Distributed digital collaborative control system for starch deep processing workshop

Abstract

The application relates to the technical field of industrial automation control, and discloses a distributed digital collaborative control system of a starch deep processing workshop, which comprises a rheological state indirect observation module, a time delay addressing matrix construction module, an enthalpy difference and enzymatic execution module and a multi-dimensional joint self-correction module.The rheological state indirect observation module collects operation parameters to calculate an apparent shear resistance characteristic value; the time delay addressing matrix construction module combines a section area parameter to calculate a transmission time, and relates the apparent shear resistance characteristic value to a predicted arrival time to queue; the enthalpy difference and enzymatic execution module calculates an enthalpy difference deviation value, and combines a feedforward gain to calculate a frequency increment when a trigger condition is met; the flash phase change characteristic extraction module collects a pressure signal to extract a phase change characteristic value and records a time stamp; and the multi-dimensional joint self-correction module corrects a section area parameter according to a phase deviation, and corrects a feedforward gain according to a slope difference value.The technical scheme of indirectly calculating the apparent shear resistance and combining fluid transmission time delay to construct an addressing data queue achieves the technical effect of accurately tracking the state of a specific material unit to implement feedforward intervention.

Description

Technical Field

[0001] This invention relates to the field of industrial automation control technology, specifically a distributed digital collaborative control system for a starch deep processing workshop. Background Technology

[0002] Starch deep processing is a fundamental industry for food manufacturing, biopharmaceuticals, and fine chemicals. In the process of converting raw starch into various syrups or modified starches, the slurry preparation and liquefaction processes are crucial. Materials continuously flow through different pipelines and reaction equipment on the production line. To ensure final product yield and reduce overall energy consumption, companies need to rigorously monitor and control the rheological state, heat exchange, and biochemical reaction progress of each stage within the workshop. This necessitates the deployment of a reliable digital control system within the production line.

[0003] Currently, starch deep-processing workshops generally employ distributed control technology based on single-loop feedback. Various sensors are typically deployed at the ends of pipelines and on the main reaction vessels to monitor parameters such as temperature, pressure, and instantaneous flow rate in real time. The control system receives the monitoring signals and compares them with the operating setpoints. The opening of the upstream steam regulating valve or the operating frequency of the material pump is adjusted individually using a classic PID algorithm. This approach has a relatively simple hardware structure and control logic, resulting in lower initial deployment and commissioning costs. Under static operating conditions and stable material properties, the end-feedback mechanism can effectively maintain local process parameters in a basic equilibrium state.

[0004] Starch slurry delivery pipelines are typically long. By the time end-point sensors detect abnormal physicochemical properties, the problematic material has already passed through the core reaction zone. At this point, the system issues adjustment commands to the front-end actuators, altering the state of newly fed material. Abnormal material is difficult to correct. This pure lag phenomenon often leads to disconnected control actions. The liquefaction process involves heat transfer and enzymatic biochemical reactions. Existing systems implement independent closed-loop control for steam heating and enzyme dosage. When a sudden increase in feed viscosity causes poor local heat transfer, the temperature control loop continuously opens the steam valve. Localized high temperatures at the contact surface can easily lead to thermal inactivation of the enzymes. The lack of coordination between independent loops and mutual interference can easily cause system oscillations and energy waste. After prolonged continuous operation of industrial equipment, physical scaling easily forms on the inner walls of pipelines. The pipe diameter and flow cross-sectional area decrease accordingly, altering the actual material flow rate and process transmission time. Pre-set system control parameters gradually deviate from actual physical conditions. The existing control architecture lacks a post-hoc evaluation mechanism for deteriorating operating conditions. Fixed predictive models cannot adapt to equipment wear evolution, and the dynamic adjustment accuracy often declines significantly after long-term operation. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a distributed digital collaborative control system for starch deep processing workshops. This system solves problems such as the disconnection of control actions caused by pure time delay in long-distance pipeline transmission, the interference between heat input and biochemical reactions caused by independent closed-loop circuits, and the physical degradation caused by the inability of the fixed control model to adapt to long-term operation of the equipment.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a distributed digital collaborative control system for a starch deep processing workshop, comprising:

[0007] The rheological indirect observation module is used to collect electrical operating parameters of the slurry mixing tank and calculate the apparent shear resistance characteristic value of the target material unit.

[0008] The delay addressing matrix construction module is used to calculate the transmission time by combining the equivalent effective cross-sectional area parameter, associate the apparent shear resistance characteristic value of the target material unit with the expected arrival time and dequeue it;

[0009] The enthalpy difference and enzyme-catalyzed execution module is used to calculate the enthalpy difference deviation value. When the apparent shear resistance characteristic value or apparent characteristic change increment of the target material unit meets the trigger condition with the enthalpy difference deviation value, the frequency execution increment is calculated in combination with the biochemical mapping feedforward total gain parameter.

[0010] The flash phase change feature extraction module is used to acquire absolute pressure signals, extract flash phase change feature values, and record absolute timestamps;

[0011] The multidimensional joint self-calibration module is used to calculate the time phase deviation based on the absolute timestamp matched to the target material unit, the expected arrival time, and the comprehensive equivalent process delay, so as to correct the equivalent effective cross-sectional area parameter, and to correct the biochemical mapping feedforward total gain parameter based on the slope difference between the characteristic slope lower limit threshold and the flash phase change characteristic value.

[0012] Preferably, the rheological indirect observation module is specifically used for:

[0013] The effective values ​​of the line voltage, line current, and power factor of the mixing motor in the slurry mixing tank are collected as electrical operating parameters of the slurry mixing tank, and the active power is calculated.

[0014] The net active power is obtained by subtracting the reference loss power using the comprehensive power transfer correction factor.

[0015] The apparent shear resistance index is calculated based on the net active power, the operating frequency of the mixing motor in the slurry tank, and the rheological characteristic correction coefficient.

[0016] The apparent shear resistance indexes obtained continuously within a material time slice are statistically processed to obtain the apparent shear resistance characteristic value of the target material unit.

[0017] Preferably, the delay addressing matrix construction module is specifically used for:

[0018] The instantaneous mass flow rate and fluid density of the main feed pipe are collected, and the transient pure delay transmission time is calculated as the transmission time by combining the physical length of the pipe and the equivalent effective cross-sectional area parameter.

[0019] Establish a data queue that operates according to the first-in, first-out (FIFO) rule;

[0020] The estimated arrival time is obtained by adding the generation time of the apparent shear resistance characteristic value of the target material unit to the transmission time.

[0021] The apparent shear resistance characteristic value, the generation time, and the expected arrival time are written into the data queue for storage.

[0022] When the local system time is greater than or equal to the expected arrival time, the apparent shear resistance characteristic value is dequeued and output from the data queue.

[0023] Preferably, the method by which the enthalpy difference and the enzyme-catalyzed execution module calculate the enthalpy difference deviation value is as follows:

[0024] Collect the instantaneous mass flow rate and effective heat release specific enthalpy of steam, and calculate the heat release of steam;

[0025] Collect the instantaneous mass flow rate of the material, the actual temperature rise of the material, and the specific heat capacity of the starch milk, and calculate the increase in sensible heat of the material;

[0026] The difference between the heat released by the steam and the increase in the sensible heat of the material is calculated to obtain the enthalpy difference deviation value.

[0027] Preferably, the triggering conditions specifically include:

[0028] The apparent shear resistance characteristic value of the target material unit exceeds the preset rheological resistance upper limit, or the apparent characteristic change increment generated by adjacent target material units is higher than the preset rheological change judgment threshold.

[0029] Furthermore, the enthalpy difference deviation value is higher than the positive dead zone threshold for multiple consecutive control cycles.

[0030] Preferably, the method by which the enthalpy difference and enzyme-catalyzed execution module calculates the frequency execution increment is as follows:

[0031] When the triggering condition is met, the integral term of the closed-loop control of the steam regulating valve is maintained and the output upper limit is limited;

[0032] The apparent characteristic change increment and the enthalpy difference deviation are normalized respectively to obtain the relative deviation of apparent shear resistance and the relative deviation of enthalpy difference;

[0033] The frequency execution increment of the enzyme metering pump is obtained by weighting the relative deviation of the apparent shear resistance, the relative deviation of the enthalpy difference, the preset weight, and the total gain parameter of the biochemical mapping feedforward.

[0034] Preferably, the flash phase transition feature extraction module is specifically used for:

[0035] The feature acquisition window is opened when the flash feed control valve receives an opening command or when the change in the setpoint of the flash feed control valve reaches the minimum effective change threshold.

[0036] Within the feature acquisition window, the absolute pressure signal of the flash tank is acquired at a set sampling period and then subjected to moving average filtering to obtain the filtered absolute pressure signal.

[0037] The transient first-order differential slope is obtained by performing transient first-order differential operation on the filtered absolute pressure signal using a discrete difference method.

[0038] Preferably, the method by which the flash phase change feature extraction module extracts the flash phase change feature values ​​is as follows:

[0039] From the transient first-order differential slopes, effective slope samples whose directions conform to the effective pressure response direction are selected;

[0040] Extract the effective slope sample with the largest absolute value as the flash phase change feature value, and record the absolute timestamp when the flash phase change feature value appears;

[0041] The flash phase change characteristic value and the absolute timestamp are matched with the target material unit within a preset effective time range.

[0042] Preferably, the method by which the multidimensional joint self-calibration module corrects the equivalent effective cross-sectional area parameter is as follows:

[0043] The theoretical predicted response time is obtained by adding the estimated arrival time matched to the target material unit and the comprehensive equivalent process delay.

[0044] The difference between the absolute timestamp and the theoretically predicted response time is calculated as the time phase deviation;

[0045] When the time phase deviation is negative and exceeds the negative time addressing fault tolerance dead zone, the equivalent effective cross-sectional area parameter is subjected to a downward bounded correction according to the iterative correction step size.

[0046] Preferably, the method by which the multidimensional joint self-calibration module corrects the total gain parameter of the biochemical mapping feedforward is as follows:

[0047] The transient response intensity of the lowest flash pressure corresponding to the target liquefaction state is set as the lower limit threshold of the characteristic slope.

[0048] When the flash phase change characteristic value is lower than the characteristic slope lower limit threshold, and the target material unit matched by the flash phase change characteristic value has undergone enzyme compensation and enzyme-catalyzed evaluation delay, the slope difference between the characteristic slope lower limit threshold and the flash phase change characteristic value is calculated.

[0049] The overall gain parameter of the biochemical mapping feedforward is adjusted upward by multiplying the gain smoothing correction coefficient by the slope difference.

[0050] This invention provides a distributed digital collaborative control system for a starch deep-processing workshop. It has the following beneficial effects:

[0051] 1. This invention employs a technical solution that indirectly calculates apparent shear resistance and combines it with fluid transport delay to construct an addressing data queue, achieving the technical effect of accurately tracking the state of specific material units and implementing feedforward intervention. Compared to conventional control technologies that rely solely on feedback from end-sensors for adjustment, this invention solves the problem of pure lag in long-distance pipeline transport, where the controller cannot respond in advance to specific batches of materials and is prone to control lag.

[0052] 2. This invention employs a technical solution that combines enthalpy difference deviation and rheological characteristics for trigger control. When certain conditions are met, the integral term of the steam regulating valve is locked, and the enzyme addition frequency increment is simultaneously calculated via feedforward, achieving a multi-dimensional synergistic control effect of thermodynamics and biochemical reactions. Compared to existing technologies that implement independent single-loop closed-loop control for the steam valve and enzyme addition pump, this invention overcomes the shortcomings of existing technologies that suffer from system oscillations and severe fluctuations in system energy consumption caused by mutual interference between individual loops under conditions of sudden changes in material properties.

[0053] 3. This invention employs a multi-dimensional joint self-calibration technique by extracting the differential characteristics of flash phase change pressure. The system utilizes the time phase deviation and characteristic slope difference to perform online bounded correction of the time delay prediction parameters and biochemical feedforward gain, achieving the technical effect of adaptive dynamic optimization of control model parameters. Compared to existing technologies that rely on static engineering experience to solidify controller parameters, this invention overcomes the shortcomings of prior model distortion and gradual decrease in feedback accuracy caused by internal pipe scaling or long-term equipment wear. Attached Figure Description

[0054] Figure 1 This is a schematic diagram of the system architecture according to an embodiment of the present invention;

[0055] Figure 2 This is a schematic diagram of the method flow according to an embodiment of the present invention;

[0056] Figure 3 This is a flowchart of the rheological indirect observation logic according to an embodiment of the present invention;

[0057] Figure 4 This is a flowchart illustrating the delay addressing and time axis alignment logic of an embodiment of the present invention;

[0058] Figure 5 This is a diagram illustrating the thermodynamic and biochemical synergistic control logic of an embodiment of the present invention.

[0059] Figure 6 This is a logic diagram for extracting flash phase transition features according to an embodiment of the present invention;

[0060] Figure 7 This is a multi-dimensional joint self-correcting closed-loop logic diagram of an embodiment of the present invention;

[0061] Figure 8 The upstream rheological state observation and cross-node time alignment process of this invention embodiment is shown in (a) the apparent shear resistance characteristic value sequence generated by the slurry adjustment and buffer control node, and (b) the dequeue characteristic value sequence obtained by the jet liquefaction and enzyme addition control node.

[0062] Figure 9 The coordinated control response of the jet liquefaction and enzyme addition control nodes in this embodiment of the invention is shown in (a) the process of enthalpy difference deviation, (b) the process of steam regulating valve opening, and (c) the process of enzyme addition metering pump actual operating frequency.

[0063] Figure 10 The present invention describes the flash phase change feature extraction and multidimensional joint self-calibration process, wherein (a) shows the absolute pressure signal of the flash tank and the absolute value curve of the transient first-order differential slope, and (b) shows the first bounded correction process of the model parameters.

[0064] Among them, 10 is the rheological indirect observation module; 20 is the time delay addressing matrix construction module; 30 is the enthalpy difference and enzyme-catalyzed execution module; 40 is the flash phase change feature extraction module; and 50 is the multidimensional joint self-calibration module. Detailed Implementation

[0065] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] Reference Figure 1The present invention provides a distributed digital collaborative control system for a starch deep processing workshop, comprising: a rheological indirect observation module 10, a time delay addressing matrix construction module 20, an enthalpy difference and enzyme-catalyzed execution module 30, a flash phase change feature extraction module 40, and a multi-dimensional joint self-calibration module 50.

[0067] The starch deep processing workshop adopts a continuous production method, and all process equipment is connected to a distributed control system via industrial Ethernet. According to the equipment function and process location, the on-site control structure is divided into slurry preparation and buffer control nodes, jet liquefaction and enzyme addition control nodes, and flash cooling control nodes.

[0068] Each module runs on a programmable logic controller or DCS control server at its corresponding control node, and exchanges state variables, time information, and control parameters via a data bus. To achieve state association between continuous materials at different control nodes, the control system divides the continuous materials into sequentially arranged virtual material units according to a material time slice period independent of the edge computing cycle, and assigns a unique identifier to each material unit.

[0069] Within a material time slice, the control system averages, medians, or performs preset filtering on the apparent shear resistance indices obtained from multiple edge calculation cycles to obtain the apparent shear resistance characteristic value corresponding to that material unit. The material time slice period is determined based on the axial dispersion of the material in the pipeline, the width of the flash evaporation characteristic acquisition window, and the distinguishability of the expected arrival windows of adjacent material units. The material unit identifier is used throughout the slurry preparation, jet liquefaction, enzyme addition, and flash cooling processes to associate relevant status data, control records, and feedback characteristics.

[0070] The rheological indirect observation module 10 is deployed at the slurry conditioning and buffer control node. This module reads electrical parameters such as current, voltage, operating frequency, and power factor from the frequency converter of the mixing motor in the slurry conditioning tank, and calculates the apparent shear resistance index based on the active power of the mixing shaft and the power criterion model. The apparent shear resistance characteristic value formed after statistical processing of the material time slice is written as a feedforward variable into the shared variable data area and stored together with the material unit identifier, generation time, and data validity flag.

[0071] The delay addressing matrix construction module 20 adopts a cross-node distributed deployment method. The delay calculation and data enqueueing logic runs on the slurry adjustment and buffer control node, while the time comparison and data dequeueing logic runs on the jet liquefaction and enzyme addition control node. The two nodes exchange addressing data via industrial Ethernet and a shared variable data area.

[0072] The time-delay addressing matrix construction module 20 receives the instantaneous mass flow rate and material density output by the mass flow meter of the main feed pipeline, and calculates the transient pure delay transmission time of the material from the slurry adjustment and buffer control node to the jet liquefaction and enzyme addition control node by combining the equivalent effective cross-sectional area parameter in the time-delay addressing model.

[0073] The controller establishes a first-in, first-out (FIFO) data queue within the shared variable data area. Each queue record includes at least the material unit identifier, apparent shear resistance characteristic value, generation time, expected arrival time, and data validity flag. When the local time of the jet liquefaction and enzyme addition control nodes reaches the corresponding expected arrival time, and the flow rate, density, clock synchronization status, and data validity meet preset conditions, the time-delay addressing matrix construction module 20 performs data dequeueing, ensuring that the feedforward variables correspond in time to the material units expected to arrive at the jet liquefaction unit.

[0074] The enthalpy difference and enzyme-catalyzed execution module 30 is deployed at the jet liquefaction and enzyme addition control node. The enthalpy difference and enzyme-catalyzed execution module 30 collects the steam mass flow rate and effective exothermic specific enthalpy entering the ejector, and reads the temperature rise of the material before and after flowing through the ejector. It then calculates the enthalpy difference deviation value of the heat exchange process by combining the material mass flow rate and specific heat capacity.

[0075] When the apparent shear resistance characteristic value of the outgoing material exceeds the upper limit of the rheological resistance, or the increment of the characteristic value between adjacent material units exceeds the rheological mutation threshold, and the enthalpy difference deviation exceeds the positive dead zone for multiple consecutive control cycles, the enthalpy difference and the enzyme-catalyzed execution module 30 trigger coordinated control. After triggering, the integral term of the steam regulating valve closed-loop controller enters a holding state, while limiting the control output of the steam regulating valve from further increasing.

[0076] While limiting steam output, the enthalpy difference and enzymatic execution module 30 calculates the frequency execution increment of the enzyme metering pump based on the increment of the apparent shear resistance characteristic value and the enthalpy difference deviation value, and adjusts the injection amount of heat-resistant α-amylase to change the biochemical degradation intensity of the corresponding material unit. The enzyme addition command, execution time, and pump operating parameters are bound and recorded with the material unit identifier for downstream quality evaluation and gain correction.

[0077] The flash phase change feature extraction module 40 is deployed at the flash cooling control node and acquires the absolute pressure signal inside the flash tank through an absolute pressure transmitter. When the flash feed regulating valve generates a valve position action that meets the triggering conditions, the flash phase change feature extraction module 40 opens the feature acquisition window and performs filtering and first-order differential calculation on the absolute pressure signal within the window.

[0078] The flash phase change feature extraction module 40 extracts the absolute peak value from the differential slope corresponding to the effective pressure response direction and records the absolute timestamp when the peak value appears. This peak value is used to characterize the flash pressure transient response intensity of the corresponding material unit within the preset operating condition range. After completing the validity judgment and material unit matching, the maximum peak value, absolute timestamp, material unit identifier, flash operating condition parameters, and data validity flag are packaged and output.

[0079] The multi-dimensional joint self-calibration module 50 is deployed in the controller or DCS control server corresponding to the global collaborative control layer. Before performing parameter correction, this module performs consistency verification on the rheological state data generated by the slurry adjustment and buffer control nodes, the control execution records generated by the jet liquefaction and enzyme addition control nodes, and the phase change characteristic data generated by the flash cooling control node, based on the material unit identifier. If the data chain is incomplete, the material unit matching fails, or any data is marked as invalid, the parameter correction is prohibited.

[0080] After successful verification, the multidimensional joint self-calibration module 50 calculates the time phase deviation based on the actual response time and the theoretically predicted response time of the flash phase change characteristic, and performs bounded correction on the equivalent effective cross-sectional area parameter in the time delay addressing model based on this deviation. When the flash phase change characteristic value is lower than the lower limit threshold of the characteristic slope, and the corresponding material unit has already undergone enzyme compensation and been evaluated for delay by enzymatic action, the multidimensional joint self-calibration module 50 calculates the slope difference and performs bounded incremental correction on the total gain of the biochemical mapping feedforward.

[0081] Reference Figure 2 This invention provides a distributed digital collaborative control method for a starch deep processing workshop, comprising the following steps:

[0082] S100 collects the electrical operating parameters of the frequency converter of the mixing motor of the slurry tank, performs rheological indirect observation without the need to install a contact physical viscosity sensor, and calculates the apparent shear resistance index of the material.

[0083] S200 calculates the transient pure delay transmission time based on real-time mass flow rate and material density, constructs a dynamic delay addressing matrix, and achieves time axis alignment of apparent shear resistance characteristic values ​​through a first-in-first-out data queue.

[0084] S300 determines the state of heat transfer obstruction based on the apparent shear resistance characteristic value and its change increment and the enthalpy difference deviation value. While maintaining the integral term of the steam regulating valve closed-loop controller and limiting the control output from continuing to increase, it calculates the frequency execution increment of the enzyme metering pump.

[0085] S400: Within the feature acquisition window corresponding to the action of the flash feed regulating valve, the absolute pressure signal of the flash tank is processed by transient differentiation, the peak value of the first-order differential slope is extracted and the corresponding absolute timestamp is recorded.

[0086] S500, after confirming that the flash phase change characteristics, absolute timestamps and upstream control data correspond to the same material unit, calculates the time phase deviation and slope difference, and performs bounded corrections on the equivalent effective cross-sectional area parameter in the dynamic delay addressing matrix and the total gain of the biochemical mapping feedforward, respectively.

[0087] The following explanation, based on the calculation models and execution logic of each control node, further clarifies steps S100 to S500.

[0088] Reference Figure 3 Step S100 involves indirectly observing the rheological state of the material based on the electrical operating parameters output by the frequency converter of the mixing motor in the slurry tank, and calculating the apparent shear resistance index. Specifically, this includes steps S110 to S130.

[0089] S110 collects the electrical operating parameters of the frequency converter of the mixing motor in the slurry mixing tank and calculates the instantaneous active power.

[0090] The rheological indirect observation module 10 is deployed at the slurry conditioning and buffer control node, and reads the operating data of the slurry conditioning tank agitator motor frequency converter according to a preset edge computing cycle. The edge computing cycle can be set from 50 milliseconds to 200 milliseconds depending on the frequency converter communication rate and the controller's computational load. The parameters collected in each computing cycle include the effective value of the stator current. Operating frequency Line voltage and power factor .

[0091] When the rheological resistance of the material in the mixing tank changes, the load torque and input active power of the agitator motor also change accordingly. The rheological indirect observation module 10 calculates the real-time active power of the agitator motor based on the three-phase AC power relationship. :

[0092] ;

[0093] in, This represents the effective value of the line voltage on the output side of the frequency converter. This indicates the corresponding effective value of the line current. This indicates the motor's operating power factor. The electrical parameters in the inverter register are mapped to the programmable logic controller via industrial Ethernet; parameter reading can be accomplished using standard communication function blocks.

[0094] S120, after deducting the reference loss of the stirring drive system, calculate the net active power acting on the material.

[0095] Starch slurry is characterized by high temperature, high viscosity, and a tendency to scale, which can easily lead to adhesion, drift, or hysteresis in contact viscosity sensors during long-term operation. To reduce the impact of these factors on rheological state detection, the rheological indirect observation module 10 indirectly characterizes the material resistance based on changes in the electrical load of the stirring drive system.

[0096] The real-time active power obtained in step S110 This includes motor copper losses, iron losses, mechanical losses, and reduction gear transmission losses. The rheological indirect observation module 10 calculates the net active power acting on the starch slurry based on the comprehensive power transfer correction coefficient and no-load calibration results. :

[0097] ;

[0098] in, This represents the comprehensive power transmission correction factor determined jointly by the stirring motor and the reduction gear mechanism. This represents the reference loss power of the stirring system corresponding to the current operating frequency.

[0099] It can be determined based on motor efficiency parameters, reducer nameplate parameters, or equipment joint calibration results; The calibration was obtained through no-load calibration: when the mixing tank was empty, the stirring motor was controlled to run at one or more preset frequency points, and the steady-state power corresponding to each frequency point was recorded; when multiple calibration frequency points were set, the frequency was determined by looking up a table or interpolation based on the current operating frequency. .

[0100] By deducting the losses of the stirring drive system itself It mainly reflects the power change generated by the blades overcoming the rheological resistance of the material.

[0101] S130, calculate the apparent shear resistance index based on the power criterion model.

[0102] When the structure of the agitator blades, material temperature, solids concentration, and operating frequency are within the calibrated operating range, the net active power of the agitation system changes with the rheological resistance of the material. The rheological indirect observation module 10 calculates the net active power based on this change. With operating frequency The relationship between the apparent shear resistance exponents is used to calculate the apparent shear resistance exponent. :

[0103] ;

[0104] in, Indicates operating frequency cubed; This represents the rheological characteristic correction coefficient determined based on the structure of the mixing tank, the geometric parameters of the agitator blades, and the equipment calibration results.

[0105] It is a relative rheological characteristic obtained through equipment calibration, used to reflect the changing trend of material shear resistance under the same mixing equipment and similar process conditions, and is not used as a direct measurement result of the absolute viscosity of the material.

[0106] Calibration can be performed during the equipment commissioning phase using a standard fluid or a representative starch slurry. During calibration, net active power is collected at a preset operating frequency, and then inversely calculated using known rheological state parameters. The calibration results obtained using clean water can be used as an initial benchmark, and then corrected based on the actual operating data of the starch slurry.

[0107] To avoid division by zero, numerical amplification, or abnormal jumps when the stirring motor stops or starts at low speed, the rheological indirect observation module 10 calculates... First, determine the operating frequency. When... If the frequency falls below the starting frequency threshold, skip the current calculation and mark the data for this cycle as invalid. The starting frequency threshold can be set to 5 Hz.

[0108] Invalid data is not treated as zero values ​​in the differential calculation between adjacent cycles, nor is it involved in downstream coordinated control. The controller can retain the previous valid result for status display. Once the operating frequency recovers to above the starting frequency threshold and remains stable for a preset time, it is reactivated. calculate.

[0109] Within a material time slice, the rheological indirect observation module 10 continuously acquires multiple effective data. The average value, median value, or preset filtering process is performed to form the apparent shear resistance characteristic value corresponding to the material unit.

[0110] The apparent shear resistance characteristic value is used as an upstream feedforward variable in the collaborative control system. It is written into the data bus and shared variable data area, and stored together with the material unit identifier, generation time and validity flag, for time-delay addressing and feedforward control of the jet liquefaction and enzyme addition control nodes.

[0111] Reference Figure 4 Step S200 calculates the material transport delay based on the real-time mass flow rate and material density of the main feed pipe, and uses a first-in-first-out data queue to correlate the apparent shear resistance characteristic value generated by the slurry adjustment and buffer control node with the material unit expected to arrive at the jet liquefaction unit in time. Step S200 includes steps S210 to S230.

[0112] S210: Collect fluid parameters from the main feed pipe and calculate the transient pure delay transmission time.

[0113] When the starch slurry is transported from the slurry conditioning and buffering control node to the jet liquefaction and enzyme addition control node, it needs to pass through the main feed pipe connecting the two control nodes. Due to the influence of pipe length and material conveying speed, there is a transmission lag between the rheological state data generated by the slurry conditioning and buffering control node and the time when the corresponding material arrives at the jet liquefaction unit.

[0114] The delay addressing matrix construction module 20 reads the instantaneous mass flow rate output by the Coriolis mass flow meter via the data bus. With fluid density Based on the relationship between the effective conveying volume of the pipeline and the current volumetric flow rate, the transient pure time delay of the material in the main feed pipeline is calculated. :

[0115] ;

[0116] in, This indicates the total length of the centerline of the pipeline between the outlet of the slurry mixing tank and the inlet of the ejector. This represents the equivalent effective cross-sectional area parameter in the time-delay addressing model. It is used to characterize the effect of the effective conveying volume of a pipeline on the material transport time, and is not directly equivalent to the actual cross-sectional area of ​​the pipeline obtained through geometric measurements.

[0117] When the system is first put into operation, It can be determined based on the construction drawings or on-site measurement results. The pipe design cross-sectional area can be used as the initial value. After the system is put into operation, the multi-dimensional joint self-calibration module 50 adjusts the time phase deviation accordingly. Perform bounded corrections.

[0118] Transient pure delay propagation time Suitable for continuous conveying applications. Calculation. Instantaneous mass flow rate and fluid density The material should be kept within a preset stable range during the expected delivery period of the corresponding material unit.

[0119] To avoid division by zero and abnormal addressing caused by shutdown, cleaning, incomplete pipeline filling, or low flow fluctuations in the flow meter, the time-delay addressing matrix construction module 20 calculates... Pre-determination of instantaneous mass flow rate . When the flow rate drops below the dead zone threshold, the controller stops updating. The arrival timing and dequeueing operations of the relevant data records are suspended, and the current conveying status is marked as invalid. The lower limit dead zone threshold can be set to 3% to 5% of the rated mass flow rate of the feed pump.

[0120] Once the flow rate recovers to above the lower dead zone threshold and remains stable for a preset time, the controller regenerates the material unit data record. Data records that were not addressed before the material stoppage are marked as invalid and will no longer participate in enzyme feedforward control and model parameter correction.

[0121] During material transportation, if or If the rate of change exceeds the preset stability threshold, the controller marks the data record of the corresponding material unit as invalid for time delay prediction. After the conveying state returns to stability, the dynamic time delay addressing record is re-established.

[0122] S220 constructs a dynamic delay addressing matrix and writes the material unit data into the first-in-first-out data queue.

[0123] The delay-addressing matrix construction module 20 establishes a data queue operating according to the first-in, first-out rule within the shared variable data area. The dynamic delay-addressing matrix uses this data queue as its storage structure. Each data record in the queue corresponds to a virtual material unit.

[0124] The capacity of the data queue is determined based on the maximum allowable transmission delay of the process and the material time slice period. The maximum transmission delay can be divided by the material time slice period and rounded up to obtain the basic number of records in the queue, and then a certain amount of capacity is reserved as an operational margin.

[0125] When the number of records in the queue reaches the capacity limit, the controller stops writing new valid data records and outputs a capacity alarm; for data records that have timed out and been marked as invalid, they can be deleted according to preset cleanup rules to free up storage space.

[0126] The delay addressing matrix construction module 20 reads the material unit identifier, apparent shear resistance characteristic value, and generation time generated in step S100, and records the generation time as . The delay addressing matrix construction module 20 will... Compared with the currently calculated Adding these together, we obtain the estimated arrival time of the corresponding material unit at the jet liquefaction unit:

[0127] .

[0128] Subsequently, the time-delay addressing matrix construction module 20 generates material unit data records. Each record includes at least the material unit identifier, the apparent shear resistance characteristic value, and the generation time. Estimated arrival time And data validity indicators. The generated data records are written to the first-in, first-out (FIFO) data queue in chronological order.

[0129] Only one time-delay addressing record is generated for the same material unit. The multiple apparent shear resistance indices obtained within the material time slice have already been statistically processed in step S100 and are no longer repeatedly written into the queue on a per-edge calculation cycle basis.

[0130] S230 performs cross-node time comparison and outputs material unit data at the expected arrival time.

[0131] The slurry preparation and buffering control nodes, as well as the jet liquefaction and enzyme addition control nodes, are clock-synchronized via industrial Ethernet. The control system has a pre-set upper limit for node clock synchronization error. When the clock deviation between two control nodes exceeds this upper limit, data dequeueing and model parameter correction are paused to prevent material unit matching errors caused by clock deviation.

[0132] During each control scan cycle, the delay addressing matrix construction module 20 reads the local system time of the jet liquefaction and enzyme addition control nodes. And will The estimated arrival time recorded at the head of the first-in-first-out (FIFO) data queue. Compare them.

[0133] when Greater than or equal to When the node clock synchronization status, instantaneous mass flow rate, fluid density, material conveying status and data validity flag all meet the preset conditions, the delay addressing matrix construction module 20 determines that the material unit corresponding to the first record of the queue has reached the expected position of the jet liquefaction and enzyme addition control node, and executes data dequeue.

[0134] The outgoing data includes material unit identifier, apparent shear resistance characteristic value, generation time, estimated arrival time, and validity status. The enthalpy difference and enzyme-catalyzed execution module 30 calls the corresponding apparent shear resistance characteristic value based on the material unit identifier and uses this characteristic value for heat transfer obstruction judgment and enzyme addition feedforward calculation.

[0135] when If the expected arrival time is reached, but the flow rate, density, clock synchronization status, or data validity does not meet the dequeue conditions, the delay addressing matrix construction module 20 will not output the record and will mark it as an addressing anomaly. Addressing anomaly records will not participate in subsequent cooperative control and model parameter correction.

[0136] Through the above time addressing process, the apparent shear resistance characteristic value formed upstream can be invoked when the corresponding material unit is expected to arrive at the jet liquefaction unit, thereby compensating for the time lag caused by pipeline transportation between the slurry adjustment node and the jet liquefaction node.

[0137] Reference Figure 5 Step S300 determines whether heat transfer is obstructed in the jet liquefaction process based on the apparent shear resistance characteristic value, the increment of the apparent shear resistance characteristic value change, and the enthalpy difference deviation value, and restricts the output of the steam regulating valve when the triggering condition is met, while simultaneously calculating the frequency execution increment of the enzyme metering pump. Step S300 includes steps S310 to S330.

[0138] S310, calculate the enthalpy difference deviation during the jet liquefaction heat transfer process.

[0139] The enthalpy difference and enzyme-catalyzed execution module 30 is deployed at the jet liquefaction and enzyme addition control node. After the starch slurry enters the ejector, it mixes directly with saturated steam. When the rheological resistance of the material increases, the steam dispersion state and the degree of gas-liquid mixing may change, which may cause the actual temperature rise to deviate from the normal heat exchange state.

[0140] The enthalpy difference and enzyme-catalyzed execution module 30 collects the instantaneous mass flow rate of saturated steam entering the ejector via the field instrument network. and effective exothermic specific enthalpy Simultaneously read the inlet and outlet temperatures of the ejector to calculate the actual temperature rise of the material before and after passing through the ejector. .

[0141] Effective exothermic specific enthalpy This represents the specific enthalpy difference released when a unit mass of saturated steam changes from the ejector inlet state to the condensed state or the mixed state. The enthalpy difference and the enzymatic execution module 30 can retrieve the specific enthalpy corresponding to the steam inlet state and the condensed state or the mixed state from the steam inlet pressure, steam inlet temperature, and the temperature of the mixture after injection, and calculate the effective exothermic specific enthalpy accordingly. .

[0142] Enthalpy difference and enzyme-catalyzed execution module 30 combined with the instantaneous mass flow rate of material in the main feed pipeline Specific heat capacity of starch milk Calculate the enthalpy difference deviation value :

[0143] ;

[0144] in, This represents the instantaneous mass flow rate of saturated steam. This indicates the specific enthalpy of effective heat release from saturated steam. Indicates the instantaneous mass flow rate of the material. This indicates the specific heat capacity of starch milk. This indicates the actual temperature rise of the material before and after it passes through the ejector.

[0145] Starch milk specific heat capacity The value can be determined based on the dry starch concentration, material temperature, and production formula. When the range of process conditions is small, pre-calibrated empirical values ​​can be used; when the range of process conditions is large, the corresponding values ​​can be obtained by looking up tables or interpolation.

[0146] Enthalpy difference deviation This reflects the deviation between the actual heat release of steam and the increase in sensible heat of the material. Enthalpy difference deviation value. It may be affected by factors such as material rheological state, steam state, equipment heat dissipation and instrument error. Therefore, it is not used as a criterion for heat exchange obstruction on its own, but is used in conjunction with the characteristic value of apparent shear resistance corresponding to the same material unit.

[0147] S320 limits the output of the steam regulating valve based on the rheological state and enthalpy difference deviation.

[0148] In conventional temperature closed-loop control, when the ejector outlet temperature is lower than the set temperature, the steam regulating valve closed-loop controller will increase the opening of the steam regulating valve. If the temperature deviation is caused by insufficient mixing due to high rheological resistance, continuously increasing the steam volume may cause local temperature increases and a decrease in enzyme activity.

[0149] The enthalpy difference and enzymatic execution module 30 reads the apparent shear resistance characteristic value output by the delay addressing matrix construction module 20, and performs differential analysis on the apparent shear resistance characteristic values ​​of adjacent material units to obtain the increment of the apparent shear resistance characteristic value change. .

[0150] When the apparent shear resistance characteristic value of the outgoing line The increase in rheological resistance above the preset upper limit, or the incremental change in the apparent shear resistance characteristic value The value exceeds the preset rheological mutation detection threshold, and the enthalpy difference deviates from the value. When the enthalpy difference and enzyme-catalyzed execution module 30 are higher than the positive dead zone threshold for multiple consecutive control cycles, the current material unit is determined to be in a state of suspected high rheological resistance and heat transfer obstruction.

[0151] The enthalpy difference and enzymatic execution module 30 triggers a thermodynamic boundary suppression command, causing the integral term of the steam regulating valve closed-loop controller to enter a holding state. During the holding period of the integral term, the integral state value stops accumulating, while the proportional and derivative terms still participate in the calculation according to the preset control rules.

[0152] The enthalpy difference and the enzyme-catalyzed execution module 30 simultaneously limit the upper limit of the output of the steam regulating valve closed-loop controller. The upper limit of the output can be set to the opening degree of the steam regulating valve when the thermodynamic boundary suppression command is triggered, or it can be set to a pre-calibrated safe opening degree. In this state, the steam regulating valve cannot continue to increase its opening degree, but it can decrease its opening degree according to temperature feedback.

[0153] When the apparent shear resistance characteristic value drops below the rheological resistance relief threshold, and the enthalpy difference deviates from the value... When the temperature falls back to within the dead zone range after several consecutive control cycles, the enthalpy difference and enzyme-catalyzed execution module 30 release the integral term holding state and output upper limit limit. The steam regulating valve closed-loop controller restores normal closed-loop control through output tracking, reducing valve position abrupt changes caused by control state switching.

[0154] When the integral term holding time exceeds the maximum allowable duration, the ejector outlet material temperature is lower than the process safety lower limit, the enzyme metering pump malfunctions, or the steam regulating valve feedback is abnormal, the enthalpy difference and enzyme-catalyzed execution module 30 stops the current coordinated control and outputs the corresponding fault alarm.

[0155] S330, calculates the frequency increment of the enzyme metering pump.

[0156] After triggering the thermodynamic boundary inhibition command, the enthalpy difference and enzymatic execution module 30, while limiting the steam regulating valve from further increasing its output, initiates enzymatic feedforward compensation. Thermoresistant α-amylase can hydrolyze the glycosidic bonds in starch molecules, converting some large starch molecules into short-chain dextrins, thereby reducing the rheological resistance of the material.

[0157] To eliminate the dimensional difference between the increment of the apparent shear resistance characteristic value and the deviation of the enthalpy difference, the enthalpy difference and the enzyme-catalyzed execution module 30 respectively measure the increment of the apparent shear resistance characteristic value. Deviation from enthalpy difference Perform per-unit processing.

[0158] Increment of change in apparent shear resistance eigenvalue Dividing by the corresponding maximum permissible deviation reference value of the rheological parameter yields the dimensionless relative deviation of the apparent shear resistance. Enthalpy difference deviation value Divide by the corresponding maximum permissible deviation benchmark value of enthalpy difference to obtain the dimensionless relative deviation of enthalpy difference. .

[0159] The enthalpy difference and enzyme-catalyzed execution module 30 perform a weighted calculation on the two dimensionless relative deviations to obtain the frequency execution increment of the enzyme metering pump. :

[0160] ;

[0161] in, This represents the total feedforward gain of the biochemical mapping, and its physical unit is consistent with the frequency unit of the enzyme metering pump inverter. The weight representing the relative deviation of apparent shear resistance; This represents the weight of the relative deviation of the enthalpy difference. and The sum is 1, and the specific value can be calibrated according to the type of starch raw material, dry basis concentration and liquefaction process parameters.

[0162] Frequency execution increment The final frequency setpoint is formed by superimposing the base operating frequency of the enzyme metering pump. The base operating frequency is determined based on the base enzyme dosage per unit mass of dry starch under normal operating conditions.

[0163] The enthalpy difference and enzymatic execution module 30 performs amplitude limiting processing before outputting the final frequency setpoint. When the frequency is incremented... When the frequency is less than zero, the frequency will be incremented. The final frequency setpoint is limited to zero; when the final frequency setpoint is higher than the maximum operating frequency allowed by the enzyme metering pump frequency converter, the final frequency setpoint is limited to the maximum operating frequency.

[0164] When the final frequency setpoint is lower than the minimum stable operating frequency allowed by the enzyme metering pump, the enthalpy difference and enzyme-catalyzing execution module 30, according to the preset control mode, makes the enzyme metering pump operate at the minimum stable operating frequency, or stops the enzyme metering pump.

[0165] During the equipment commissioning phase, the actual enzyme solution output flow rate of the enzyme metering pump was measured under different operating frequencies and pipeline back pressures to establish a calibration relationship between operating frequency, pipeline back pressure, and enzyme solution output flow rate. Based on this calibration relationship, the enthalpy difference and enzyme-catalyzed execution module 30 converted the final frequency setpoint into the actual enzyme solution injection volume.

[0166] The enthalpy difference and enzyme-catalyzed execution module 30 records the basic operating frequency, frequency execution increment, final frequency setpoint, actual enzyme solution injection volume, and execution time of the enzyme metering pump, along with the current material unit identifier.

[0167] The control system determines the enzyme-catalyzing evaluation delay based on the pipeline volume, material flow rate, and process residence time between the enzyme addition location and the flash cooling control node. Before the enzyme-catalyzing evaluation delay expires, the feedback data generated by the flash cooling control node is not used to repeatedly correct the total biochemical mapping feedforward gain corresponding to the same material unit.

[0168] Reference Figure 6 In step S400, within the acquisition window corresponding to the action of the flash feed regulating valve, the absolute pressure signal of the flash tank is filtered and transiently differentiated to extract the peak value of the pressure change rate and record the absolute timestamp when the peak value occurs. Step S400 includes steps S410 to S430.

[0169] S410, determine the feature acquisition trigger event, and acquire the absolute pressure signal of the flash tank.

[0170] The flash phase change feature extraction module 40 is deployed at the flash cooling control node. After the starch material completes the isothermal reaction in the laminar flow insulated pipe section, it enters the flash tank, where it is rapidly cooled under negative pressure to reduce the material temperature to the range required by subsequent processes and to reduce the rate of subsequent enzymatic reactions.

[0171] The flash phase change feature extraction module 40 continuously monitors the position setpoint and valve position feedback value of the flash feed control valve. When the flash feed control valve receives an opening command, performs a process step action, or the change in the valve position setpoint relative to the previous steady state reaches the minimum effective change threshold, the flash phase change feature extraction module 40 determines this valve position action as a feature acquisition trigger event and opens the time gating window.

[0172] During continuous and stable production, if no natural valve position movement meets the minimum effective change threshold, the control system can issue a diagnostic valve position change command to the flash feed regulating valve according to a preset cycle, provided that the flash tank pressure, liquid level, and feed flow rate are all within the allowable range. The amplitude and duration of the diagnostic valve position change command are limited, and it does not change the safety boundary under normal production conditions.

[0173] When the triggering time of the diagnostic valve position change command is entirely determined by the predicted arrival time given by the current time-delay addressing model, the flash phase change characteristic values ​​obtained in this acquisition are... It is only used for liquefaction quality evaluation and biochemical mapping feedforward total gain correction, and not for equivalent effective cross-sectional area parameters. Time phase correction. Used to correct the equivalent effective cross-sectional area parameter. The triggering event should originate from natural process actions, fixed-cycle diagnostic actions, or other events with independent time bases.

[0174] The width of the time gating window can be determined based on the flash tank volume, feed flow rate, and pressure response duration. In one embodiment, the time gating window is set to 2 to 5 seconds to cover the main pressure transient process and reduce the impact of background fluctuations during the steady-state discharge phase on the feature extraction results.

[0175] After the time gating window is opened, the flash phase change feature extraction module 40 acquires the absolute pressure inside the flash tank at a millisecond-level sampling period through the absolute pressure transmitter. During the data acquisition process, the material temperature before flash evaporation, feed mass flow rate, initial absolute pressure of the flash tank, liquid level in the flash tank, valve position change range, and vacuum system operating status are recorded simultaneously.

[0176] The data collected is considered valid only when the material temperature before flash evaporation, feed mass flow rate, initial absolute pressure of the flash tank, and changes in the liquid level and valve position of the flash tank are all within the preset effective range, and when there are no fault indicators in the absolute pressure transmitter, flash feed regulating valve, and vacuum system. If any condition is not met, the data collected is considered invalid and will not be used for subsequent model parameter correction.

[0177] To suppress power frequency interference, mechanical vibration, and high-frequency fluctuations in the pressure signal, the flash phase change feature extraction module 40 performs a moving average filter on the raw absolute pressure data. The moving average filter window can contain 5 to 10 consecutive sampling points, and the specific window length is determined based on the absolute pressure transmitter's response speed, sampling period, and the duration of the flash pressure change.

[0178] S420 calculates the transient first-order differential slope of the absolute pressure signal.

[0179] The degree of material liquefaction affects the rheological state, bubble formation, and water vapor escape process during flash evaporation. When the material temperature, feed mass flow rate, initial absolute pressure of the flash tank, and changes in the liquid level and valve position of the flash tank are within comparable ranges before flash evaporation, the transient rate of change of the absolute pressure of the flash tank can be used as an indirect characteristic for evaluating the liquefaction state of the previous process stage.

[0180] Within the time-gated window, the flash phase change feature extraction module 40 extracts the features based on the filtered absolute pressure. Calculate the transient first-order differential slope :

[0181] ;

[0182] in, Indicates time.

[0183] The programmable logic controller (PLC) uses a discrete differential method to perform the actual calculations. The filtered absolute pressure value of the current sampling period is denoted as... The filtered absolute pressure value of the previous sampling period is denoted as The fixed sampling period is denoted as Then the transient first-order differential slope Calculate according to the following formula:

[0184] .

[0185] Discrete difference computation is compatible with the periodic scanning mode of programmable logic controllers, and can obtain the pressure change rate sequence without performing continuous numerical differentiation.

[0186] S430 extracts the peak value of the transient first-order differential slope and records the corresponding absolute timestamp.

[0187] The flash phase change feature extraction module 40 extracts the transient first-order differential slope within the time gating window based on the effective flash pressure response direction determined during the equipment commissioning phase. Directional filtering is performed. Only slope data whose direction of change matches the effective pressure response direction and whose number of consecutive effective sampling points reaches a set value are considered valid slope samples.

[0188] The flash phase change feature extraction module takes the absolute value of 40 effective slope samples and performs continuous comparisons within a time gating window. The effective slope with the largest absolute value is determined as the flash phase change feature value. :

[0189] ;

[0190] Flash phase transition characteristic values This is used to characterize the transient response intensity of flash pressure generated by the corresponding material unit within a preset operating condition range. The flash phase change feature extraction module 40 is used in... When an event occurs, the controller's global system time is read and recorded as an absolute timestamp. .

[0191] The flash phase change feature extraction module 40 extracts the flash phase change feature values ​​based on the material unit's expected arrival window provided by the jet liquefaction and enzyme addition control nodes. and absolute timestamp Match with the material unit identifier. Only if an absolute timestamp is used. A correspondence is established only when the expected arrival window of the only valid material unit is reached.

[0192] When absolute timestamp If a material unit does not fall within any of its expected arrival windows, or falls within two or more overlapping expected arrival windows, the material unit match is marked as invalid, and the flash phase change characteristic value is not displayed. and absolute timestamp It does not participate in the correction of the equivalent effective cross-sectional area parameter and the total gain of the biochemical mapping feedforward.

[0193] During the equipment trial operation phase, multiple groups of material samples with different liquefaction states can be selected. When the material temperature, feed mass flow rate, initial absolute pressure of the flash tank, liquid level in the flash tank, and valve position changes are within the preset range before flash evaporation, the characteristic values ​​of flash phase change are recorded simultaneously. And the offline viscosity, glucose equivalent value, or other liquefaction quality test results of the corresponding material samples. Based on the flash phase change characteristic value The correlation between the results and the liquefaction quality test results was used to determine the lower limit threshold of the characteristic slope. .

[0194] The flash phase change feature extraction module 40 performs a validity check before outputting the data. When the flash phase change feature value... If the noise level is below the baseline noise threshold, no effective flash pressure response is formed within the decision time gating window, and the current characteristic data is discarded. The baseline noise threshold can be determined based on the statistical results of the pressure change rate under normal flash conditions, for example, set to 5% of the normal characteristic peak value.

[0195] When the characteristic value of flash phase change When the noise level exceeds the baseline noise threshold, and the acquisition conditions, equipment status, and material unit matching results are all valid, the flash phase change feature extraction module 40 stops high-speed sampling after the time gating window ends, and records the flash phase change feature values. Absolute timestamp The material unit identifier, flash evaporation condition parameters, and data validity flags are packaged together.

[0196] The packaged data is sent to the multidimensional joint self-calibration module 50 for bounded correction of the equivalent effective cross-sectional area parameter and the total gain of the biochemical mapping feedforward.

[0197] Reference Figure 7 Step S500, after confirming that the flash phase change characteristic value, absolute timestamp, and upstream control data all correspond to the same material unit, calculates the time phase deviation and characteristic slope difference, and performs bounded correction on the equivalent effective cross-sectional area parameter and the total gain parameter of the biochemical mapping feedforward in the dynamic delay addressing matrix. Step S500 includes steps S510 and S520.

[0198] S510 corrects the equivalent effective cross-sectional area parameter in the dynamic delay addressing matrix based on the time phase deviation.

[0199] The multidimensional joint self-calibration module 50 is deployed in the controller or control server of the distributed control system corresponding to the global collaborative control layer of the distributed control system, and has the authority to read the running data of each control node and write the model correction parameters.

[0200] When starch slurry is transported continuously over a long period, material adhesion, scaling, or deposition may occur on the inner wall of the pipeline, causing changes in the actual transport volume and flow resistance. If the dynamic time-delay addressing matrix still uses the initial equivalent effective cross-sectional area parameter to calculate the transmission time, the estimated arrival time of the material unit may gradually deviate from the actual characteristic response time.

[0201] The multidimensional joint self-calibration module 50 reads the flash phase change feature values ​​output by the flash phase change feature extraction module 40. Absolute timestamp And the corresponding material unit identifier, and retrieve the generation time of the same material unit at the slurry preparation and buffer control node based on the material unit identifier. Transient pure time delay propagation time and comprehensive equivalent process delay .

[0202] Comprehensive equivalent process delay This indicates the material's movement from the jet liquefaction and enzyme addition control node to the flash phase change characteristic value. The total time delay between processes includes the residence time of the ejector and connecting pipe sections, the residence time of the laminar flow insulation pipe section, the conveying time from the outlet of the laminar flow insulation pipe section to the flash feed position, the response time of the flash feed regulating valve, the flash pressure response time, and the delay in absolute pressure signal acquisition and filtering. This is the comprehensive equivalent process delay. It can be calculated based on the equipment's structural parameters, or it can be calibrated using trial operation data during the equipment commissioning phase.

[0203] The multidimensional joint self-calibration module 50 calculates the time phase deviation between the actual response time and the theoretically predicted response time of the flash phase transition characteristic according to the following formula. :

[0204] ;

[0205] Among them, absolute timestamp Indicates the characteristic value of flash phase transition The actual characteristic response time at the time of occurrence is not equivalent to the time when the material's geometric front first enters the flash tank. Generation time Transient pure time delay propagation time and comprehensive equivalent process delay Together, they constitute the theoretical prediction of response time.

[0206] When the time phase deviation A negative value indicates that the actual response time of the flash phase change characteristic is earlier than the theoretically predicted response time. If multiple consecutive effective material units show negative time phase deviations, and the node clock synchronization status, mass flow rate signal, fluid density signal, flash feed regulating valve response status, and vacuum system operation status are all normal, the multi-dimensional joint self-correction module 50 determines that the equivalent conveying volume used in the dynamic delay addressing matrix may be too large.

[0207] This judgment is only used to correct the equivalent effective cross-sectional area parameter in the dynamic time delay addressing matrix and is not used as the sole basis for determining whether scaling or pipe diameter reduction has occurred. When a negative time phase deviation is accompanied by an increase in pipe pressure differential, an increase in the outlet pressure of the delivery pump, or other abnormal delivery resistance, the control system may output a warning indicating pipe adhesion, scaling, or abnormal effective delivery volume.

[0208] To reduce the impact of single pressure fluctuations, instrument errors, and local flow disturbances on the correction results, the multidimensional joint self-calibration module 50 performs a moving average processing on the time phase deviations corresponding to three to five consecutive effective material units to obtain the filtered time phase deviations.

[0209] When the filtered time phase deviation exceeds the negative time addressing fault tolerance dead zone, the multi-dimensional joint self-calibration module 50 further checks the data link integrity, node clock synchronization status, mass flow rate and fluid density signal validity, material conveying continuity, flash feed regulating valve operating status, and vacuum system operating status of the same material unit. If any condition is not met, the equivalent effective cross-sectional area parameter is maintained. constant.

[0210] When all verification conditions are met, the multidimensional joint self-calibration module 50 uses an iterative correction step size. Update equivalent effective cross-sectional area parameters :

[0211] ;

[0212] in, This represents the equivalent effective cross-sectional area parameter before correction; This represents the corrected equivalent effective cross-sectional area parameter, and the iterative correction step size. It is a positive value. Due to the negative time phase deviation. Less than zero, updated equivalent effective cross-sectional area parameter The corresponding decrease.

[0213] Iterative correction step size The correction amount used to convert the time phase deviation into an equivalent effective cross-sectional area parameter. Iterative correction step size. The fixed value determined during the equipment commissioning phase can be used, or it can be divided into multiple correction levels based on the absolute value of the time phase deviation.

[0214] The multidimensional joint self-calibration module 50 pairs the corrected equivalent effective cross-sectional area parameters. Apply upper and lower bound constraints. Equivalent effective cross-sectional area parameters. The upper limit can be set to the pipe design cross-sectional area, and the lower limit can be set to 60% of the pipe design cross-sectional area. When the calculation result exceeds the allowable range, the corresponding boundary value shall be used.

[0215] When the equivalent effective cross-sectional area parameter If the pressure remains near the lower limit for several consecutive correction cycles, and the pipeline differential pressure, the outlet pressure of the delivery pump, or other delivery resistance parameters simultaneously reach the abnormal judgment conditions, the control system outputs an equipment maintenance alarm indicating pipeline adhesion, scaling, or abnormal effective delivery volume.

[0216] S520 corrects the total gain parameter of the biochemical mapping feedforward based on the slope difference of the flash phase transition characteristic.

[0217] Different batches of corn feedstock may vary in starch structure, dry basis concentration, and substrate reactivity. The actual activity of thermostable α-amylases is also affected by storage time, storage temperature, and usage conditions. Fixed biochemical mapping feedforward total gain parameters are difficult to adapt to these changes over a long period.

[0218] The multidimensional joint self-calibration module 50 receives the flash phase change feature values ​​output by the flash phase change feature extraction module 40. The control system pre-sets a lower limit threshold for the characteristic slope. Feature slope lower limit threshold Used to characterize the transient response intensity of the lowest flash pressure corresponding to the target liquefaction state.

[0219] During the equipment trial operation phase, when the material temperature before flash evaporation, feed mass flow rate, initial absolute pressure of the flash tank, liquid level in the flash tank, and the variation range of the flash feed regulating valve are within the preset range, multiple sets of material samples that have reached the target glucose equivalent value or other target liquefaction quality indicators are collected, and the flash phase change characteristic values ​​corresponding to each material sample are recorded. Based on the characteristic values ​​of flash phase transition The lower limit threshold of the characteristic slope is determined by the average value, the range of dispersion, the repeatability error, and the allowable process fluctuation. .

[0220] When the characteristic value of flash phase change Below the lower limit threshold of the feature slope At that time, the multidimensional joint self-calibration module 50 confirms, based on the material unit identifier, that the corresponding material unit has already undergone enzyme addition compensation in step S300, and confirms that the enzyme catalytic effect evaluation delay has occurred since the enzyme addition execution time. When both conditions are met, it is determined that the liquefaction effect of the material unit is lower than the preset target.

[0221] When the vacuum pump, flash feed regulating valve, absolute pressure transmitter, temperature sensor, mass flow meter, or enzyme metering pump has a fault indicator, or when the corresponding material unit experiences a material stoppage, process change, data loss, or material unit mismatch during the conveying process, the multidimensional joint self-calibration module 50 will not update the biochemical mapping feedforward total gain parameter. .

[0222] Each material unit is only allowed to trigger a single update of the total gain parameter of the biochemical mapping feedforward, and a maximum allowable change value is set for the amount of a single update to avoid abrupt changes in gain caused by continuous repeated corrections.

[0223] The multidimensional joint self-calibration module 50 calculates the lower limit threshold of the feature slope. With flash phase change characteristic value The slope difference between them is used to update the total gain parameter of the biochemical mapping feedforward according to the following formula. :

[0224] ;

[0225] in, This represents the total gain parameter of the biochemical mapping feedforward before the update. This represents the updated biochemical mapping feedforward total gain parameter, and the gain smoothing correction coefficient. It is a positive value.

[0226] Gain smoothing correction factor The correction factor used to convert the difference in flash pressure change rate into the total gain parameter of the biochemical mapping feedforward. Gain smoothing correction coefficient. It can be determined based on the calibration relationship between the enzyme metering pump frequency, enzyme flow rate, and liquefaction mass.

[0227] Since the total gain parameter of the biochemical mapping feedforward is only present in the flash phase transition eigenvalues Below the lower limit threshold of the feature slope Since the update is performed in real time, the update method in this embodiment is a bounded incremental compensation, which is used to improve the enzyme feedforward strength of subsequent material units.

[0228] The multidimensional joint self-calibration module has 50 pairs of updated biochemical mapping feedforward total gain parameters. Apply upper and lower bound constraints. Biochemical mapping feedforward total gain parameter. The lower limit can be set to zero, while the upper limit is determined based on the maximum allowable operating frequency of the enzyme metering pump, the calibration relationship between the enzyme metering pump frequency and the enzyme flow rate, and the maximum allowable amount of enzyme added by the process.

[0229] When the calculated biochemical mapping feedforward total gain parameter When the value is below the lower limit or above the upper limit, the multidimensional joint self-calibration module 50 adopts the corresponding boundary value and records the current parameter limiting status.

[0230] By correcting for time phase deviation and characteristic slope difference, the equivalent effective cross-sectional area parameter and the total gain parameter of biochemical mapping feedforward in the dynamic delay addressing matrix can be adjusted within the preset effective operating range to adapt to long-term changes in conveying status and raw material reaction characteristics.

[0231] To illustrate the collaborative relationships among the control nodes in a continuous production process, the following section combines... Figures 8 to 10 The following example illustrates the situation where the rheological resistance of a virtual material cell in a continuous material supply is increased. During the example operation, the production line maintains continuous feeding, and the mass flow rate and fluid density are within the stable range allowed by the time-delay addressing model.

[0232] Reference Figure 8 ,like Figure 8 As shown in part (a) of the diagram. Before time t1, the apparent shear resistance characteristic values ​​corresponding to each material unit remained within the baseline range. From time t1 onwards, affected by changes in material properties, the apparent shear resistance index calculated by the rheological indirect observation module 10 based on the electrical operating parameters of the stirring motor gradually increased.

[0233] The rheological indirect observation module 10 statistically processes the continuously obtained apparent shear resistance indices according to material time slices, forming sequentially arranged apparent shear resistance characteristic values ​​of material units, and writes the apparent shear resistance characteristic values, material unit identifiers, generation times, and validity flags into the shared variable data area. The time-delay addressing matrix construction module 20 reads the corresponding data and writes it into the first-in-first-out data queue according to the generation time order.

[0234] like Figure 8 As shown in section (b), the delay addressing matrix construction module 20 calculates the transient pure delay transmission time based on the instantaneous mass flow rate, fluid density, and equivalent effective cross-sectional area parameters of the main feed pipe.

[0235] At time t2, the local system time of the jet liquefaction and enzyme addition control node reaches the expected arrival time of the corresponding material unit. After the node clock synchronization status, mass flow rate, fluid density, and data validity checks all pass, the delay addressing matrix construction module 20 performs data dequeueing.

[0236] Figure 8 Part (a) and Figure 8 In part (b), the change in the apparent shear resistance characteristic value begins at a time difference of one transient pure delay transmission time. Figure 8 The curve in part (b) is formed by connecting the apparent shear resistance characteristic values ​​of the continuous material units exiting sequentially. It is used to represent the output process of the upstream rheological state data at the jet liquefaction and enzyme addition control nodes after time delay compensation.

[0237] Reference Figure 9 ,like Figure 9 As shown in section (a), after the corresponding material unit reaches the jet liquefaction unit, the increased rheological resistance of the material affects the mixing and heat exchange of steam and material. After the response time of the jet and temperature detection stage, the deviation between the enthalpy difference and the enthalpy difference calculated by the enzyme-catalyzed execution module 30 begins to increase and exceeds the positive dead zone threshold.

[0238] like Figure 9As shown in section (b) of the diagram. After the enthalpy difference deviation exceeds the positive dead zone threshold, the control system continues to perform state confirmation for multiple control cycles. When the apparent shear resistance characteristic value exceeds the rheological resistance upper limit, and the enthalpy difference deviation is higher than the positive dead zone threshold for multiple consecutive control cycles, the enthalpy difference and the enzyme-catalyzed execution module 30 trigger a thermodynamic boundary suppression command at time t4.

[0239] After the thermodynamic boundary suppression command is triggered, the integral term of the steam regulating valve closed-loop controller enters a holding state, and the upper limit of the steam regulating valve control output is limited to the valve opening corresponding to the trigger moment. In this example, the temperature feedback after triggering does not require the steam regulating valve to reduce its opening; therefore, the steam regulating valve opening remains at the value corresponding to the trigger moment.

[0240] like Figure 9 As shown in section (c), the enthalpy difference and enzyme-catalyzed execution module 30 calculates the frequency execution increment based on the increment of the apparent shear resistance characteristic value and the enthalpy difference deviation value, and superimposes the frequency execution increment on the base operating frequency to form a new frequency setpoint.

[0241] Due to the acceleration time of the frequency converter of the enzyme metering pump and the mechanical response of the metering pump, the actual operating frequency of the enzyme metering pump gradually increases from time t4 and tends to a new stable operating frequency. After the operating frequency of the enzyme metering pump increases, the injection amount of high-temperature α-amylase increases accordingly.

[0242] Reference Figure 10 ,like Figure 10 As shown in part (a) of the diagram. At time t3, the flash feed regulating valve performs a natural process action or a fixed-cycle diagnostic action with an independent time base, and the flash phase change feature extraction module 40 opens the time gating window.

[0243] The upper curve in the figure represents the absolute pressure signal of the flash tank after filtering, and the lower curve represents the absolute value of the transient first-order differential slope calculated from the filtered absolute pressure signal. The absolute pressure curve first responds after t3, then undergoes a process of gradually increasing and decreasing rate of change. The absolute value of the transient first-order differential slope... The value reaches its maximum at a given time, and this maximum value is denoted as the characteristic value of flash phase transition. The flash phase change feature extraction module 40 synchronously reads the global system time and... The global system time record at the time of occurrence is an absolute timestamp. .

[0244] After the material temperature, feed mass flow rate, initial absolute pressure of the flash tank, liquid level in the flash tank, variation range of the flash feed regulating valve, and operating status of the vacuum system all meet the validity conditions before flash evaporation, the flash phase change feature extraction module 40 will extract the flash phase change feature values. Absolute timestamp The corresponding material unit identifier is sent to the multi-dimensional joint self-calibration module 50.

[0245] like Figure 10 As shown in section (b), the multidimensional joint self-calibration module 50 first verifies the material unit correspondence, data link integrity, and equipment operating status.

[0246] The upper curve represents the change process of the equivalent effective cross-sectional area parameter. (When the absolute timestamp...) When the response time is earlier than the theoretically predicted time and the filtered time phase deviation of the continuous material unit exceeds the negative time addressing fault tolerance dead zone, the multidimensional joint self-calibration module 50 performs a downward bounded correction on the equivalent effective cross-sectional area parameter at parameter correction time t5 according to the iterative correction step size.

[0247] The flash feed control valve action used in this correction of the equivalent effective cross-sectional area parameter has a time base independent of the current time delay prediction result. If the flash feed control valve action is triggered entirely based on the predicted arrival time of the current time delay addressing model, then this data will not be used to correct the equivalent effective cross-sectional area parameter.

[0248] The lower curve represents the change process of the total gain parameter of the biochemical mapping feedforward. The multidimensional joint self-calibration module 50 confirms that the corresponding material unit has undergone enzyme addition compensation, and that an enzyme-catalyzed effect evaluation delay has occurred since the enzyme addition execution time, before determining the flash phase transition characteristic value. With feature slope lower limit threshold Compare them.

[0249] When the characteristic value of flash phase change Below the lower limit threshold of the feature slope When this occurs, it indicates that the liquefaction effect of the corresponding material unit has not reached the preset liquefaction quality target. The multi-dimensional joint self-calibration module 50, at parameter correction time t5, adjusts the parameter based on the characteristic slope lower limit threshold. With flash phase change characteristic value The difference between them is used to perform an upward bounded correction on the total gain parameter of the biochemical mapping feedforward, so as to increase the frequency of the enzyme metering pump under subsequent similar operating conditions.

[0250] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A distributed digital collaborative control system for a starch deep processing workshop, characterized in that, include: The rheological indirect observation module is used to collect electrical operating parameters of the slurry mixing tank and calculate the apparent shear resistance characteristic value of the target material unit. The delay addressing matrix construction module is used to calculate the transmission time by combining the equivalent effective cross-sectional area parameter, associate the apparent shear resistance characteristic value of the target material unit with the expected arrival time and dequeue it; The enthalpy difference and enzyme-catalyzed execution module is used to calculate the enthalpy difference deviation value. When the apparent shear resistance characteristic value or apparent characteristic change increment of the target material unit meets the trigger condition with the enthalpy difference deviation value, the frequency execution increment is calculated in combination with the biochemical mapping feedforward total gain parameter. The flash phase change feature extraction module is used to acquire absolute pressure signals, extract flash phase change feature values, and record absolute timestamps; The multidimensional joint self-calibration module is used to calculate the time phase deviation based on the absolute timestamp matched to the target material unit, the expected arrival time, and the comprehensive equivalent process delay, so as to correct the equivalent effective cross-sectional area parameter, and to correct the biochemical mapping feedforward total gain parameter based on the slope difference between the characteristic slope lower limit threshold and the flash phase change characteristic value.

2. The distributed digital collaborative control system for a starch deep processing workshop according to claim 1, characterized in that, The rheological indirect observation module is specifically used for: The effective values ​​of the line voltage, line current, and power factor of the mixing motor in the slurry mixing tank are collected as electrical operating parameters of the slurry mixing tank, and the active power is calculated. The net active power is obtained by subtracting the reference loss power using the comprehensive power transfer correction factor; The apparent shear resistance index is calculated based on the net active power, the operating frequency of the mixing motor in the slurry tank, and the rheological characteristic correction coefficient. The apparent shear resistance indexes obtained continuously within a material time slice are statistically processed to obtain the apparent shear resistance characteristic value of the target material unit.

3. The distributed digital collaborative control system for a starch deep processing workshop according to claim 1, characterized in that, The time-delay addressing matrix construction module is specifically used for: The instantaneous mass flow rate and fluid density of the main feed pipe are collected, and the transient pure delay transmission time is calculated as the transmission time by combining the physical length of the pipe and the equivalent effective cross-sectional area parameter. Establish a data queue that operates according to the first-in, first-out (FIFO) rule; The estimated arrival time is obtained by adding the generation time of the apparent shear resistance characteristic value of the target material unit to the transmission time. The apparent shear resistance characteristic value, the generation time, and the expected arrival time are written into the data queue for storage. When the local system time is greater than or equal to the expected arrival time, the apparent shear resistance characteristic value is dequeued and output from the data queue.

4. The distributed digital collaborative control system for a starch deep processing workshop according to claim 1, characterized in that, The method by which the enthalpy difference and enzyme-catalyzed execution module calculate the enthalpy difference deviation value is as follows: Collect the instantaneous mass flow rate and effective heat release specific enthalpy of steam, and calculate the heat release of steam; Collect the instantaneous mass flow rate of the material, the actual temperature rise of the material, and the specific heat capacity of the starch milk, and calculate the increase in sensible heat of the material; The difference between the heat released by the steam and the increase in the sensible heat of the material is calculated to obtain the enthalpy difference deviation value.

5. The distributed digital collaborative control system for a starch deep processing workshop according to claim 1, characterized in that, The triggering conditions specifically include: The apparent shear resistance characteristic value of the target material unit exceeds the preset rheological resistance upper limit, or the apparent characteristic change increment generated by adjacent target material units is higher than the preset rheological change judgment threshold. Furthermore, the enthalpy difference deviation value is higher than the positive dead zone threshold for multiple consecutive control cycles.

6. The distributed digital collaborative control system for a starch deep processing workshop according to claim 5, characterized in that, The method by which the enthalpy difference and enzyme-catalyzed execution module calculates the frequency execution increment is as follows: When the triggering condition is met, the integral term of the closed-loop control of the steam regulating valve is maintained and the output upper limit is limited; The apparent characteristic change increment and the enthalpy difference deviation are normalized respectively to obtain the relative deviation of apparent shear resistance and the relative deviation of enthalpy difference; The frequency execution increment of the enzyme metering pump is obtained by weighting the relative deviation of the apparent shear resistance, the relative deviation of the enthalpy difference, the preset weight, and the total gain parameter of the biochemical mapping feedforward.

7. The distributed digital collaborative control system for a starch deep processing workshop according to claim 1, characterized in that, The flash phase transition feature extraction module is specifically used for: The feature acquisition window is opened when the flash feed control valve receives an opening command or when the change in the setpoint of the flash feed control valve reaches the minimum effective change threshold. Within the feature acquisition window, the absolute pressure signal of the flash tank is acquired at a set sampling period and then subjected to moving average filtering to obtain the filtered absolute pressure signal. The transient first-order differential slope is obtained by performing transient first-order differential operation on the filtered absolute pressure signal using a discrete difference method.

8. The distributed digital collaborative control system for a starch deep processing workshop according to claim 7, characterized in that, The method for extracting the flash phase transition feature values ​​by the flash phase transition feature extraction module is as follows: From the transient first-order differential slopes, effective slope samples whose directions conform to the effective pressure response direction are selected; Extract the effective slope sample with the largest absolute value as the flash phase change feature value, and record the absolute timestamp when the flash phase change feature value appears; The flash phase change characteristic value and the absolute timestamp are matched with the target material unit within a preset effective time range.

9. The distributed digital collaborative control system for a starch deep processing workshop according to claim 1, characterized in that, The method by which the multidimensional joint self-calibration module corrects the equivalent effective cross-sectional area parameter is as follows: The theoretical predicted response time is obtained by adding the estimated arrival time matched to the target material unit and the comprehensive equivalent process delay. The difference between the absolute timestamp and the theoretically predicted response time is calculated as the time phase deviation; When the time phase deviation is negative and exceeds the negative time addressing fault tolerance dead zone, the equivalent effective cross-sectional area parameter is subjected to a downward bounded correction according to the iterative correction step size.

10. The distributed digital collaborative control system for a starch deep processing workshop according to claim 1, characterized in that, The method by which the multidimensional joint self-calibration module corrects the total gain parameter of the biochemical mapping feedforward is as follows: The transient response intensity of the lowest flash pressure corresponding to the target liquefaction state is set as the lower limit threshold of the characteristic slope. When the flash phase change characteristic value is lower than the characteristic slope lower limit threshold, and the target material unit matched by the flash phase change characteristic value has undergone enzyme compensation and enzyme-catalyzed evaluation delay, the slope difference between the characteristic slope lower limit threshold and the flash phase change characteristic value is calculated. The overall gain parameter of the biochemical mapping feedforward is adjusted upward by multiplying the gain smoothing correction coefficient by the slope difference.

Images