a sampling window corresponding to the control segment. After the actuator completes the movement from the starting position to the target position in the first part of the sampling window, the central control computer keeps the actuator running at the target position until the end of the sampling window, and uses the remaining time of the sampling window to collect the response results of the process operation parameters to the control segment. If the actuator has not reached the target position by the end of the sampling window, the control segment is recorded as an incomplete chain segment.

[0104] After the first control segment command is issued, the control segment becomes an executing control segment. During the execution of the control segment, the central control computer does not issue new adjustment commands to other actuators, waits for the actuators to move to the target position within the expected execution time of the control segment, and continues to maintain the position after the actuators reach the target position until the end of the sampling window.

[0105] After the sampling window expires, the central control computer synchronously reads the current process operating parameters and actuator position feedback values ​​from the field controllers, forming a current operating record. This record includes values ​​for slurry feed flow rate, slurry feed temperature, pressure before and after the freezing cylinder, freezing cylinder outlet temperature, agitator current, agitator speed, air flow rate, back pressure valve position, refrigeration valve position, and discharge flow rate, as well as the actual positions of the refrigeration valve, slurry pump, air valve, back pressure valve, and agitator.

[0106] The current operation record is compared with the actuator position status of the corresponding time period in the operating status page to obtain the position verification result. The specific comparison is as follows: check whether the actual position value of each actuator in the current operation record matches the target position value set in the control segment of the candidate operation chain. If the deviation between the actual position value and the target position value is within the allowable tolerance range, the control segment is determined to have been successfully executed.

[0107] In this embodiment, the position tolerance of the refrigeration valve and back pressure valve is ±2% opening, the frequency tolerance of the slurry pump is ±0.5 Hz, the position tolerance of the air valve is ±3% opening, and the speed tolerance of the agitator is ±5 revolutions per minute. If the deviation between the actual position value and the target position value of the actuator exceeds the allowable tolerance range, the central control computer further queries the health status indicators of the actuator, including the cumulative running time, the running cycle after the last maintenance, and the current fault alarm status, to determine whether the deviation is caused by health problems such as aging, jamming, or sensor drift of the actuator, and writes the deviation record and health assessment results into the position verification result.

[0108] Based on the position verification results, completed chain segments are confirmed, and a chain segment execution result is generated. If the position verification result indicates that the current control segment has been successfully executed, then the control segment is marked as a completed chain segment. If the position verification result indicates that the current control segment has not been successfully executed, then the control segment and all subsequent unexecuted control segments are marked as incomplete chain segments, and all control segments preceding the completed chain segments are confirmed as completed chain segments. The chain segment execution result includes a list of completed chain segments, a list of incomplete chain segments, and a note explaining why they were incomplete.

[0109] Based on the chain segment execution results, completed and incomplete chain segments are extracted to obtain chain segment separation results. The completed chain segments in the chain segment separation results are then matched with the current operation record to obtain the current position record. The current position record reflects the actual position of each actuator after the candidate operation chain is partially executed. Specific content includes the actual opening degree of the refrigeration valve, the actual operating frequency of the slurry pump, the actual opening degree of the air valve, the actual opening degree of the back pressure valve, and the actual rotational speed of the agitator.

[0110] The central control computer corrects the time period position record in the operating condition page according to the current position record. The correction method is as follows: update the average position value of the actuator corresponding to the current time period in the operating condition page to the actual position value in the current position record. At the same time, append all the running record frames collected during the execution of this control segment to the operating condition page, update the frame sequence number and timestamp information, and the corrected operating condition page is the updated operating condition page.

[0111] Simultaneously, the central control computer extracts the remaining control actions from incomplete chain segments, forming a set of successor control segments. The extraction method involves: extracting the list of incomplete chain segments from the chain segment separation results, retaining the names of the actuators, starting positions, target positions, and direction of action within each incomplete chain segment, and arranging them in their original order to form the set of successor control segments. If the current control segment fails to reach its target position due to actuator health issues or external disturbances, the first control segment in the successor control segment set remains the incomplete control segment itself, with its target position unchanged. If the current control segment has been successfully executed, the successor control segment set represents the sequence of subsequent control segments in the original candidate operation chain that have not yet been issued.

[0112] The execution order of the remaining control actions is rearranged based on the updated operating status page and the set of successive control segments to form a successive operation chain. The specific rearrangement method is as follows: based on the current position records of each actuator in the updated operating status page, reachability verification is performed on each control segment in the successive control segment set. If the target position of a control segment is still within the migration direction allowed by the reachable path list under the updated current equipment position, the control segment is retained; if the target position is no longer reachable due to a change in the current actual position, the target position of the control segment is replaced with the adjacent position in the reachable path list that is closest to the current actual position and meets the action direction requirements. After the reachability verification, the control segments in the successive control segment set are arranged in their original order to form a successive operation chain. The successive operation chain is the new operation sequence after dynamic correction during execution, maintaining the results of completed actions unchanged, and only continuing the unexecuted parts locally based on actual operational feedback.

[0113] The rule generation module determines the stable production segment that returns to the target production zone based on the updated work status page and the continuous operation chain, extracts the corresponding parameter migration relationship, and forms a production rule table.

[0114] Extract the continuous running segments corresponding to the successive operation chain from the updated operating status page to obtain a segment set; based on the start time and execution duration of each control segment in the successive operation chain, determine the time interval covered by this adjustment operation on the time axis of the updated operating status page, and extract all running record frames within the time interval to form a segment set.

[0115] In this embodiment, the starting point of the time interval is the time when the instruction of the first control segment in the continuous operation chain is issued, and the ending point of the time interval is the time after the last control segment is completed and then extended by one sampling window duration, so as to fully cover the entire process response caused by the adjustment action.

[0116] The stable production segment is obtained by checking the continuity of the outlet status after returning to the target production zone according to the segment set. The outlet status returning to the target production zone means that the three core parameters of the freezer outlet temperature, discharge flow rate and agitator current simultaneously enter their respective target production zone ranges.

[0117] In this embodiment, the target production range for the outlet temperature of the freezer cylinder is -4.5°C to -5.5°C, the target production range for the discharge flow rate is 180 liters per hour to 220 liters per hour, and the target production range for the agitator current is 28 amperes to 32 amperes.

[0118] The specific method for checking the continuity is as follows: On the timeline of the segment set, starting from the sampling moment when the exit state first enters the target production zone, check frame by frame whether the three core parameters are continuously kept within the target production zone range. When the three core parameters do not exceed the boundary of the target production zone within 60 consecutive sampling moments, and there is no abrupt change in the parameter change between adjacent frames, the continuous interval is determined to be a stable production segment.

[0119] The starting point of the stable production segment is the sampling moment when the three core parameters first enter the target production zone simultaneously, and the ending point of the stable production segment is the last sampling moment of the segment set, or the previous sampling moment when the parameters deviate from the target production zone again.

[0120] If there is no interval in the fragment set that meets the above continuous maintenance conditions, then this adjustment operation will not generate stable production fragments and will not form a new production rule table record.

[0121] The rule record is obtained by extracting the parameter migration order, execution duration, filling cycle time correspondence, and discharge holding time from the stable production segment.

[0122] The parameter migration sequence refers to the actual sequence of actions performed by each actuator from the starting point of the segment set to the starting point of the stable production segment. This sequence of actions is directly extracted from the continuous operation chain and includes the names of the actuators that actually performed the adjustment actions in the refrigeration valve, slurry pump, air valve, back pressure valve, and agitator, as well as the order of their actions.

[0123] Execution time refers to the length of time it takes for each actuator action in the parameter migration sequence to reach the target value from the issuance of the instruction to the actual position. This time is determined by combining the expected execution time of the control segment in the continuous operation chain with the actual position status in the position verification result.

[0124] The filling cycle time correspondence refers to the matching between the average discharge flow rate during the stable production segment and the working cycle time of the downstream filling equipment. Specifically, it records the arithmetic mean of the discharge flow rate within the stable production segment and the corresponding cycle time gear of the filling equipment.

[0125] The discharge holding time refers to the total number of sampling moments that a stable production segment lasts from the start to the end, i.e., the number of operation record frames contained in the stable production segment.

[0126] In this embodiment, the recording format for the parameter migration order is a sequence of actuator names arranged according to the order in which the actions occur, such as refrigeration valve → slurry pump → back pressure valve.

[0127] The extracted parameters—migration order, execution duration, filling cycle time correspondence, and discharge holding time—are combined into a single rule record. Simultaneously, the current batch's recipe identifier and frozen section category are appended to the rule record. The recipe identifier includes the ice cream product code and the syrup recipe version number. The frozen section category is the frozen section category of the stable production segment in the frozen state table, i.e., the stable production segment.

[0128] The rules are categorized and organized according to their records to form a production rule table. The categorization is based on the formula identifier and the freezing section category, with rule records sharing the same formula identifier and freezing section category grouped together. Within each rule group, the rule records are arranged in descending order of discharge holding time. The production rule table is stored in the process database of the central control computer, and each rule record is associated with a corresponding updated operating condition page segment index for easy retrieval and retrieval later.

[0129] When the same operating conditions with the same formula, cycle time, start-up stage, or switching stage occur again, the central control computer will first retrieve the parameter migration sequence and successive operation chain structure associated with the corresponding stable production segment from the production rule table, and use it directly as a reference for the initial adjustment sequence, instead of re-executing the entire process of operating condition page construction and sequence optimization from scratch. This will transform the result of this sequence parameter adjustment into the immediate starting point for the next similar operating condition.

[0130] In summary, this invention forms a working condition page by corresponding the time-series process parameters of the continuous freezing section with the current position of each actuator according to the operating stage. Based on this, it extracts freezing state characteristics, forms a freezing state table and a list of reachable paths, thus limiting the adjustment process to the range of reachable paths consistent with the current freezing section category, actuator migration direction, and historical successful action sequence. During sequence optimization, the adjustment parameters are progressively expanded according to the actuator sequence, and a next sampling window prediction segment is immediately established after each level of parameter is added. Forward checks are performed on the outlet state change direction, continuous material discharge maintenance, and filling cycle time maintenance. Migration sequences that are far from the target production zone, disrupt continuous material discharge, or do not meet the actuator position migration relationship are continuously deleted. This transforms sequence convergence from a unified screening after full expansion to an expansion-based process. The system employs a step-by-step contraction process to reduce the accumulation of invalid searches under multi-condition migration and improve the consistency between candidate operation chains and the executable states of field equipment. Simultaneously, during the execution of candidate operation chains, the system verifies the running position at the end of each sampling window and reorders incomplete chain segments, ensuring that completed actions are retained and incomplete parts can be continued locally by updating the condition page. Furthermore, by organizing the stable production segments that successfully return to the target production zone, the parameter migration order, execution duration, and the corresponding filling cycle time into a production rule table, the system forms a closed-loop self-optimization mechanism that includes current condition identification, restricted sequence generation, step-by-step forward convergence, execution feedback correction, and stable production rule rewriting. This makes the system more suitable for multi-condition migration control scenarios where the establishment segment, stable production segment, switching segment, and disturbance segment alternate during the continuous freezing of ice cream.

[0131] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A self-optimizing control system for ice cream production parameters, characterized by: include, The working condition construction module collects the process operation parameters of the continuously frozen section, divides the time period according to different operation stages, and associates the process operation parameters of each time period with the current position of each actuator to form a working condition page; The status classification module extracts the frozen status features from the working status page, completes the classification of frozen segments and matches historical operation fragments, and forms a frozen status table and a list of reachable paths. The sequence optimization module uses the working condition migration sequence optimization algorithm to expand the adjustment parameters corresponding to the frozen state table and reachable path list step by step, and uses the sequence forward convergence pruning algorithm to screen out sequences that deviate from the target production zone and disrupt continuous material output, forming candidate operation chains. The continuation execution module issues control actions segment by segment according to the candidate operation chain and checks the running position at the end of each sampling window. It performs continuation rearrangement for incomplete chain segments to form an updated working condition page and continuation operation chain. The rule generation module determines the stable production segment that returns to the target production zone based on the updated work status page and the continuous operation chain, extracts the corresponding parameter migration relationship, and forms a production rule table.

2. The ice cream production parameter self-optimizing control system of claim 1, wherein: The formation of the operating condition page specifically includes: Synchronously read the slurry feed flow rate, slurry feed temperature, pressure before and after the freezing cylinder, freezing cylinder outlet temperature, agitator current, agitator speed, air flow rate, back pressure valve position, refrigeration valve position, and discharge flow rate to obtain a set of timing parameters; The timing parameter group is encapsulated according to the unified sampling time to obtain the running record frame; The operation is divided into stages based on the continuous changes of the operation record frames on the time axis, resulting in a stage sequence; The operation record frames in the stage sequence are matched with the current positions of the refrigeration valve, slurry pump, air valve, back pressure valve and agitator for each time period to form an operation status page.

3. The ice cream production parameter self-optimizing control system of claim 2, wherein: The division of the operation into stages, resulting in a stage sequence, specifically includes: The slurry feeding status, freezing cylinder outlet status, filling cycle status, and formula switching status in the operation record frame are continuously identified to obtain the stage discrimination results. Based on the stage discrimination results, the stages of start-up freezing establishment, stable material output, filling cycle change, and before and after formula switching are distinguished to obtain a stage label set; Arrange the stage label set in chronological order to obtain the stage sequence.

4. The ice cream production parameter self-optimizing control system of claim 3, wherein: The formation of the frozen state table and the reachable path list specifically includes: Extract the characteristics of outlet temperature change, freezing cylinder pressure difference, agitator current fluctuation, continuous discharge, and air flow rate and slurry flow rate matching from the operating condition page to obtain the freezing state feature set. Based on the feature set of the frozen state, each time period is classified to obtain the frozen segment category results; The frozen status table is obtained by arranging the time periods in order according to the frozen segment category results. Based on the frozen status table, retrieve historical operation segments with the same frozen segment category and similar actuator location status to obtain a set of historical segments; Extract the action sequence and location migration relationships of the actuators from the historical fragment set to form a list of reachable paths.

5. The self-optimizing control system for ice cream production parameters as described in claim 4, characterized in that: The extraction of the action sequence and positional migration relationship of the actuator specifically includes: The actions of the actuators in the historical fragment set are segmented and identified to obtain a set of action segments; The action sequence results are obtained by statistically analyzing the action segments and the sequential action order of the refrigeration valve, slurry pump, air valve, back pressure valve, and agitator in different freezing stages. Based on the action sequence results, the next position that each actuator can enter from its current position is extracted to obtain the position migration result; Based on the location migration results, the movable gears and adjacent gear relationships of each actuator are organized to obtain a path unit set, and the path unit set is arranged in order to form a list of reachable paths.

6. The self-optimizing control system for ice cream production parameters as described in claim 4, characterized in that: The step-by-step sequential expansion of the adjustment parameters corresponding to the frozen state table and reachable path list using the working condition migration sequence optimization algorithm specifically includes: Based on the freeze status table, determine the export status corresponding to the current freeze segment and the preset target production zone to obtain the optimization target; Extract the actuators and corresponding adjustment parameters of the allowed actions of the current frozen segment from the list of reachable paths to obtain the first-level adjustment parameter set; The primary adjustment parameter set is arranged step-by-step according to the optimization objective to obtain the initial migration sequence set; For each migration sequence in the initial migration sequence set, a next sampling window prediction segment is established to obtain the first-level prediction segment set; According to the first-level prediction segment set, check the direction of change of exit status, the maintenance of continuous material discharge and the maintenance of filling cycle time one by one, delete the migration sequence that is far away from the target production zone and disrupts continuous material discharge, and obtain the first-level retained sequence set; Following the order of cooling valve, slurry pump, air valve, back pressure valve and agitator, subsequent adjustment parameters are added to the first-level retained sequence set step by step to form the current-level migration sequence set, and a next sampling window prediction segment is established for the current-level migration sequence set after each level is added; According to the next sampling window prediction segment established at each level, the direction of change of the outlet status, the maintenance of continuous material discharge, and the maintenance of the filling cycle are checked step by step. Migration sequences that are far away from the target production zone and disrupt continuous material discharge are deleted, and migration sequences that satisfy the position migration relationship of the actuator are retained to obtain an extended migration sequence set. The sequence evaluation results are obtained by statistically analyzing the action length, execution order, and rhythm changes of each migration sequence in the extended migration sequence set. Based on the sequence evaluation results, the migration sequences that satisfy the location migration relationship of the implementing agency are retained to form a set of sequences to be screened.

7. The self-optimization control system for ice cream production parameters as described in claim 6, characterized in that: The process of screening out sequences that deviate from the target production zone and disrupt continuous output specifically includes: For each migration sequence in the sequence set to be screened, a set of prediction segments is established, consisting of the next sampling window prediction segments corresponding to each action step. According to the predicted segment set, the direction of change in outlet status, the maintenance of continuous discharge, and the maintenance of filling cycle time are checked one by one from the previous level action step to the next level action step to obtain the sequence discrimination result; Based on the sequence discrimination results, the migration sequences that are far from the target production zone are deleted, and the migration sequences that cause material discharge interruption, pressure change and air entrainment instability are also deleted, resulting in a converged sequence set; Based on the total motion length, cycle time perturbation degree, and retention status after reaching the target production zone, candidate ranking results are obtained by statistically analyzing the convergent sequence set; Based on the candidate ranking results, the migration sequence with the highest ranking is selected to form a candidate operation chain.

8. The self-optimization control system for ice cream production parameters as described in claim 7, characterized in that: The process of issuing control actions segment by segment according to the candidate operation chain and verifying the running position at the end of each sampling window specifically includes: The candidate operation chain is decomposed into action to obtain a set of sequential control segments, and control actions are sent to the field controller segment by segment according to the set of sequential control segments to obtain the control segments in execution; A sampling window is maintained for the control segment in execution. The current operation record at the end of the sampling window is obtained. The current operation record is compared with the position status of the actuator in the corresponding time period in the working condition page to obtain the position verification result. Based on the location verification results, the completed chain segment is confirmed, and the chain segment execution result is generated.

9. The self-optimizing control system for ice cream production parameters as described in claim 8, characterized in that: The process of performing continuation rearrangement on incomplete chain segments specifically includes: Based on the execution results of the chain segments, complete and incomplete chain segments are extracted to obtain the chain segment separation results. The completed chain segments in the chain segment separation results are then matched with the current running record to obtain the current position record. Correct the time period position record in the working condition page according to the current position record, and at the same time extract the remaining control actions in the incomplete chain segment to form the updated working condition page and the set of the continuation control segment respectively. The execution order of the remaining control actions is rearranged based on the updated operating condition page and the set of successive control segments to form a successive operation chain.

10. The self-optimization control system for ice cream production parameters as described in claim 9, characterized in that: The formation of the production rule table specifically includes: Extract the consecutive running segments corresponding to the continuous operation chain from the updated work status page to obtain a segment set; Based on the segment set, check the continuity of the export status after returning to the target production zone to obtain the stable production segment; Extract the parameter migration sequence, execution duration, filling cycle time correspondence, and discharge holding time from the stable production segment to obtain rule records. Then, classify and organize the rule records to form a production rule table.

Images