Fragrant oil-extracted fragrance-preserving dewatering and drying parameter self-adaptive regulation system

Abstract

This invention relates to the field of food processing equipment technology, and discloses an adaptive control system for aroma preservation, dehydration, and drying parameters of a strong-aroma oil press. The system includes a processing and execution unit, a multi-dimensional sensing unit, and a central control and calculation unit. Based on data from the multi-dimensional sensing unit, the central control and calculation unit calculates the average adsorption saturation of the cake (used as an adsorbent) in real time and compares it with a preset threshold. It then executes two branch control strategies: when the saturation is high, the system determines that the adsorbent efficiency has decreased and decides to perform a new cake addition operation to restore dehydration momentum; when the saturation is in the high-efficiency range, the system calculates and adjusts the optimal stirring speed by solving a multi-objective optimization problem targeting dehydration rate, aroma emission intensity, and stirring power. This invention can significantly improve dehydration efficiency while effectively preserving the product aroma, and improve the automation level of the dehydration process and the consistency of product quality.

Description

Technical Field

[0001] This invention relates to the field of food processing equipment technology, specifically to an adaptive control system for parameters of aroma preservation, dehydration, and drying in the extraction of aromatic oils. Background Technology

[0002] Strongly aromatic vegetable oils, such as fragrant rapeseed oil and peanut oil, are highly favored by consumers for their unique flavor. In the production of these oils, removing residual trace amounts of moisture is a crucial step in ensuring their storage stability and final quality. Currently, physical adsorption is commonly used industrially, involving the addition of dried oil cake as an adsorbent. As a byproduct of oil pressing, oil cake not only has a porous structure that can absorb moisture but also retains the unique aroma of the raw materials. Therefore, in the dehydration and drying process, using oil cake can theoretically enhance the aroma of the oil, or at least ensure that the aroma components remaining in the oil cake can compensate for the flavor loss during processing, achieving the effect of "preserving aroma" or even "enhancing aroma."

[0003] Current production operations mostly rely on fixed process parameters. For example, based on the approximate batch size of the oil, the estimated total amount of cake is added at once or in two batches, and a constant stirring speed is used throughout the process. This one-size-fits-all control method cannot adapt to the fluctuations in the initial water content of different batches of crude oil, and it also ignores the dynamic changes in the system state during the dehydration process.

[0004] This extensive control method leads to several problems. First, to ensure dehydration, operators often tend to use strong stirring intensity and long processing times. However, this intensifies mass transfer between the oil phase and the top gas phase, causing a large amount of valuable volatile aroma compounds to escape. This not only fails to enhance the aroma of the cake itself but also results in a bland final product that lacks the characteristics of a strong-aroma oil. Second, reducing stirring intensity to preserve aroma leads to uneven suspension of the cake in the oil, low mass transfer efficiency, a lengthy dehydration process, and increased energy consumption. Furthermore, due to a lack of real-time monitoring of the adsorbent's working status, operators cannot accurately determine when the cake's adsorption capacity approaches saturation. Often, the adsorbent has already failed, but the equipment continues to operate ineffectively, resulting in wasted energy and time. The entire process is highly dependent on the operator's experience, making it difficult to guarantee consistent product quality.

[0005] Therefore, how to achieve a balance between the dehydration process and the aroma preservation goal, fully explore the aroma preservation potential of cake as an adsorption medium, and dynamically optimize the operating parameters according to the real-time status of the process are technical problems that urgently need to be solved in this field. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a system that can overcome the defects of aroma substance loss and poor dehydration efficiency caused by the crude control method in the existing dehydration and drying process of strong aroma oil, and realize the dynamic optimization and adaptive control of dehydration process parameters.

[0007] To address the aforementioned technical problems, this invention provides an adaptive control system for aroma preservation, dehydration, and drying parameters of a strong-aroma oil press. This system includes a processing and execution unit, a multi-dimensional sensing unit, and a central control and computing unit. These three components constitute a closed-loop control system. The processing and execution unit includes a processing vessel for containing the oil to be processed and cake oil as an adsorbent, as well as a cake oil feeder and a variable-speed stirring module connected to the processing vessel. The multi-dimensional sensing unit is used to collect process parameters within the processing vessel in real time. The central control and computing unit is connected to the processing and execution unit and the multi-dimensional sensing unit, and performs data processing, calculation decisions, and command issuance.

[0008] The core technical solution of this system lies in the fact that the central control and computing unit is configured to perform the following operations:

[0009] The system receives data on the real-time water content of the oil from the multi-dimensional sensing unit, and calculates the average adsorption saturation of the currently added cake based on the data and the existing cake addition records.

[0010] The average adsorption saturation is compared with a preset saturation efficiency threshold.

[0011] When the average adsorption saturation is greater than or equal to the saturation efficiency threshold, the system determines that the adsorption capacity of the current adsorbent has approached saturation. At this time, the focus of the control strategy is to replenish new, highly active adsorbent, and therefore a cake feeding instruction is generated and sent to the cake feeder.

[0012] When the average adsorption saturation is less than the saturation efficiency threshold, the system determines that the current adsorbent is still in the high-efficiency adsorption range. At this time, the focus of the control strategy is to optimize the mass transfer efficiency of the existing system. Therefore, by solving a multi-objective optimization problem, a stirring speed adjustment command is generated and sent to the variable speed stirring module.

[0013] Through the above methods, this invention achieves intelligent switching between two core control strategies based on real-time evaluation of the adsorbent's working state. When the adsorbent fails, new material is added to restore the dehydration momentum, and when the adsorbent is highly efficient, the stirring intensity is optimized to enhance mass transfer, thereby achieving refined and adaptive control of the entire dehydration process.

[0014] Preferably, the central control and computing unit includes a hybrid prediction model module and a multi-objective adaptive decision module. The hybrid prediction model module is used to receive sensor data and perform state estimation to output the real-time water content of the oil and the average adsorption saturation of the cake. The multi-objective adaptive decision module is used to execute the aforementioned comparison judgment and decision branches to generate corresponding cake addition or stirring speed adjustment commands.

[0015] In a preferred embodiment of the present invention, the mixed prediction model module is configured to calculate the average adsorption saturation using the principle of material balance. Specifically, it is to compare the total mass of water reduced in the oil from the start of the process to the current moment with the maximum mass of water that can be adsorbed by all the cakes added up to the current moment, thereby obtaining a dimensionless index that quantifies the proportion of adsorption capacity utilized by the adsorbent.

[0016] In a preferred embodiment of the present invention, to achieve high-precision online estimation of the real-time water content of oil, the multi-dimensional sensing unit includes a near-infrared spectral probe and a high-frequency dielectric constant sensor. The central control and computing unit utilizes a preset data fusion model to fuse the signals from these two sensors based on different physical principles. Simultaneously, it incorporates the real-time oil temperature collected by a temperature sensor as a compensation input to eliminate the influence of temperature changes on the measurement, thereby obtaining a more accurate and robust water content estimation result compared to a single sensor.

[0017] Preferably, when the average adsorption saturation is less than the saturation efficiency threshold, the multi-objective optimization problem solved by the central control and computing unit aims to find an optimal stirring speed that minimizes the comprehensive performance index function. This comprehensive performance index function is defined as the weighted sum of the normalized values ​​of the dehydration rate, aroma emission intensity, and stirring power. This optimization aims to balance the three mutually constraining objectives of dehydration efficiency, aroma retention, and energy saving.

[0018] Furthermore, preferably, the constraints in the solution process of the multi-objective optimization problem include ensuring that the aroma emission intensity monitored by sensors such as electronic noses does not exceed a preset upper limit for aroma emission intensity. This constraint ensures that while pursuing a high dehydration rate, the flavor quality of the oil is effectively protected, directly achieving the core technological objective of aroma preservation and dehydration.

[0019] In a preferred embodiment of the present invention, when the average adsorption saturation is greater than or equal to the saturation efficiency threshold, the amount of cake added calculated by the central control and calculation unit is not a fixed value, but a dynamically calculated, efficient addition amount that can reduce the overall average adsorption saturation of the newly added cake and the original cake in the pot to the preset target saturation.

[0020] Furthermore, to coordinate with the aforementioned dynamic addition, the central control and processing unit, while sending a addition command to the cake feeder, simultaneously sends a command to the variable-speed stirring module to temporarily increase the stirring speed to a preset instantaneous mixing speed, and then restores it after a preset time. This coordinated operation ensures that the newly added cake feed can be quickly and evenly dispersed into the oil phase, avoiding clumping, thereby immediately exerting its adsorption effect.

[0021] Preferably, the system also includes process termination logic. The central control and processing unit terminates the main control loop when it determines that the real-time water content of the oil has reached the preset process target value. Furthermore, to ensure the safety and economy of system operation, the system will also forcibly terminate the control loop and provide a prompt when the total processing time exceeds the maximum allowable processing time, or when the dehydration rate remains below the minimum threshold for a continuous period (indicating process stagnation).

[0022] In a preferred embodiment of the present invention, after the control cycle terminates, the central control and computing unit is further configured to issue a command to the variable-speed stirring module to adjust its stirring speed to a preset low-speed settling preparation speed. This operation does not completely stop the process, but rather prevents the water-saturated cake from rapidly settling at the bottom of the processing vessel and forming hard, difficult-to-handle cake blocks through slow stirring, thus facilitating subsequent solid-liquid separation processes.

[0023] This invention provides an adaptive control system for aroma preservation, dehydration, and drying parameters in the extraction and drying of aromatic oils. It offers the following advantages:

[0024] 1. The system of this invention can accurately determine the working state of the adsorbent by calculating and monitoring the average adsorption saturation of the cake in real time. When the system determines that the adsorbent is approaching saturation, it immediately decides to replenish the cake to quickly restore the mass transfer driving force; when the adsorbent is in the high-efficiency range, the system enhances the mass transfer process by optimizing the stirring speed. This event-driven dynamic control strategy ensures that the entire dehydration process is maintained at a high reaction rate, thereby effectively shortening the batch processing time and improving equipment utilization.

[0025] 2. When optimizing the stirring speed, this invention takes the aroma emission intensity as a key optimization objective and sets an upper limit as a hard constraint. This means that when the system seeks the optimal stirring speed to enhance dehydration, it will actively avoid or limit violent stirring conditions that may lead to excessive aroma emission, thereby achieving an effective balance between dehydration efficiency and product flavor quality, and ensuring the rich aroma characteristics of the final product.

[0026] 3. The closed-loop adaptive control system of this invention integrates online sensing, state estimation, decision optimization, and automatic execution, achieving unmanned intelligent management of the entire dehydration process. It replaces intermittent operation and subjective judgment relying on human experience, and through precise algorithms for continuous dynamic optimization, it not only significantly reduces energy consumption and the ineffective use of cake residue, thus reducing operating costs, but also ensures that each batch of product completes processing under similar optimal paths, thereby guaranteeing a high degree of consistency in the moisture content and flavor quality of the final product. Attached Figure Description

[0027] Figure 1 This is a system logic functional block diagram of one embodiment of the present invention;

[0028] Figure 2 This is a flowchart of a method according to an embodiment of the present invention.

[0029] Among them, 100 is the processing and execution unit; 110 is the processing vessel; 111 is the jacket temperature control module; 112 is the variable speed stirring module; 120 is the cake feeder; 200 is the multi-dimensional sensing unit; 211 is the near-infrared spectroscopy probe; 212 is the high-frequency dielectric constant sensor; 221 is the electronic nose; 222 is the temperature sensor; 223 is the stirring power or torque sensor; 300 is the central control and computing unit; 310 is the hybrid prediction model module; and 320 is the multi-objective adaptive decision-making module. Detailed Implementation

[0030] To better understand this invention, it will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other. The dried cake used in this invention is a dried cake that has undergone crushing and drying treatment, and will be referred to as dried cake throughout this document.

[0031] See attached document Figure 1 , Figure 1 This is a system logic functional block diagram according to an embodiment of the present invention. The present invention provides an adaptive control system for aroma preservation, dehydration, and drying parameters of a strong-aroma oil press. The system includes a processing and execution unit 100, a multi-dimensional sensing unit 200, and a central control and calculation unit 300. The three units are interconnected to form a closed-loop system for feedback control.

[0032] The processing and execution unit 100 is a collection of devices for performing a physical adsorption dehydration process. Specifically, it includes a processing vessel 110 for containing the oil to be processed. The processing vessel 110 is a sealable container made of food-grade stainless steel and is used to contain the oil to be processed and cake as an adsorbent.

[0033] To control the process conditions inside the reactor, the processing reactor 110 is equipped with a jacket temperature control module 111. This module circulates a heat exchange medium (such as water or heat transfer oil) in the outer jacket or internal coil of the processing reactor 110 and is linked with the external heating and cooling units. According to the instructions of the central control and calculation unit 300, it maintains the oil temperature within a preset low temperature range.

[0034] The processing vessel 110 is also equipped with a variable speed stirring module 112, which includes a stirring shaft and stirring blades driven by a variable frequency motor. The stirring blades are selected to be suitable for solid-liquid suspension systems, such as a three-bladed propeller or a folding-blade turbine, to ensure that the cake is uniformly suspended in the oil without excessive shearing. The variable frequency motor achieves stepless adjustment of the stirring speed by receiving a frequency signal output from the central control and computing unit 300.

[0035] The processing and execution unit 100 also includes a cake feeder 120 connected to the processing vessel 110. The cake feeder 120 is preferably a loss-in-weight screw feeder, which can feed cake into the processing vessel 110 in batches at a controlled rate and with precise metering according to the instructions of the central control and calculation unit 300.

[0036] The multidimensional sensing unit 200 is a collection of online analytical and measuring instruments used to acquire key process parameters in real time without intervention. These sensors are installed on the tank wall, top cover, or dedicated external circulation pipeline of the processing vessel 110 to ensure the representativeness and continuity of measurements.

[0037] The multidimensional sensing unit 200 includes a near-infrared spectral probe 211 and a high-frequency dielectric constant sensor 212. The near-infrared spectral probe 211 is a fiber optic probe that emits near-infrared light into the oil and detects the absorption spectral intensity at a specific wavelength (e.g., near the absorption peak of water molecules). The high-frequency dielectric constant sensor 212 reflects the water content by measuring the change in the dielectric constant of the oil-water mixture under a high-frequency electric field. Together, they provide complementary information about the trace water content in the oil.

[0038] The multi-dimensional sensing unit 200 also includes an electronic nose 221, a temperature sensor 222, and a stirring power or torque sensor 223. The gas sampling port of the electronic nose 221 is located in the top gas phase space of the processing vessel 110. Its internal gas sensor array responds to volatile aroma substances and uses the changing trend of the response spectrum as a quantitative indicator of aroma emission intensity. The temperature sensor 222, such as a Pt100 platinum resistance thermometer, is immersed in the oil to measure the real-time temperature. The stirring power or torque sensor 223 reflects the actual energy consumption of the stirring process by monitoring the output data of the variable frequency motor.

[0039] The central control and processing unit 300 is the core unit for receiving data, performing calculations, and issuing control commands. It is typically an industrial computer or a programmable logic controller (PLC) with corresponding computing capabilities. Its internal functional modules include:

[0040] The hybrid prediction model module 310 is used to receive and process data from the multi-dimensional sensing unit 200, estimate the current oil condition based on the built-in algorithm, and predict the future trend of key parameters.

[0041] The multi-objective adaptive decision module 320 is used to calculate the optimal control operation command for the next step based on the output of the hybrid prediction model module 310 and in combination with the preset process objectives through an optimization algorithm.

[0042] During system operation, the multi-dimensional sensing unit 200 transmits real-time collected oil and process data to the central control and computing unit 300. The central control and computing unit 300 analyzes, predicts, and optimizes the received data, then generates control commands and outputs them to the processing and execution unit 100. The variable-speed stirring module 112 and the cake feeder 120 of the processing and execution unit 100 adjust the stirring speed or perform cake feed feeding operations according to the received commands, thereby achieving dynamic control of the dehydration process.

[0043] See attached document Figure 2 , Figure 2 This is a flowchart of a method according to an embodiment of the present invention. The overall workflow of the adaptive control method of the present invention includes the following steps:

[0044] Step S100: System initialization and baseline characterization. The strong-aroma crude oil to be processed is fed into the processing vessel 110. The central control and computing unit 300 instructs the multi-dimensional sensing unit 200 to collect the initial state data of the oil, wherein the data from the multi-dimensional sensing unit 200 is used to estimate the initial water content. Data from the electronic nose 221 was used to calibrate the initial aroma background baseline. .

[0045] Step S200: Perform the initial dosing operation and start the process. The hybrid prediction model module 310 within the central control and computing unit 300, based on the initial water content... Calculate the initial amount of dried cake added. and initial stirring speed Subsequently, the central control and computing unit 300 issues a command to drive the cake feeder 120 and the variable speed mixing module 112 to perform the operation.

[0046] In step S300, the system enters the main control loop. This loop continues to run until the process termination condition is met. Its specific internal steps are as follows:

[0047] Step S310: Perform online monitoring and data acquisition. During the process, the multi-dimensional sensing unit 200 continuously collects real-time data and transmits it to the central control and processing unit 300.

[0048] Step S320: Perform state estimation and trend prediction. The hybrid prediction model module 310 of the central control and computing unit 300 runs, fusing real-time data from the multi-dimensional sensing unit 200 to estimate the water content at the current moment. The average adsorption saturation of the cake was predicted based on the mass transfer model. .

[0049] Step S330: Execute decision optimization and instruction generation. The multi-objective adaptive decision module 320 determines the adsorption saturation based on the predicted adsorption saturation. The saturation level is compared with a preset threshold. If the saturation level is high, a new instruction to add dry cake is generated; if the saturation level is in the efficient range, an instruction to adjust the stirring rate is generated by solving a multi-objective optimization function.

[0050] Step S340: Execute control commands. The central control and arithmetic unit 300 sends the commands generated in the previous step to the corresponding modules (variable speed stirring module 112 or cake feeder 120) of the processing and execution unit 100 to perform physical operations.

[0051] Step S400: Execution process termination judgment. At the end of each main control cycle, the system determines the current water content. Whether the preset process target value has been achieved If the condition is not met, the process returns to step S310 and continues to execute the next control loop; if the condition is met, the process ends the main loop.

[0052] Step S500: Process completion operation. The central control and arithmetic unit 300 sends a process completion signal, instructing the variable speed stirring module 112 to stop or switch to low speed operation to indicate that the dehydration and drying process for this batch has been completed.

[0053] The steps of the method of the present invention will be described in detail below.

[0054] Step S100 is performed before the physical adsorption dehydration process officially starts, aiming to obtain the initial state parameters of the oil to be treated, providing an accurate starting point and reference for subsequent closed-loop control. This step may include the following steps:

[0055] Step S101: Initial state data acquisition. A batch of strong-aroma crude oil to be processed is pumped into the processing vessel 110 via pipeline. After the oil level stabilizes, the central control and processing unit 300 issues a command to activate the multi-dimensional sensing unit 200 to perform an initial state scan of the oil in the processing vessel 110. This scan simultaneously acquires the following data:

[0056] Initial near-infrared spectral data acquired by near-infrared spectral probe 211 ;

[0057] The initial dielectric constant value collected by the high-frequency dielectric constant sensor 212 ;

[0058] Initial gas fingerprint data collected by the electronic nose 221 located in the top air phase space of the processing vessel 110;

[0059] The initial temperature of the oil collected by temperature sensor 222 .

[0060] All collected initial data are transmitted to the central control and processing unit 300 for processing.

[0061] Step S102 involves calculating and storing the initial baseline parameters. The hybrid prediction model module 310 within the central control and computation unit 300 receives the initial data collected in step S101 and performs calculations to establish the initial baseline parameters: initial water content. and initial aroma background baseline .

[0062] initial moisture content Through a pre-built data fusion model This model achieves information fusion from sensor data derived from different physical principles, specifically as follows:

[0063] ;

[0064] in, The initial water content of the oil to be treated can be expressed as a percentage by mass (%) or mg / kg. The initial near-infrared spectral data can be the absorbance values ​​at the water molecule absorption peaks (e.g., around 1450 nm and 1940 nm) or the spectral feature vectors after preprocessing (e.g., differentiation, smoothing). This is the initial dielectric constant value; The initial oil temperature is used as the compensation input for the model to correct the effect of temperature on near-infrared spectroscopy and dielectric constant measurements.

[0065] As a data fusion model, in one specific embodiment, This can be a multivariate calibration model, such as a partial least squares regression (PLS) model or a principal component regression (PCR) model established using chemometric methods. The method for establishing this model involves preparing a series of standard oil samples with known moisture content, accurately determining their moisture content using national standard methods such as Karl Fischer titration as the true value, simultaneously collecting their near-infrared spectra and dielectric constant data at different temperatures, and finally establishing the mathematical relationship between the input data and the true moisture content through regression analysis. The establishment of such multivariate calibration models is a well-known technique in this field and will not be elaborated further here.

[0066] Initial aroma background baseline This is accomplished by processing the initial gas fingerprint data collected by the electronic nose 221. The electronic nose 221 contains an array of various types of gas sensors, and its initial output is a response vector. ,in For the number of sensors, For the first The response values ​​of each sensor at the initial moment. For ease of subsequent comparison, this response vector can be... Processed as a scalar value as the aroma background baseline In one specific embodiment, The calculation method can be the Euclidean norm of the response vector:

[0067] ;

[0068] In another embodiment, It can also be a weighted sum of the response values ​​of one or more sensors in the vector that are most sensitive to a specific aroma component.

[0069] After the calculation is completed, the initial water content and initial aroma background baseline The data is stored in the storage area of ​​the central control and computing unit 300, serving as the basis for the initial process parameter calculation in the subsequent step S200, and as the benchmark for process monitoring and comparison in the main control loop.

[0070] In step S200, physical operations are calculated and executed based on the initial state parameters to bring the system into a controllable initial operating state. This step may include the following steps:

[0071] Step S201: Calculate the initial process parameters. The hybrid prediction model module 310 within the central control and computing unit 300 calls the initial moisture content obtained in step S101. Based on the pre-set process objectives and material property parameters within the system, the initial amount of cake added during this dewatering process was calculated. and initial stirring speed .

[0072] In one specific embodiment, the calculation process may include:

[0073] First, calculate the total amount of cake theoretically required to complete this dehydration task. This calculation is based on the principle of material balance:

[0074] ;

[0075] in, The total mass of the oil to be processed in the treatment vessel 110 can be measured in advance by a level gauge or a weighing sensor and input into the system. The preset target moisture content for the dehydration process, for example, 0.05%; The maximum moisture absorption capacity per unit mass of dried cake is a parameter that is determined through offline experiments and pre-stored in the system as an inherent property of the material. As a safety factor, it is usually taken to be greater than 1 (e.g., 1.1 to 1.3) to compensate for non-ideal factors in the actual mass transfer process and ensure that moisture can be completely removed.

[0076] Then, calculate the initial amount of dried cake added. The initial addition is not the entire amount, but a specific proportion of the total theoretical amount, to avoid drastic changes in the system's properties and uneven mixing caused by a sudden addition of a large amount of powder. The calculation formula is as follows:

[0077] ;

[0078] in, The initial dosage ratio is a preset value between 0 and 1, such as 0.2 to 0.5. The choice is designed to quickly establish an effective mass transfer drive while leaving room for subsequent adaptive batching.

[0079] Initial stirring speed This is the preset reference speed within the system. This speed is an optimized value determined through prior hydrodynamic simulation or small-scale experiments, based on the oil viscosity, the geometry of the processing vessel, and the characteristics of the selected impeller. The selection of this value aims to provide sufficient turbulent kinetic energy to achieve rapid and uniform dispersion of the initial feed cake during the initial process startup, while controlling the stirring shear force to a level that is insufficient to cause significant aroma substance loss.

[0080] Step S202: Perform the initial process operation. The central control and calculation unit 300 converts the parameters calculated in step S201 into control commands and sends them to the corresponding modules of the processing and execution unit 100. Specifically:

[0081] A command is sent to the cake feeder 120 to precisely feed a mass of [amount missing] into the processing vessel 110. Dry cakes;

[0082] A command is sent to the variable speed mixing module 112 to start it and set its speed. Stir.

[0083] In the main control loop of step S300, step S310 synchronously collects various dynamic parameters of the oil and process environment in the processing vessel 110 at a preset time frequency, providing real-time data input for subsequent state estimation and decision optimization.

[0084] At the beginning of each control cycle, the central control and computing unit 300 sends data acquisition commands to all sensor modules within the multidimensional sensing unit 200. Upon receiving the commands, the sensors under the multidimensional sensing unit 200 operate in parallel to acquire the physical or chemical signals at the current moment.

[0085] In one specific embodiment, the synchronous acquisition process specifically acquires the following data:

[0086] The near-infrared spectral probe 211 in the multi-dimensional sensing unit 200 collects data at the current moment. Near-infrared spectral data .

[0087] The high-frequency dielectric constant sensor 212 in the multidimensional sensing unit 200 measures the current time. Macroscopic dielectric constant of oil-water mixture .

[0088] The electronic nose 221 in the multi-dimensional sensing unit 200 collects the current time. The gas fingerprint data in the headspace gas phase of reactor 110 is processed, and its output is a vector containing the response values ​​of each member of the gas sensor array. .

[0089] The temperature sensor 222 in the multi-dimensional sensing unit 200 measures the current time. Real-time temperature of oil .

[0090] The stirring power or torque sensor 223 in the multi-dimensional sensing unit 200 measures the current moment. Real-time energy consumption parameters of variable speed stirring module 112 .

[0091] The above-described acquisition process is performed at a preset sampling interval. It is executed repeatedly in a periodic manner.

[0092] After the data collection is completed, all at the same time The data obtained includes , , , as well as Packed into timestamps The data frame is transmitted to the central control and processing unit 300 via a data bus or wireless communication for processing in subsequent steps S320.

[0093] In step S320 of the main control loop, the raw sensor data acquired in step S310 is transformed into state variables and predicted indicators. This step may specifically include the following steps:

[0094] Step S321: Perform online fusion estimation of oil water content. The hybrid prediction model module 310 calls the data fusion model established during the initialization phase. and apply it to the current moment. The real-time sensor data is used to calculate the water content of the oil at that moment. This calculation process is repeated in each control cycle, providing continuous tracking of water content. The calculation formula is consistent with the model used during initialization, but the input is dynamic data that changes over time:

[0095] ;

[0096] in, For at any time Estimated water content of the oil; , and At time respectively Simultaneously acquired near-infrared spectral data, high-frequency dielectric constant values, and oil temperature. As mentioned earlier, The pre-defined multivariate calibration model integrates measurements from different physical principles to obtain water content estimation results.

[0097] Step S322: Dynamically predict the adsorption saturation of the cake oil. Obtain the current moisture content. Subsequently, the hybrid prediction model module 310 further calculates the average adsorption saturation of the currently added cake. This indicator quantifies the proportion of adsorption capacity utilized by the adsorbent and serves as the basis for judging different control strategy branches in subsequent step S330.

[0098] Average adsorption saturation The calculation is based on the principle of material balance, which is defined as the time from the start of the process to the current moment. The total mass of water adsorbed, and up to the current moment The ratio of the total mass of cake oil added to the maximum water mass that can be absorbed. The calculation formula is:

[0099] ;

[0100] in, For at any time The average adsorption saturation of the cake is a dimensionless value. The initial water content (mass fraction) of the oil determined in step S102 is a constant throughout the entire batch processing. The real-time moisture content (mass fraction) is estimated in step S321 of the current control cycle, and is a variable that changes over time. The total volume of the oil in reactor 110 is a known constant. The density of the oil can be considered a constant or determined based on the real-time temperature. The variables to be corrected; From the start of the process to the current moment The total mass of the dried cake added is obtained by the central control and calculation unit 300 in real time based on the feeding records of the dried cake dispenser 120. The maximum hygroscopic capacity per unit mass of dried cake is denoted as , and is a constant determined in advance through material experiments.

[0101] By utilizing the latest [data / technology] in each control cycle Values ​​are calculated and updated in real time. The system can clearly define the process of depletion of the adsorption capacity of the cake.

[0102] After step S320 is completed, the current water content of the oil is recorded. and average adsorption saturation of cake The input is passed to the subsequent step S330 for decision-making and optimization of the control strategy.

[0103] In step S330 of the main control loop, the optimal control operation command is generated based on the current system state and guided by a preset process objective. This step may specifically include the following steps:

[0104] Step S331 involves event-driven decision-making. The central control and computing unit 300 first processes the real-time average adsorption saturation of the dried cake calculated in step S322. , and the preset saturation efficiency threshold Compare. The threshold. The critical point representing the start of a significant decrease in cake adsorption efficiency can be set between 0.7 and 0.9 in one specific embodiment. This comparison forms the basis of event-driven decision-making, and the system autonomously selects to enter different control strategy branches based on the comparison results.

[0105] Step S332: Based on the decision judgment result, perform the corresponding control quantity calculation. This step has two parallel logical branches:

[0106] Branch 1: When the judgment result is At this point, the system determines that the adsorption capacity of the current adsorbent is approaching saturation, and further enhancing mass transfer through methods such as increased stirring will have limited effect on improving the dehydration rate. Therefore, the control strategy focuses on supplementing the adsorbent with new, highly active material. The system then performs the following operations:

[0107] Generate a new dry cake addition instruction, that is, determine the dry cake addition amount for the next control cycle. .

[0108] In one specific embodiment, the dosage can be a preset fixed batch value. .Should The value is determined during system configuration based on the minimum precise dosage of the cake feeder 120 per cycle and the minimum amount of powder required to effectively restore the adsorption rate, as determined in process experiments.

[0109] In another embodiment, It can also be dynamically calculated based on the remaining amount of water to be removed. The goal is to [addition / removal / etc.]. Afterwards, the overall average adsorption saturation of the system can be immediately reduced to the preset target saturation. (For example, set to) (70%-90%). Its calculation formula can be obtained by solving the following equation:

[0110] ;

[0111] Solve We can obtain:

[0112] ;

[0113] The dosage calculated in this way can restore the average adsorption capacity of all the old and new cakes to a high efficiency level, achieving more precise control.

[0114] Meanwhile, to ensure that the newly added cake is quickly and evenly dispersed into the oil phase and to prevent clumping, the system implements a stirring strategy linked to the addition action. Specifically, when the central control and computing unit 300 drives the cake feeder 120 to perform the addition operation, it simultaneously sends a command to the variable speed stirring module 112 to change the stirring speed from the current speed... Temporarily increase to the preset instantaneous mixing speed .Should The stirring speed is predetermined according to fluid dynamics principles, and is the minimum rotational speed sufficient to overcome the surface tension of the powder and rapidly wet and disperse it, but lower than the speed that would cause violent eddying or material splashing on the liquid surface. After this high-speed stirring state is maintained for a preset time (e.g., 30 to 180 seconds), the stirring speed is restored to the speed before the acceleration rate was increased. It waits for the next control cycle to recalculate and set the optimal operating speed.

[0115] Branch 2: When the judgment result is At this point, the system determines that the adsorbent is still within its highly efficient adsorption range and no additional material is needed. The control strategy then focuses on optimizing the mass transfer efficiency of the existing system, i.e., finding the optimal stirring intensity while meeting aroma preservation and energy-saving constraints. Therefore, the system performs the following operations:

[0116] Set the amount of dried cake added in the next control cycle to zero, that is... .

[0117] The optimal stirring speed for the next control cycle is calculated by solving a multi-objective optimization problem. The optimization problem aims to find a stirring speed that results in a comprehensive performance index function. To reach the minimum value. In a specific embodiment, this function can be defined as:

[0118] ;

[0119] This optimization problem needs to be solved under the following constraints:

[0120] 1. ;

[0121] 2. ;

[0122] in:

[0123] The decision variable is the stirring speed;

[0124] , , These are the weighting coefficients for aroma retention, energy saving, and dehydration efficiency, respectively, and their sum is 1. They can be preset according to production priorities.

[0125] , , These are the normalized values ​​for aroma emission intensity, stirring power, and dehydration rate at the current moment, respectively. Normalization is a dimensionless operation that linearly maps each physical quantity to an interval to eliminate the influence of different units. For example, for aroma emission intensity... Its normalized value It is possible Calculation, where This is the highest peak ever recorded in history;

[0126] The dehydration rate can be calculated from the change in moisture content between two consecutive control cycles: ;in The estimated water content for the current control cycle. The moisture content estimated for the previous control cycle. To control the time interval of the cycle;

[0127] , These are the lower and upper limits of the permissible operating speed for stirring.

[0128] The upper limit of the aroma emission intensity is a safety constraint used to ensure the flavor quality of the product.

[0129] The multi-objective optimization problem here is solved by the central control and computing unit 300 calling its built-in numerical optimization algorithm program to solve the aforementioned constrained optimization problem. For solving such optimization problems, those skilled in the art can use sequential quadratic programming (SQP), interior point methods, or heuristic algorithms such as particle swarm optimization (PSO) to achieve the solution. The specific algorithm implementations are well-known technologies in this field and will not be described in detail here.

[0130] After step S330 is completed, it outputs the amount of dry cake to be added in the next control cycle. and stirring speed .

[0131] In step S340 of the main control loop, the decision result from step S330 is converted into the actual physical action of the processing and execution unit 100. This step may specifically include the following steps:

[0132] Step S341 involves converting and sending control commands. The central control and processing unit 300 calculates the stirring speed for the next control cycle. And the amount of dried cake added This is converted into hardware control signals that can be directly recognized and executed by the processing and execution unit 100.

[0133] In one specific embodiment, regarding the stirring speed This conversion process can generate an analog voltage or current signal proportional to the target rotational speed. For the amount of cake added... If the cake feeder 120 is a screw feeder, the conversion process can be based on a pre-calibrated feed rate per unit time to calculate the feeder running time required to achieve the target feed rate. These control signals are sent to the corresponding controller of the processing and execution unit 100 via a data interface.

[0134] Step S342: Perform physical control operations. After receiving hardware control signals from the central control and arithmetic unit 300, each module of the processing and execution unit 100 performs the corresponding physical operations.

[0135] The controller of the variable speed stirring module 112 receives the representative Upon receiving the signal, the operating frequency of the drive motor is adjusted, thereby adjusting the speed of the agitator to the target value. .

[0136] The controller of the dry cake dispenser 120 receives a representative Upon receiving the signal, the corresponding dosing action is executed. If The value is greater than zero, and the controller drives the feeding mechanism for a calculated duration to deliver a mass of [value missing]. The dried cake is added to the processing vessel 110. If If the value is zero, the cake feeder 120 will not perform any operation during this control cycle.

[0137] Through the execution of step S340, the decision of the central control and computing unit 300 is materialized into physical intervention in the dehydration process, forming a complete closed-loop control cycle.

[0138] In step S400, at the end of each main control loop cycle, it is determined whether the main control loop S300 should continue execution or terminate the entire adaptive control process. This step may specifically include the following steps:

[0139] Step S401: Determine the termination condition. At the end of each control cycle, the central control and arithmetic unit 300 will update the latest estimated real-time water content from step S321. Compared with the preset process target value Comparison. Here. This represents the upper limit of moisture content required for the final product quality, such as 0.05% or 500 mg / kg. This value is set by the operator or loaded from the process formulation before the process begins.

[0140] Step S402: Jump to the next step based on the judgment result.

[0141] If the comparison result of step S401 is If the result is negative, it indicates that the dehydration process is not yet complete and needs to continue. Therefore, the main control loop continues, and the process returns to step S310 to begin the next control cycle of online monitoring, state estimation, decision optimization, and instruction execution.

[0142] If the comparison result is If the moisture content reaches or falls below the process target value, the dehydration process is considered complete. At this point, the termination condition of the main control loop S300 is met, the process exits the loop, and proceeds to step S500 to execute the process termination operation.

[0143] To ensure the robustness and security of the system, in one specific embodiment, the process termination judgment here may include one or more auxiliary judgment conditions in addition to the judgment based on the water content target mentioned above.

[0144] One auxiliary criterion is the total processing time, which is recorded by the system's internal timer from the start of the process. And compare it with the preset maximum allowed processing time. Compare. If Even if the moisture content does not reach The system will also forcibly terminate the main control cycle and simultaneously output an alarm signal to prevent indefinite operation due to unforeseen circumstances.

[0145] Another auxiliary criterion is that the dehydration rate has stagnated. The system continuously monitors the dehydration rate. If, within a continuous time window (e.g., five or more consecutive control cycles), the absolute value of this rate remains below a preset minimum threshold... The system can determine that the process has stalled (e.g., because the adsorbent has completely failed and cannot be replenished), and at this time the cycle can be terminated in advance to avoid unnecessary energy consumption.

[0146] After the main control loop exits due to the fulfillment of the termination condition, the method of the present invention immediately executes step S500, i.e., the process termination operation. This step aims to safely and orderly bring the entire system into the final state of batch processing and prepare for subsequent material handling processes. This step may specifically include the following steps:

[0147] Step S501: Issue process end instructions. After confirming that the main control loop has terminated, the central control and arithmetic unit 300 generates and issues a set of process end instructions to the processing and execution unit 100.

[0148] Step S502: Perform final adjustment of the device state. Each module of the processing and execution unit 100 performs the corresponding final operation according to the received end command. In one specific embodiment, this operation includes:

[0149] The central control and processing unit 300 sends a command to the variable speed stirring module 112 to adjust its stirring speed to a preset final state. This final state can be a complete stop or a speed prepared for slow settling. The purpose of this low-speed operation (e.g., 5-20 RPM) is to slowly agitate the oil while waiting for subsequent solid-liquid separation operations, preventing the water-saturated cake from settling rapidly at the bottom of the treatment vessel 110 and forming hard, difficult-to-handle cake lumps.

[0150] At the same time, the central control and computing unit 300 sends a command to the jacket temperature control module 111 to stop heating or cooling, so that the material temperature in the processing vessel 110 changes naturally.

[0151] In addition, the system will also confirm that other auxiliary equipment, such as the dry cake feeder 120, are in a stopped and locked safe state.

[0152] Step S503: Perform final data recording and status indication. The central control and processing unit 300 executes data archiving and operation instructions. The specific implementation method for this process termination operation can be as follows:

[0153] The system organizes and records complete process data for this batch of processing, such as total processing time and cumulative total amount of cake oil added. Initial moisture content With final moisture content The key information is packaged into a batch processing report and stored in the database.

[0154] The system displays a "dehydration and drying complete" status message on the human-machine interface and can illuminate indicator lights or trigger audible prompts to inform on-site operators. In another embodiment, the system can also send a digital signal indicating the process is complete to the upper-level manufacturing execution system via industrial Ethernet to achieve automated scheduling at the factory level.

[0155] At this point, a complete execution cycle of the adaptive control system for aroma preservation, dehydration, and drying parameters of the strong-aroma oil extraction system provided by this invention has been completed.

[0156] 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 system for adaptive control of parameters of a full-bodied oil-pressing aroma-preserving dewatering and drying, characterized in that, include: The processing and execution unit includes a processing vessel for containing the oil to be processed and cake, a cake feeder connected to the processing vessel, and a variable speed stirring module for stirring inside the processing vessel. The multi-dimensional sensing unit is used to collect process parameters inside the processing vessel in real time and output data including real-time water content of oil, aroma emission intensity and stirring energy consumption. A central control and computing unit, connected to the processing and execution unit and the multi-dimensional sensing unit, is configured as follows: When no dry cake is initially added, the initial moisture content is estimated based on the data from the multi-dimensional sensing unit. The total dry cake amount required theoretically is calculated based on the initial moisture content. A preset proportion of the total dry cake amount is used as the first dry cake addition amount. An initial dry cake addition command is generated and sent to the dry cake adder. After the initial addition of cake oil, if the real-time water content of the oil exceeds the preset process target value, the following control loop is executed: The system receives data from the multi-dimensional sensing unit and calculates the average adsorption saturation of the currently added cake based on the real-time water content of the oil and the executed cake addition records. The average adsorption saturation is compared with a preset saturation efficiency threshold. When the average adsorption saturation is greater than or equal to the saturation efficiency threshold, a cake feeding instruction for batch replenishment is generated and sent to the cake feeder. Furthermore, when the average adsorption saturation is less than the saturation efficiency threshold, a stirring speed adjustment command is generated and sent to the variable speed stirring module by solving a multi-objective optimization problem. The multi-objective optimization problem is to find the stirring speed that minimizes the comprehensive performance index function under the constraint that the aroma emission intensity does not exceed a preset upper limit. The comprehensive performance index function is defined as the weighted sum of the normalized values ​​of the dehydration rate, aroma emission intensity, and stirring power.

2. The lour oil deodorization drying parameter adaptive control system according to claim 1, characterized in that, The central control and computing unit includes: A hybrid prediction model module is used to receive data from the multi-dimensional sensing unit to estimate the real-time water content of the oil and calculate the average adsorption saturation; and The multi-objective adaptive decision module is used to compare the average adsorption saturation with the saturation efficiency threshold, and generate the cake addition command based on the comparison result, or to solve the multi-objective optimization problem to generate the stirring speed adjustment command.

3. The lour oil deodorization drying parameter adaptive control system according to claim 2, characterized in that, The hybrid prediction model module is configured as follows: Based on material balance, the average adsorption saturation is calculated by comparing the total mass of water adsorbed from the start of the process to the current moment with the maximum mass of water that can be adsorbed by the total mass of cake cake added up to the current moment.

4. The lour oil deodorization drying parameter adaptive control system according to claim 1, characterized in that, The multidimensional sensing unit includes a near-infrared spectral probe and a high-frequency dielectric constant sensor; and the central control and processing unit is configured as follows: Using a pre-set data fusion model, data collected by the near-infrared spectral probe and the high-frequency dielectric constant sensor are fused together, and the real-time oil temperature collected by the temperature sensor is used as a compensation input to estimate the real-time water content of the oil.

5. The lour oil deodorization drying parameter adaptive control system according to claim 1, characterized in that, When the average adsorption saturation is greater than or equal to the saturation efficiency threshold, the amount of cake added calculated by the multi-objective adaptive decision module is determined to be just enough to reduce the overall average adsorption saturation of the newly added cake and the original cake in the pot to the preset target saturation.

6. The lour oil deodorization drying parameter adaptive control system according to claim 5, characterized in that, The central control and computing unit is also configured as follows: While sending a feeding instruction containing the amount of cake to the cake feeder, a command is simultaneously sent to the variable speed mixing module to temporarily increase the mixing speed to a preset instantaneous mixing speed, and then restore it after a preset time.

7. The adaptive control system for aroma preservation, dehydration, and drying parameters of strong-aroma oil extraction according to claim 1, characterized in that, The central control and computing unit is also configured as follows: The control loop terminates when the real-time water content of the oil reaches the process target value; or... The control loop is forcibly terminated when the total processing time exceeds the maximum allowable processing time, or when the dehydration rate falls below the minimum threshold within a continuous time window.

8. The adaptive control system for aroma preservation, dehydration, and drying parameters of strong-aroma oil extraction according to claim 7, characterized in that, The central control and computing unit is also configured as follows: After terminating the control cycle, a command is sent to the variable speed stirring module to adjust the stirring speed to a preset low-speed settling preparation speed to prevent the water-saturated cake from settling and clumping at the bottom of the processing vessel.

Images