A real-time temperature control feedback control method and device for cod intestine gelatinization degree

Abstract

The application discloses a real-time temperature control feedback control method and equipment for the gelatinization degree of cod intestines, and relates to the technical field of food processing automation. The method comprises the following steps: S1, multi-modal data acquisition: acquiring multi-modal data of the cod intestines through a near-infrared spectrometer, an embedded sensor, a temperature sensor and a pressure transmitter; S2, decision execution: constructing a state vector according to the multi-modal data, taking the state vector as the input of a reinforcement learning decision model, outputting a target temperature setting value and a steam regulating valve opening degree, setting a control strategy according to the target temperature setting value and the steam regulating valve opening degree, and regulating and controlling a PID controller; and S3, intelligent recovery: adjusting the steam valve opening degree in advance according to the steam flow, and recovering the steam through condensation according to the temperature of the steam. The method realizes the triple objectives of gelatinization precision control, energy consumption optimization and starch retrogradation inhibition.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of food processing automation, in particular to a real-time temperature control feedback control method and device for cod sausage gelatinization degree. BACKGROUND

[0002] As a popular instant or fast food meat product, the taste (such as Q-bounce, tender, chewy) of cod sausage is a key factor in determining product quality and consumer acceptance. In the production process of cod sausage, cooking / heating is a core process, which is mainly to make the protein in surimi denature and gelatinize, and make the starch and other auxiliary materials added to gelatinize. The gelatinization degree of starch has a crucial influence on the texture properties of the final product, especially the gel strength, elasticity and chewiness.

[0003] Starch gelatinization is a complex physicochemical process involving starch granule water absorption, swelling, rupture, and finally forming a viscous and uniform gel network. Incomplete gelatinization will result in astringent taste, sticky teeth, and insufficient elasticity of the product; while over-gelatinization may result in loose structure, water release (water separation), soft texture, and loss of elasticity and toughness.

[0004] The gelatinization process is highly dependent on temperature and time. Different starch types, concentrations, and fish paste matrix compositions have different gelatinization starting temperatures, peak temperatures, and completion temperatures. For multi-component, non-homogeneous composite systems such as cod sausage, achieving sufficient and moderate gelatinization of starch is the key to ensuring product quality stability.

[0005] Chinese invention patent CN110515403A discloses a quinoa meal replacement powder material intelligent cooking control system. Based on the traditional material cooking system, it includes a material information system, an information sensor system, a compensation power calculation system, and a compensation power adjustment system. Through the detection of multiple sensors in the information sensing system, the control system completes the full-automatic material cooking temperature control and cooking completion decision. The temperature adjustment and cooking completion judgment during the cooking process are stable, accurate, rapid, and do not require human intervention, which is beneficial for mass replication, expansion of production scale, and better realization of cooking temperature control. The initial cooking temperature model and the process cooking temperature model are used, and through a large amount of offline training, the model performance is better than that of employees with rich experience. The cooking power adjustment strategy is simple and effective, which is beneficial for real-time and rapid updating of cooking power, and better realization of cooking temperature control. The full-automatic control system can make the material cooking process in a clean and closed environment, which is beneficial for further improving the hygiene of the material.

[0006] However, in the gelatinization process of cod intestines, although high temperature can accelerate gelatinization, the energy waste rate will increase due to steam loss caused by excessively high temperature. Although long-time low-temperature treatment can reduce energy consumption, it will increase the risk of starch retrogradation, thereby further increasing the difficulty of balancing energy consumption and gelatinization stability. SUMMARY

[0007] The purpose of the present application is to provide a real-time temperature control feedback control method and device for the gelatinization degree of cod intestines to solve the problems raised in the background art.

[0008] To achieve the above-mentioned purpose, the present application provides the following technical solution: a real-time temperature control feedback control method for the gelatinization degree of cod intestines, comprising:

[0009] S1: Multi-modal data acquisition: acquiring multi-modal data of cod intestines through a near-infrared spectrometer, an embedded sensor, a temperature sensor and a pressure transmitter;

[0010] S2: Decision execution: constructing a state vector according to the multi-modal data, and taking the state vector as the input of a reinforcement learning decision model, outputting a target temperature setting value and a steam regulating valve opening degree, and simultaneously setting a control strategy according to the target temperature setting value and the steam regulating valve opening degree, and regulating a PID controller, comprising:

[0011] S2.1: Determining the control strategy: combining the target temperature setting value and the steam regulating valve opening degree with the real-time gelatinization degree, and determining the corresponding control strategy according to different gelatinization stages;

[0012] S2.2: Synergistic regulation: dynamically adjusting the electromagnetic heating power according to the corresponding final real-time instruction of the control strategy, and starting an ultrasonic generator according to the size of the real-time gelatinization degree and the loss factor;

[0013] S3: Intelligent recycling: adjusting the steam valve opening degree in advance according to the steam flow, and conducting condensation recycling according to the size of the steam temperature.

[0014] Further, the multi-modal data of cod intestines is acquired, comprising:

[0015] S1.1: Gelatinization degree detection: setting the absorbance ratio reference value through the prepared standard sample group, and determining the real-time gelatinization degree size according to the real-time absorbance acquired by the near-infrared spectrometer at different wavelengths;

[0016] S1.2: Dielectric property monitoring: determining the loss factor reference value according to the prepared standard samples of the un-gelatinized group, the completely gelatinized group and the retrogradation group, and determining the real-time retrogradation risk index according to the actual loss factor acquired by the embedded sensor;

[0017] S1.3: Environmental parameter acquisition: Real-time temperature value is obtained through temperature sensor, and real-time steam pressure is determined through current signal from pressure transmitter.

[0018] Furthermore, the real-time regeneration risk index is compared with the regeneration threshold and the complete regeneration threshold, and the state of the cod intestine is determined based on the comparison results, including:

[0019] S1.2.1: Determine the retrogradation risk index of the samples: Take the average loss factor of the standard samples of the ungelatinized group and the fully gelatinized group as the loss factor benchmark value, and determine the real-time retrogradation risk index of the standard samples of the ungelatinized group, the fully gelatinized group and the retrogradation group based on the loss factor benchmark value.

[0020] S1.2.2: Determine the critical threshold: Based on the real-time retrograde risk index of the standard samples of the ungelatinized group, the fully gelatinized group, and the retrograde group, construct the consumption factor-time data curve, and determine the retrograde critical value and the complete retrograde threshold based on the curve slope, DSC enthalpy value, and microstructure distribution of the consumption factor-time data curve.

[0021] S1.2.3: State Judgment: Based on the regeneration critical value and complete regeneration threshold, the state of the cod intestine is determined, specifically as follows:

[0022] When the loss factor is greater than the warning threshold, the cod intestine is in normal condition; when the loss factor is not greater than the warning threshold, a warning signal is triggered, and the loss factor is compared with the intervention threshold. Based on the comparison result, the monitoring status of the cod intestine is determined, specifically as follows:

[0023] When the loss factor exceeds the intervention threshold, monitoring of the cod intestines continues; when the loss factor does not exceed the intervention threshold, forced intervention is applied to the cod intestines. Simultaneously, both the loss factor and the DSC enthalpy value are compared with the complete regeneration threshold. Based on the results, the regeneration status of the cod intestines is determined, specifically as follows:

[0024] When the loss factor is less than the loss factor in the complete recovery threshold and the DSC enthalpy is greater than the DSC enthalpy in the complete recovery threshold, the cod intestine is in a complete recovery state; otherwise, the cod intestine continues to be monitored.

[0025] Furthermore, the corresponding control strategies are determined, including:

[0026] S2.1.1: Determine the adjustment command: Use the state vector as the input to the reinforcement learning decision model, and output the obtained target temperature setpoint and steam regulating valve opening according to the objective function of the reinforcement learning decision model;

[0027] S2.1.2: Determine Real-Time Command: Based on the real-time gelatinization degree, divide the gelatinization process of the cod sausage into sections, and determine the final real-time command based on the division results, the target temperature setpoint, and the steam regulating valve opening, specifically as follows:

[0028] During the rapid heating phase, the set upper temperature threshold and lower steam regulating valve opening limit are compared with the target temperature setpoint and steam regulating valve opening to determine the final real-time command.

[0029] During the isothermal gelatinization stage: Based on the temperature difference between the target temperature setpoint and the real-time temperature, the opening degree of the steam regulating valve is set as follows:

[0030] ;

[0031] in: The steam valve opening is output by the PID controller. Based on the opening degree, This is the temperature deviation gain coefficient. Set the target temperature value. This is the real-time temperature value;

[0032] When in the regeneration inhibition stage, the temperature setting and the steam regulating valve opening are fixed.

[0033] Furthermore, the real-time gelatinization degree is compared with a preset gelatinization degree threshold range, and based on the comparison result, the gelatinization process of the cod sausage is divided into sections, specifically:

[0034] When the real-time gelatinization degree is less than the lower limit of the preset gelatinization degree threshold range, it is in the rapid heating stage; when the real-time gelatinization degree is within the preset gelatinization degree threshold range, it is in the isothermal gelatinization stage; when the real-time gelatinization degree is greater than the upper limit of the preset gelatinization degree threshold range, it is in the reversion inhibition stage.

[0035] Furthermore, the ultrasonic generator is activated, including:

[0036] S2.2.1: Temperature control: Based on the temperature deviation between the target temperature setpoint in the final real-time command and the real-time temperature value, determine the corresponding adjustment stage and proportional gain coefficient, and based on the adjustment stage and proportional gain coefficient, determine the power adjustment amount;

[0037] S2.2.2: Anti-retrogression intervention: When the gelatinization process of cod intestines is in the retrogression inhibition stage, the sound intensity of the ultrasonic generator is set according to the comparison between the loss factor and the preset loss threshold range, specifically:

[0038] When the loss factor is less than the lower limit of the preset loss threshold range, the sound intensity of the ultrasonic generator is set to 3 W / cm. 2 When the loss factor is within the preset loss threshold range, the sound intensity of the ultrasonic generator is set to 5 W / cm. 2 When the loss factor is greater than the upper limit of the preset loss threshold range, the sound intensity of the ultrasonic generator is set to 8 W / cm. 2 .

[0039] Furthermore, the temperature deviation is compared with a preset deviation threshold range, and based on the comparison result, the corresponding adjustment stage and proportional gain coefficient are determined, specifically as follows:

[0040] When the temperature deviation is less than the lower limit of the preset deviation threshold range, it is in the fine-tuning stage, and the proportional gain coefficient is set to 1.5; when the temperature deviation is within the preset deviation threshold range, it is in the steady-state adjustment stage, and the proportional gain coefficient is set to 1.2; when the temperature deviation is greater than the upper limit of the preset deviation threshold range, it is in the rapid cooling stage, and the proportional gain coefficient is set to 0.8.

[0041] Furthermore, based on the aforementioned adjustment stage and proportional gain coefficient, the power adjustment amount is determined, specifically as follows:

[0042] When in the fine-tuning stage, the corresponding power adjustment amount is as follows:

[0043] ;

[0044] in: For power regulation, This is the proportional gain coefficient. For temperature deviation, Reference power;

[0045] When in the steady-state adjustment phase, the corresponding power adjustment is as follows:

[0046] ;

[0047] in: For power regulation, This is the proportional gain coefficient. Reference power;

[0048] During the rapid cooling phase, the corresponding power adjustment is as follows:

[0049] ;

[0050] in: For power regulation, This is the proportional gain coefficient. As the reference power, This refers to temperature deviation.

[0051] Furthermore, condensation recovery is carried out, including:

[0052] S3.1: Predictive control: The steam flow rate over a continuous period of historical data is used as the input to the LSTM prediction model, and the predicted steam flow rate is output. Based on the predicted steam flow rate and the baseline steam flow rate, the opening degree of the heat exchange valve is determined.

[0053] S3.2: Graded Recovery: The real-time steam temperature is compared with a preset temperature threshold range to classify the steam type into low-temperature waste gas, medium-temperature steam, and high-temperature steam. Based on the steam type, a recovery strategy is determined, specifically:

[0054] When the steam type is low-temperature exhaust gas, the exhaust gas inlet pipe is tilted to limit the raw material flow rate. At the same time, the induced draft fan is started, and auxiliary electric heating is performed according to the preheating temperature of the raw material.

[0055] When the steam type is medium-temperature steam, the steam is pressurized to the upper limit of the preset temperature threshold range and then transmitted to the condenser to heat the clean water at the same time.

[0056] When the steam type is high-temperature steam, it is condensed and recovered through a condenser.

[0057] A real-time temperature feedback control device for the gelatinization degree of cod intestines uses any one of the above-mentioned real-time temperature feedback control methods for the gelatinization degree of cod intestines.

[0058] Compared with the prior art, the beneficial effects of the present invention are:

[0059] Firstly, this invention enables timely ultrasonic intervention through real-time monitoring of the dielectric loss factor, reducing the risk of retrogradation. Furthermore, the ultrasonic intensity can be graded and controlled according to the degree of retrogradation, thereby destroying the starch recrystallization network while extending the shelf life of cod sausage. Simultaneously, by predicting steam flow rate using LSTM and performing graded steam recovery, not only is steam loss reduced, but the energy efficiency of the product is also improved.

[0060] Secondly, this invention constructs a state vector through multimodal data fusion and uses the state vector as the input to a reinforcement learning decision model, outputting the target temperature setpoint and steam valve opening, thereby realizing real-time monitoring and dynamic control of the degree of gelatinization.

[0061] Thirdly, in the constant-temperature gelatinization stage, the opening degree of the steam valve is dynamically adjusted according to the size of the temperature deviation, thereby reducing energy consumption and avoiding the energy waste of the traditional fixed opening. Attached Figure Description

[0062] Figure 1 This is a flowchart illustrating the real-time temperature control feedback control method of the present invention;

[0063] Figure 2 This is a comparison chart of the real-time monitoring and control effects of gelatinization degree in this invention;

[0064] Figure 3 This is a comparative analysis chart of steam energy consumption in this invention;

[0065] Figure 4 This is a temperature control stability analysis diagram in the present invention;

[0066] Figure 5 This is a comparison chart of production batch consistency analysis in this invention. Detailed Implementation

[0067] The technical solutions of 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.

[0068] In the gelatinization process of cod sausage, while high temperatures can accelerate gelatinization, excessive steam loss due to high temperatures leads to increased energy waste. Prolonged low-temperature treatment, while reducing energy consumption, increases the risk of starch retrogradation, further complicating the balance between energy efficiency and gelatinization stability. This application constructs a state vector from collected multimodal data and uses this vector as input to a reinforcement learning decision model. The output obtains the target temperature setpoint and steam valve opening, while real-time PID dynamic control is implemented based on the gelatinization stage division. Furthermore, an ultrasonic anti-retrogradation intervention and a steam stage recovery system are combined to achieve the triple objectives of gelatinization precision control, energy consumption optimization, and starch retrogradation inhibition. Example 1

[0069] refer to Figures 1-5 This embodiment provides a real-time temperature control feedback method for the gelatinization degree of cod intestines, which includes the following steps:

[0070] Step S1: Multimodal data acquisition. This involves real-time scanning of the cod intestines using a near-infrared spectrometer to determine the corresponding degree of gelatinization. An embedded sensor is used to measure the dielectric loss factor to assess the risk of starch retrogradation. Temperature sensors and pressure transmitters are used to monitor the material temperature and steam pressure, respectively. Details are as follows:

[0071] Step S1.1: Gelatinization Degree Detection. This involves scanning the cross-section of cod intestines using a near-infrared spectrometer (e.g., a Bruker Matrix-F) emitting light at a wavelength of 1450-1650 nm to obtain the corresponding absorbance of the product. Specifically, a standard sample set is prepared based on samples with known gelatinization degrees. In this embodiment, the gelatinization degree of the raw cod paste and ungelatinized starch mixture sample is 0%. The 0% gelatinization sample is heated in an 85°C water bath for 15 minutes to obtain a 50% gelatinization sample. The 0% gelatinization sample is then treated with steam at 100°C for 30 minutes to obtain a 100% gelatinization sample. In other words, a standard sample set is prepared based on the 0%, 50%, and 100% gelatinization samples.

[0072] Furthermore, the standard sample group was repeatedly scanned using a near-infrared spectrometer, and the corresponding baseline data were determined based on the average value of the data from the multiple repeated scans, as shown in Table 1 below:

[0073] Table 1: Baseline Data Table

[0074]

[0075] In this embodiment, a spectral probe is fixed 20cm above the conveyor belt of the production line to ensure that each cod sausage receives 3-5 effective scans as it passes through. The spectral probe uses a rotating scanning mode to cover more than 90% of the sausage's cross-section. Specifically, based on the spectral probe, the real-time absorbance of each cod sausage at 1450nm and 1650nm wavelengths is acquired. The corresponding real-time degree of gelatinization is determined based on the real-time absorbance at different wavelengths and the established baseline data.

[0076] ;

[0077] in: For real-time gelatinization degree, The real-time absorbance at a wavelength of 1450 nm. The real-time absorbance at a wavelength of 1650 nm. The absorbance ratio of ungelatinized starch serves as a benchmark. The absorbance ratio is used as a benchmark for fully gelatinized starch.

[0078] In the specific implementation process, the real-time absorbance of cod intestine at a wavelength of 1450nm is 0.7 and the real-time absorbance at a wavelength of 1650nm is 0.65. Meanwhile, in this embodiment, the absorbance ratio of ungelatinized starch is 0.93 and the absorbance ratio of fully gelatinized starch is 1.07, so the corresponding real-time gelatinization degree is 68%.

[0079] Step S1.2: Dielectric property monitoring. A parallel plate electrode sensor (such as Keysight 16451B) with an operating frequency of 1MHz and an applied voltage of 5V is installed on both sides of the production line conveyor belt and lightly touched to the surface of the cod sausage to ensure that the electric field can cover the cross-section of the cod sausage.

[0080] Furthermore, after homogenization, the same batch of cod surimi was divided into three groups, each with a weight of no less than 500g. These groups were: an ungelatinized group (stored at 4℃ and retaining its original starch structure); a fully gelatinized group (steamed at 100℃ for 30 minutes and completely gelatinized); and a regenerated group (stored at 4℃ for 24 hours after gelatinization and allowing starch recrystallization). Each cod surimi sample was then filled into a standard casing with a diameter of 20mm to prepare three different standard samples, each 10cm in length to ensure consistent density.

[0081] Furthermore, five measurement points were evenly distributed at both ends and the middle of each sample group for measurement using a parallel plate electrode sensor. Each measurement point was measured 10 times. Specifically, the average value obtained from the repeated measurements at each point was taken as the corresponding measurement value. The specific data is shown in Table 2 below.

[0082] Table 2: Data Distribution Table

[0083]

[0084] Specifically, based on the data distribution of dielectric constant and loss factor in Table 2, a loss factor-time curve is plotted, and based on the curve distribution in the loss factor-time curve, the corresponding regeneration critical value and complete regeneration threshold are determined.

[0085] In this embodiment, each cod sausage on the production line conveyor belt is scanned using a parallel plate electrode sensor to obtain the corresponding loss factor. Based on the actual loss factor and the baseline loss factor values ​​for ungelatinized raw materials and fully gelatinized raw materials, the corresponding real-time reversion risk index is determined, specifically:

[0086] ;

[0087] in: This is a real-time risk index for regeneration. The loss factor at the current moment. The loss factor baseline value is for the fully gelatinized state. This is the baseline value for the loss factor of the ungelatinized raw material.

[0088] Furthermore, the obtained real-time regeneration risk index is compared with the determined regeneration critical value and complete regeneration threshold, and the current state of the cod intestine is determined based on the comparison results.

[0089] Step S1.3: Environmental parameter acquisition, which involves inserting a PT100 platinum resistance sensor (e.g., IEC 60751 Class A) with a temperature measurement range of -50~150℃ into the center of the cod intestine to a depth not less than 1 / 2 of the intestine diameter. In other words, the corresponding temperature is determined based on the resistance value measured by the PT100 platinum resistance sensor. Specifically:

[0090] ;

[0091] in: This is the real-time temperature value. The resistance value of the PT100 sensor. The temperature coefficient of the PT100 sensor.

[0092] In the specific implementation process, the platinum resistance changes linearly with temperature, and the resistance value at 0℃ is 100Ω, while the temperature coefficient of the PT100 sensor is 0.385. Therefore, when the resistance value of the PT100 sensor is 112.5Ω, the corresponding real-time temperature value is 32.5℃.

[0093] Furthermore, a piezoresistive transmitter with a range of 0-0.5 MPa (such as the Siemens SITRANS P320) should be installed 1 meter downstream of the main steam pipeline, avoiding bends / valve disturbance areas. A siphon should also be installed to prevent high-temperature steam from directly impacting the diaphragm. In other words, the real-time steam pressure is determined from the current signal measured by the piezoresistive transmitter.

[0094] ;

[0095] in: For real-time steam pressure, This is the current signal from the piezoresistive transmitter. For the current signal span, This represents the upper limit of the range of the piezoresistive transmitter. This is the lower limit of the range of the piezoresistive transmitter.

[0096] In practical implementation, the range of the piezoresistive transmitter is 0-0.5MPa, meaning the upper limit is 0.5MPa and the lower limit is 0MPa. Simultaneously, the current signal span is 16mA. Therefore, when the current signal of the piezoresistive transmitter is 12mA, the corresponding real-time steam pressure is 0.25MPa.

[0097] Step S2: Decision Execution. This involves constructing a state vector based on the real-time gelatinization degree, loss factor, real-time temperature, and real-time steam pressure obtained in Step S1. This state vector is then used as input to the reinforcement learning decision model, outputting the corresponding target temperature setpoint and steam regulating valve opening. Simultaneously, based on the target temperature setpoint and steam regulating valve opening, a corresponding control strategy is determined, and the PID controller is adjusted according to this strategy. Specifically:

[0098] Step S2.1: Determine the control strategy. This involves combining the target temperature setpoint output by the reinforcement learning decision model and the steam regulating valve opening with the real-time degree of gelatinization to determine the corresponding control strategy based on different gelatinization stages. Specifically:

[0099] Step S2.1.1: Determine the adjustment command. This involves fusing the real-time gelatinization degree, loss factor, real-time temperature value, and real-time steam pressure obtained in step S1 to construct a state vector, which is then used as input to the reinforcement learning decision model. Further, the target temperature setpoint and steam regulating valve opening are obtained by using the objective function set in the reinforcement learning decision model.

[0100] In this embodiment, the objective function set in the reinforcement learning decision model is specifically as follows:

[0101] ;

[0102] in: To comprehensively optimize the objective function, For real-time gelatinization degree, For the weighting of gelatinization deviation, For the target degree of gelatinization, As energy consumption weight, For instantaneous energy consumption, For time.

[0103] In the specific implementation process, when the time is 30 minutes, the real-time gelatinization degree is 0.7, the target gelatinization degree is set to 0.85, the cumulative energy consumption is 80kJ, the gelatinization deviation weight is 0.6, and the energy consumption weight is 0.4. Then the corresponding comprehensive optimization objective function is 32.09.

[0104] Furthermore, when the state vector is [0.6, 0.17, 82℃, 0.28MPa], the target gelatinization degree is 0.8 and the cumulative energy consumption is 40kJ. The corresponding comprehensive optimization objective function is 16.12. At this time, the corresponding output results are the target temperature setpoint and the steam regulating valve opening, which are 85℃ and 70%, respectively.

[0105] Step S2.1.2: Determine the real-time instruction. This involves comparing the real-time gelatinization degree obtained in step S1.1 with a preset gelatinization degree threshold range (which can be specifically set according to actual data, but is not specifically described in this embodiment, e.g., 0.5-0.9), and dividing the gelatinization process of the cod sausage according to the comparison result, specifically as follows:

[0106] When the obtained real-time gelatinization degree is less than the lower limit of the preset gelatinization degree threshold range (0.5), the current stage is rapid heating. When the obtained real-time gelatinization degree is within the preset gelatinization degree threshold range (0.5-0.9), the current stage is isothermal gelatinization. When the obtained real-time gelatinization degree is greater than the upper limit of the preset gelatinization degree threshold range (0.9), the current stage is reversion inhibition.

[0107] Furthermore, during the rapid heating phase, the upper temperature threshold and the lower opening limit of the steam regulating valve are set. Specifically, the upper temperature threshold is set to 95℃, and the lower opening limit of the steam regulating valve is set to 80%. In other words, based on the target temperature setpoint of 85℃ and the steam regulating valve opening of 70% output in step S2.1.1, the final real-time command is: temperature setpoint 85℃, steam regulating valve opening 70%.

[0108] Furthermore, during the isothermal gelatinization stage, the opening degree of the steam regulating valve is set according to the temperature difference between the target temperature setpoint output in step S2.1.1 and the real-time temperature, specifically as follows:

[0109] ;

[0110] in: The steam valve opening is output by the PID controller. Based on the opening degree, This is the temperature deviation gain coefficient. Set the target temperature value. This is the real-time temperature value.

[0111] In the specific implementation process, the temperature setpoint is 85℃, the basic opening is 50%, and the temperature deviation gain coefficient is 10% / T. Therefore, when the real-time temperature is 83℃, the corresponding steam valve opening output by the PID controller is 70%. In other words, the final real-time command is: temperature setpoint 85℃, steam regulating valve opening 70%.

[0112] To elaborate further, during the regeneration suppression phase, the temperature and valve settings are fixed, with the target temperature locked at 70±0.5℃, the valve set to a pulse cycle of opening every 5 minutes and closing every 10 minutes, and the opening degree at the start being 40%. In other words, the final real-time command is: temperature setpoint 70℃, steam regulating valve opening 40%.

[0113] Step S2.2: Coordinated Adjustment. Based on the final real-time command determined in step S2.1.2, the electromagnetic heating power is dynamically adjusted. Simultaneously, based on the real-time degree of gelatinization and loss factor, ultrasonic waves are activated to disrupt starch recrystallization. Specifically:

[0114] Step S2.2.1: Temperature Control. This involves setting a corresponding proportional gain coefficient based on the temperature deviation between the target temperature setpoint in the final real-time command and the real-time temperature value. Specifically, the obtained temperature deviation is compared with a preset deviation threshold range (which can be set according to actual data, but is not specifically described in this embodiment, e.g., 0.5℃-1℃), and based on the comparison result, the corresponding adjustment stage and proportional gain coefficient are determined, as follows:

[0115] When the obtained temperature deviation is less than the lower limit of the preset deviation threshold range (0.5℃), the system is currently in the fine-tuning stage, with a proportional gain coefficient of 1.5. When the obtained temperature deviation is within the preset deviation threshold range of 0.5℃-1℃, the system is currently in the steady-state adjustment stage, with a proportional gain coefficient of 1.2. When the obtained temperature deviation is greater than the upper limit of the preset deviation threshold range (1℃), the system is currently in the rapid cooling stage, with a proportional gain coefficient of 0.8.

[0116] Furthermore, during the fine-tuning phase, the corresponding power adjustment amount is determined based on the proportional gain coefficient of 1.5, specifically as follows:

[0117] ;

[0118] in: For power regulation, This is the proportional gain coefficient. For temperature deviation, This is the reference power.

[0119] In other words, when the temperature deviation is 0.3℃ and the base power is 12kW, the corresponding power adjustment is +3.6kW, and the adjusted power is 15.6kW.

[0120] Furthermore, during the steady-state adjustment phase, the corresponding power adjustment amount is determined based on the proportional gain coefficient of 1.2, specifically as follows:

[0121] ;

[0122] in: For power regulation, This is the proportional gain coefficient. This is the reference power.

[0123] In other words, when the temperature deviation is 0.8℃ and the base power is 12kW, the corresponding power adjustment is +7.2kW, and the adjusted power is 19.2kW.

[0124] Furthermore, during the rapid cooling phase, the corresponding power adjustment is determined based on the proportional gain coefficient of 0.8, specifically as follows:

[0125] ;

[0126] in: For power regulation, This is the proportional gain coefficient. As the reference power, This refers to temperature deviation.

[0127] In other words, when the temperature deviation is 0.6℃ and the base power is 12kW, the corresponding power adjustment is -5.76kW, and the adjusted power is 6.24kW.

[0128] Step S2.2.2: Anti-retrogression intervention. That is, according to the retrogression inhibition stage determined in step S2.1.2, the ultrasonic generator is set, and the gelatinization process of the cod intestine is intervened according to the sound intensity emitted by the ultrasonic generator.

[0129] In this embodiment, the obtained loss factor is compared with a preset loss threshold range (which can be specifically set according to actual data, so it is not specifically described in this embodiment, for example, 0.2-0.22), and the sound intensity of the ultrasonic generator is set according to the comparison result, specifically as follows:

[0130] When the obtained loss factor is less than the lower limit of the preset loss threshold range (0.2), the sound intensity of the ultrasonic generator is set to 3 W / cm². 2 When the obtained loss factor is within the preset loss threshold range of 0.2-0.22, the sound intensity of the ultrasonic generator is set to 5 W / cm². 2 When the obtained loss factor is greater than the upper limit of the preset loss threshold range of 0.22, the sound intensity of the ultrasonic generator is set to 8 W / cm². 2 .

[0131] In other words, when in the retrograde inhibition stage, the intensity of the ultrasonic generator is set according to the obtained loss factor to intervene in the gelatinization process of cod intestines, thereby disrupting starch recrystallization.

[0132] Step S3: Intelligent Recovery. This involves pre-adjusting the steam valve opening based on the set steam flow rate and recovering condensate based on the steam temperature. Details are as follows:

[0133] Step S3.1: Predictive Control. This involves combining the steam flow rates over 10 consecutive minutes from historical data to construct the flow input data. This constructed flow input data is then used as input to an LSTM prediction model, which outputs the predicted steam flow rate for the next 5 minutes.

[0134] Furthermore, based on the predicted steam flow rate output by the LSTM prediction model and the baseline steam flow rate, the corresponding heat exchange valve opening is determined, specifically as follows:

[0135] ;

[0136] in: For the opening degree of the heat exchange valve, This is the predicted steam flow rate. This is the baseline value for steam flow rate.

[0137] In the specific implementation process, the steam flow rate baseline value is set to 1 kg / s. Therefore, when the predicted steam flow rate is greater than the baseline value, steam surplus is predicted, and the heat exchange valve opening is increased to facilitate the recovery of more waste heat. When the predicted steam flow rate is less than the baseline value, steam shortage is predicted, and the heat exchange valve opening is decreased to reduce heat recovery and ensure the normal operation of the main process.

[0138] Specifically, when the predicted steam flow rate is 1.2 kg / s, the corresponding heat exchange valve opening is 54%. When the predicted steam flow rate is 0.9 kg / s, the corresponding heat exchange valve opening is 48%.

[0139] Step S3.2: Graded Recovery. This involves recovering steam at different temperatures based on its temperature. Specifically, the real-time steam temperature is compared with a preset temperature threshold range (which can be set based on actual data; therefore, this embodiment does not elaborate on this, e.g., 95℃-120℃), and the corresponding steam type is determined based on the comparison result. Specifically:

[0140] When the real-time steam temperature is below the lower limit of the preset temperature threshold range (95℃), the corresponding steam stage is low-temperature exhaust gas. When the real-time steam temperature is below the preset temperature threshold range (95℃-120℃), the corresponding steam stage is medium-temperature steam. When the real-time steam temperature is above the upper limit of the preset temperature threshold range (120℃), the corresponding steam stage is high-temperature steam.

[0141] Furthermore, based on the determined steam type, a corresponding recovery strategy is determined. Specifically:

[0142] When the steam type is low-temperature exhaust gas, the exhaust gas inlet pipe is tilted at 5° to prevent condensate backflow. Simultaneously, the flow rate of the cod surimi on the raw material side is controlled at 0.5-1 m / s to prevent blockage. Furthermore, the initial wind speed of the induced draft fan is set to 3 m / s, and the preheating temperature of the raw cod surimi is monitored. When the preheating temperature of the raw cod surimi is lower than a preset preheating threshold (which can be specifically set based on actual data, but is not specifically described in this embodiment, e.g., 45°C), auxiliary electric heating is used for heating; otherwise, the current operating state of the induced draft fan is maintained.

[0143] When the steam type is medium-temperature steam, the steam is pressurized to the upper limit of the preset temperature threshold range of 120℃ and then transmitted to the condenser. It can also be used to heat cleaning water at 60℃-80℃, thereby saving energy of approximately (95℃-30℃)*4.2kJ / kg⋅℃ / 3600=0.076kWh / kg.

[0144] When the steam type is high-temperature steam, the steam is condensed and recovered through a shell-and-tube condenser.

[0145] This embodiment also provides a real-time temperature control feedback control device for the degree of gelatinization of cod intestines, which uses the above-mentioned real-time temperature control feedback control method for the degree of gelatinization of cod intestines.

[0146] refer to Figure 2 ,Depend on Figure 2 It can be seen that, compared with traditional methods, the degree of gelatinization of this scheme is stable at the target value of 85%±1%, with a standard deviation of less than 0.5%. At the same time, through real-time adjustment by reinforcement learning, the response time is less than 10 seconds, which can quickly eliminate bias.

[0147] refer to Figure 3 ,Depend on Figure 3 It can be seen that, compared with the traditional method, the steam recovery rate of this scheme is increased from 40% to 65%, an increase of 25 percentage points, while the steam consumption is reduced from 1.2 tons / hour to 0.9 tons / hour, a reduction of 25%.

[0148] refer to Figure 4 ,Depend onFigure 4 It can be observed that during the isothermal phase (20-80 minutes), the temperature stabilizes within the range of 85±0.5℃, with significantly smaller fluctuations than traditional control methods. This means that real-time adjustment via the PID algorithm effectively suppresses temperature overshoot. During the heating phase (0-20 minutes), the valve rapidly opens to 80%, achieving rapid temperature increase. Simultaneously, during the isothermal phase, the valve opening is dynamically adjusted within the range of 50%±10%, with a response time of less than 30 seconds. During the cooling phase (after 80 minutes), the valve smoothly reduces its opening, preventing a sudden temperature drop. In other words, the PID collaborative reinforcement learning decision model achieves rapid response and avoids overshoot.

[0149] refer to Figure 5 ,Depend on Figure 5 It can be seen that the standard deviation of the traditional method ranges from 0.8 to 1.5, with an average of 1.15. The standard deviation of this solution ranges from 0.1 to 0.3, with an average of 0.18, resulting in an 84% improvement in stability. In other words, by fusing multimodal data and implementing adaptive control, the consistency of the product is significantly improved. Example 2

[0150] This embodiment provides a real-time temperature control feedback method for the gelatinization degree of cod intestines. The specific implementation method is the same as in Embodiment 1, except that the obtained real-time retrograde risk index is compared with the determined retrograde critical value and complete retrograde threshold, and the current state of the cod intestines is determined based on the comparison result. The invention will be illustrated below with specific examples of this embodiment.

[0151] In this embodiment, the current state of the cod intestine is determined by comparing the real-time regeneration risk index with the regeneration threshold and the complete regeneration threshold, as follows:

[0152] Step S1.2.1: Determine the retrogradation risk index of the sample. That is, based on the loss factor data in the data distribution table in Table 2, the average loss factor of the ungelatinized group that is refrigerated at 4℃ and retains the original starch structure is taken as the baseline value of the loss factor of the ungelatinized raw material, and the average loss factor of the fully gelatinized group that is steamed at 100℃ for 30 minutes and is completely gelatinized is taken as the baseline value of the loss factor of the fully gelatinized state.

[0153] Furthermore, based on multiple loss factors corresponding to the sample at different storage times, the average value of these loss factors is taken as the final loss factor for that storage time. Simultaneously, by determining the baseline values ​​of the loss factors for ungelatinized raw materials and the baseline values ​​for the loss factors in the fully gelatinized state, the reversion risk index corresponding to the loss factors at different times is determined, as shown in Table 3 below:

[0154] Table 3: Data Table of Loss Factor and Regeneration Risk Index

[0155]

[0156] Step S1.2.2: Determine the critical threshold. Based on the data in the loss factor-regeneration risk index data table from step S1.2.1, plot the loss factor-time data curve. Simultaneously, based on the curve, obtain the characteristic data table shown in Table 4 below:

[0157] Table 4: Feature Data Table

[0158]

[0159] Based on the curve slopes in the characteristic data table in Table 4, and combining them with the corresponding DSC enthalpy values, it can be seen from the DSC enthalpy values ​​and the corresponding microstructure distribution that the regeneration critical values ​​in this embodiment are 0.16 and 0.165, respectively, i.e., the warning threshold is 0.165 and the intervention threshold is 0.16. Meanwhile, the complete regeneration thresholds in this embodiment are 0.15 and 3 J / g.

[0160] Step S1.2.3: State Judgment. Based on the regeneration critical value and complete regeneration threshold obtained in step S1.2.2, the current state of the cod intestine is judged. Specifically, the current loss factor of the cod intestine is compared with the warning threshold, and based on the comparison result, the current loss factor of the cod intestine is compared with the intervention threshold.

[0161] When the current cod intestine's loss factor exceeds the warning threshold, the current state of the cod intestine is considered normal, and monitoring of the next cod intestine's state continues. Conversely, when the current cod intestine's loss factor does not exceed the warning threshold, a warning signal is triggered. Simultaneously, the current cod intestine's loss factor is compared to the intervention threshold. Based on the comparison result, the current cod intestine's loss factor and DSC enthalpy value are compared to the complete regeneration threshold. Based on the comparison result, the final state of the current cod intestine is determined, specifically:

[0162] If the current loss factor of the cod intestine exceeds the intervention threshold, the loss factor will continue to be monitored. Conversely, if the current loss factor of the cod intestine is not greater than the intervention threshold, forced intervention will be applied to the cod intestine. At the same time, the current loss factor and DSC enthalpy value of the cod intestine will be compared with the complete regeneration threshold. If the loss factor is less than 0.15 and the DSC enthalpy value is greater than 3 J / g, the current cod intestine is considered to have completely regenerated. Otherwise, the loss factor and DSC enthalpy value of the current cod intestine will continue to be monitored.

[0163] 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 embodiments and their equivalents.

Claims

1. A real-time temperature control feedback method for the gelatinization degree of cod intestines, characterized in that, Including: S1: Multimodal Data Acquisition: Multimodal data of cod intestines are acquired using a near-infrared spectrometer, embedded sensors, temperature sensors, and pressure transmitters, including: S1.1: Gelatinization degree detection: By preparing a standard sample group, an absorbance ratio benchmark is set, and the real-time gelatinization degree is determined based on the real-time absorbance obtained by the near-infrared spectrometer at different wavelengths. S1.2: Dielectric property monitoring: Based on the prepared standard samples of ungelatinized group, fully gelatinized group and regenerated group, the baseline value of loss factor is determined, and the real-time regenerated risk index is determined based on the actual loss factor obtained by the embedded sensor. S1.3: Environmental parameter acquisition: Real-time temperature value is obtained through temperature sensor, and real-time steam pressure is determined through current signal from pressure transmitter; S2: Decision Execution: Based on the multimodal data, a state vector is constructed, and the state vector is used as the input to the reinforcement learning decision model. The output is the target temperature setpoint and the steam regulating valve opening. Simultaneously, based on the target temperature setpoint and the steam regulating valve opening, a control strategy is set, and the PID controller is adjusted, including: S2.1: Determine the control strategy: Combine the target temperature setpoint and the steam regulating valve opening with the real-time gelatinization degree, and determine the corresponding control strategy according to different gelatinization stages, including: S2.1.1: Determine the adjustment command: Use the state vector as the input to the reinforcement learning decision model, and output the obtained target temperature setpoint and steam regulating valve opening according to the objective function of the reinforcement learning decision model; S2.1.2: Determine Real-Time Command: Based on the real-time gelatinization degree, divide the gelatinization process of the cod sausage into sections, and determine the final real-time command based on the division results, the target temperature setpoint, and the steam regulating valve opening, specifically as follows: During the rapid heating phase, the set upper temperature threshold and lower steam regulating valve opening limit are compared with the target temperature setpoint and steam regulating valve opening to determine the final real-time command. During the isothermal gelatinization stage: Based on the temperature difference between the target temperature setpoint and the real-time temperature, the opening degree of the steam regulating valve is set as follows: ； in: The steam valve opening is output by the PID controller. Based on the opening degree, This is the temperature deviation gain coefficient. Set the target temperature value. This is the real-time temperature value; When in the regeneration inhibition stage, the temperature setting and the steam regulating valve opening are fixed. S2.2: Coordinated adjustment: Dynamically adjust the electromagnetic heating power according to the final real-time command corresponding to the control strategy, and start the ultrasonic generator according to the real-time degree of gelatinization and loss factor. S3: Intelligent Recovery: Adjusts the steam valve opening in advance according to the steam flow rate, and performs condensation recovery according to the steam temperature.

2. The real-time temperature control feedback method for the gelatinization degree of cod intestines according to claim 1, characterized in that, The real-time regeneration risk index is compared with the regeneration critical value and the complete regeneration threshold, and the state of the cod intestine is determined based on the comparison results, including: S1.2.1: Determine the retrogradation risk index of the samples: Take the average loss factor of the standard samples of the ungelatinized group and the fully gelatinized group as the loss factor benchmark value, and determine the real-time retrogradation risk index of the standard samples of the ungelatinized group, the fully gelatinized group and the retrogradation group based on the loss factor benchmark value. S1.2.2: Determine the critical threshold: Based on the real-time retrograde risk index of the standard samples of the ungelatinized group, the fully gelatinized group, and the retrograde group, construct the consumption factor-time data curve, and determine the retrograde critical value and the complete retrograde threshold based on the curve slope, DSC enthalpy value, and microstructure distribution of the consumption factor-time data curve. S1.2.3: State Judgment: Based on the regeneration critical value and complete regeneration threshold, the state of the cod intestine is determined, specifically as follows: When the loss factor is greater than the warning threshold, the cod intestine is in normal condition; when the loss factor is not greater than the warning threshold, a warning signal is triggered, and the loss factor is compared with the intervention threshold. Based on the comparison result, the monitoring status of the cod intestine is determined, specifically as follows: When the loss factor exceeds the intervention threshold, monitoring of the cod intestines continues; when the loss factor does not exceed the intervention threshold, forced intervention is applied to the cod intestines. Simultaneously, both the loss factor and the DSC enthalpy value are compared with the complete regeneration threshold. Based on the results, the regeneration status of the cod intestines is determined, specifically as follows: When the loss factor is less than the loss factor in the complete recovery threshold and the DSC enthalpy is greater than the DSC enthalpy in the complete recovery threshold, the cod intestine is in a complete recovery state; otherwise, the cod intestine continues to be monitored.

3. The real-time temperature control feedback method for the gelatinization degree of cod intestines according to claim 1, characterized in that, The real-time gelatinization degree is compared with the preset gelatinization degree threshold range, and the gelatinization process of cod sausage is divided according to the comparison result, specifically as follows: When the real-time gelatinization degree is less than the lower limit of the preset gelatinization degree threshold range, it is in the rapid heating stage; when the real-time gelatinization degree is within the preset gelatinization degree threshold range, it is in the isothermal gelatinization stage; when the real-time gelatinization degree is greater than the upper limit of the preset gelatinization degree threshold range, it is in the reversion inhibition stage.

4. The real-time temperature control feedback method for the gelatinization degree of cod intestines according to claim 1, characterized in that, Starting the ultrasonic generator includes: S2.2.1: Temperature control: Based on the temperature deviation between the target temperature setpoint in the final real-time command and the real-time temperature value, determine the corresponding adjustment stage and proportional gain coefficient, and based on the adjustment stage and proportional gain coefficient, determine the power adjustment amount; S2.2.2: Anti-retrogression intervention: When the gelatinization process of cod intestines is in the retrogression inhibition stage, the sound intensity of the ultrasonic generator is set according to the comparison between the loss factor and the preset loss threshold range, specifically: When the loss factor is less than the lower limit of the preset loss threshold range, the sound intensity of the ultrasonic generator is set to 3 W / cm. 2 When the loss factor is within the preset loss threshold range, the sound intensity of the ultrasonic generator is set to 5 W / cm. 2 When the loss factor is greater than the upper limit of the preset loss threshold range, the sound intensity of the ultrasonic generator is set to 8 W / cm. 2 .

5. The real-time temperature control feedback method for the gelatinization degree of cod intestines according to claim 4, characterized in that, The temperature deviation is compared with a preset deviation threshold range, and based on the comparison result, the corresponding adjustment stage and proportional gain coefficient are determined, specifically as follows: When the temperature deviation is less than the lower limit of the preset deviation threshold range, it is in the fine-tuning stage, and the proportional gain coefficient is set to 1.5; when the temperature deviation is within the preset deviation threshold range, it is in the steady-state adjustment stage, and the proportional gain coefficient is set to 1.2; when the temperature deviation is greater than the upper limit of the preset deviation threshold range, it is in the rapid cooling stage, and the proportional gain coefficient is set to 0.

8.

6. The real-time temperature control feedback method for the gelatinization degree of cod intestines according to claim 4, characterized in that, Based on the aforementioned adjustment stage and proportional gain coefficient, the power adjustment amount is determined as follows: When in the fine-tuning stage, the corresponding power adjustment amount is as follows: ； in: For power regulation, This is the proportional gain coefficient. For temperature deviation, Reference power; When in the steady-state adjustment phase, the corresponding power adjustment is as follows: ； in: For power regulation, This is the proportional gain coefficient. Reference power; During the rapid cooling phase, the corresponding power adjustment is as follows: ； in: For power regulation, This is the proportional gain coefficient. As the reference power, This refers to temperature deviation.

7. The real-time temperature control feedback method for the gelatinization degree of cod intestines according to claim 1, characterized in that, Condensation recovery includes: S3.1: Predictive control: The steam flow rate over a continuous period of historical data is used as the input to the LSTM prediction model, and the predicted steam flow rate is output. Based on the predicted steam flow rate and the baseline steam flow rate, the opening degree of the heat exchange valve is determined. S3.2: Graded Recovery: The real-time steam temperature is compared with a preset temperature threshold range to classify the steam type into low-temperature waste gas, medium-temperature steam, and high-temperature steam. Based on the steam type, a recovery strategy is determined, specifically: When the steam type is low-temperature exhaust gas, the exhaust gas inlet pipe is tilted to limit the raw material flow rate. At the same time, the induced draft fan is started, and auxiliary electric heating is performed according to the preheating temperature of the raw material. When the steam type is medium-temperature steam, the steam is pressurized to the upper limit of the preset temperature threshold range and then transmitted to the condenser to heat the clean water at the same time. When the steam type is high-temperature steam, it is condensed and recovered through a condenser.

8. A real-time temperature control feedback device for the gelatinization degree of cod intestines, characterized in that, The method for real-time temperature control feedback of cod intestine gelatinization degree as described in any one of claims 1-7 was used.

Images