Method for preparing virgin peanut oil based on near-infrared mildew sorting and low-oxygen temperature control pressing

Abstract

The application discloses a kind of based on near infrared mildew sorting and low oxygen temperature control squeezing native peanut oil preparation method, belong to edible oil processing technical field, solve the problem of not completely, oil oxidation and nutrient component loss in the preparation process of traditional peanut oil in mouldy particle screening. Including: near infrared spectroscopy detection is combined with artificial sorting and removes mouldy particle to obtain selected peanut kernel;After gradient temperature baking, use citric acid solution to soak and ultrasonic auxiliary treatment to remove impurities and inhibit enzyme activity;Utilize the roll surface to contain the roll crusher of staggered triangle tooth pattern and break peanut kernel to specific particle size to optimize oil extraction efficiency;Segmented temperature control and pressure control screw press machine presses peanut crude oil under nitrogen protection, centrifugal separation and ceramic membrane filtration purify oil and fat.The method is accurately controlled crushing parameter, pressing temperature and inert gas environment, effectively reduces oxidation side reaction, retains peanut oil natural flavor and nutrient component, suitable for producing high quality, low acid value, high stability native peanut oil.

Description

Technical Field

[0001] This invention relates to the field of edible oil technology. More specifically, this invention relates to a method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing. Background Technology

[0002] In traditional peanut oil production processes, raw material screening often relies on manual visual sorting or simple mechanical sieving, making it difficult to accurately identify and remove moldy particles. Aflatoxin in moldy peanut kernels is highly concealed, and conventional sorting methods cannot penetrate the seed coat to detect internal contamination, resulting in a persistent risk of toxin residue. This problem stems from the fact that the appearance of peanut kernels does not change significantly in the early stages of mold growth, and manual sorting is greatly affected by operator experience and fatigue, making it difficult to balance sorting efficiency and accuracy. During the roasting process, inadequate temperature and time control can lead to uneven heating of peanut kernels, resulting in both surface charring and insufficient internal dehydration. This not only reduces oil yield but may also cause localized overheating, leading to surface charring and the formation of polycyclic aromatic hydrocarbons (PAHs), while accelerating oil pre-oxidation and reducing the stability of the final finished oil. In the crushing stage, the fixed tooth shape design and roller spacing parameters of ordinary roller crushers result in a large dispersion in particle size distribution after crushing, with an excessively high proportion of fine powder, leading to oil passage blockage or fluctuations in oil extraction efficiency during subsequent pressing. In the pressing process, uneven temperature and pressure distribution in the screw press chamber can easily lead to oil oxidation due to localized high-temperature areas. Simultaneously, oxygen contact during pressing accelerates free radical generation, resulting in increased acid value and excessive peroxide value. Existing centrifugal separation and filtration technologies have limited ability to remove colloidal impurities and suspended particles, especially for particles smaller than 1 μm, leading to decreased transparency and storage stability of the finished oil. The core contradiction lies in insufficient discrete control of process parameters, a lack of real-time dynamic adjustment capabilities in key processes, and the failure to effectively block oxidation reaction pathways, ultimately hindering further improvements in the quality and safety of peanut oil. Summary of the Invention

[0003] One object of the present invention is to solve at least the above-mentioned problems and to provide at least the advantages that will be described later.

[0004] Another objective of this invention is to provide a method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing. By precisely controlling crushing parameters, pressing temperature, and inert gas environment, it effectively reduces oxidation side reactions, preserves the natural flavor and nutrients of peanut oil, and is suitable for producing high-quality, low-acid-value, and high-stability virgin peanut oil.

[0005] To achieve these objectives and other advantages according to the present invention, a method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing is provided, comprising the following steps: 1) Near-infrared hyperspectral mold identification: Raw peanut kernels are scanned by a near-infrared hyperspectral imaging system with a wavelength of 900 nm to 1700 nm. Each peanut kernel is scanned multiple times with pulses. The obtained spectral data is processed by a partial least squares discriminant analysis model to output the mold probability value in real time. When the mold probability value of a single peanut kernel is ≥0.90, it is rejected by a pneumatic nozzle and then manually sorted to obtain selected peanut kernels. 2) Pretreatment: Selected peanut kernels are roasted at 55-65℃ for 20-40 min, then soaked in 0.4-0.6% citric acid solution for 10-20 min, rinsed and dried to obtain pretreated peanut kernels; 3) Directional crushing and particle control: The pre-treated peanut kernels are fed into a roller crusher with a roller spacing of 0.8 mm to 1.2 mm, and crushed for 20 to 30 seconds until the peanut kernels have a particle size of 2 mm to 3 mm. The roller crusher has staggered triangular teeth on the roller surface with a tooth height of 0.5 mm to 0.8 mm and a tooth spacing of 2 mm to 3 mm. 4) Low-oxygen, low-temperature pressing: Peanut kernels are fed into a screw press for pressing to obtain crude peanut oil. The screw press has a length-to-diameter ratio of 8:1 to 10:1 and a screw shaft speed of 15 r / min to 18 r / min. The pressing chamber of the screw press is divided into three sections: the screw lead in the feeding section is 120 mm to 150 mm, the screw lead in the compression section decreases to 60 mm to 80 mm, and the screw lead in the cake discharge section is fixed at 40 mm to 50 mm. During the pressing process, circulating water is introduced into the pressing chamber jacket to control the pressing temperature at 50℃ to 55℃ and maintain the pressing chamber pressure at 50 MPa to 85 MPa. During the pressing process, multiple nitrogen injection points are set in the pressing chamber to continuously inject nitrogen with a purity of ≥99.5%. An oxygen content sensor set at the pressing chamber outlet adjusts the total nitrogen flow rate in real time to control the oxygen content in the pressing chamber at 0.3% to 0.5%. 5) Membrane filtration stabilization: After combining the crude peanut oils and separating them by centrifugation, the crude peanut oils are filtered through a ceramic membrane filter with a pore size of 0.1 μm to 0.5 μm to obtain virgin peanut oil.

[0006] Preferably, step 5) specifically includes: S51. The crude peanut oil is fed into a centrifugal separator and suspended impurities are separated at a speed of 3000 rpm to 4000 rpm for 15 min to 20 min. The inner wall of the centrifugal separator drum is equipped with six sets of vertical baffles, the height of which is 1 / 3 to 1 / 2 of the height of the drum, and the thickness of which is 3 mm to 5 mm. S52. The centrifuged peanut oil is filtered through a ceramic membrane filter with a pore size of 0.1 μm to 0.5 μm to obtain filtered peanut oil; wherein, the operating pressure of the ceramic membrane filter is 0.2 MPa to 0.4 MPa, and the membrane surface flow velocity is 0.8 m / s to 1.5 m / s.

[0007] Preferably, step 2) specifically includes: S21. Spread the selected peanut kernels evenly on the metal mesh belt of the hot air circulating oven, with a thickness of 3 cm to 5 cm. Turn on the transverse airflow distribution device to keep the wind speed at 0.8 m / s to 1.2 m / s. Increase the temperature in a gradient of 55℃ to 65℃: maintain 55℃ for the first 10 minutes to evaporate the surface moisture of the selected peanut kernels, raise the temperature to 60℃ for the middle 10 minutes to promote the diffusion of internal moisture, and stabilize at 65℃ for the last 10 minutes to complete the dehydration. S22. The selected roasted peanut kernels are immediately transferred to a soaking tank. The pH value of the citric acid solution in the soaking tank is 2.3-2.8, and the liquid-to-solid ratio is 5-8 L:1 kg. During the soaking process, an ultrasonic generator is used to apply mechanical vibration at a frequency of 28 kHz-40 kHz, with an ultrasonic power density of 0.3 W / cm³. 2 ~0.5 W / cm 2 ; S23. After soaking, feed the selected peanut kernels into a three-channel countercurrent washing machine. The first channel uses warm water at 40℃~45℃ with a water flow rate of 1.5 m / s. 3 / h~2 m 3 / h, the second channel is filled with cold water with an ozone concentration of 1.5 mg / L to 3.0 mg / L, and the third channel is purged with compressed air to remove residual moisture from the surface; S24. After being cleaned and drained by a vibrating screen, the selected peanut kernels are fed into a track dryer with finned heat dissipation tubes. The drying temperature is 35℃~40℃ and the track conveying speed is 0.5 m / min~0.8 m / min until the moisture content of the peanut kernels drops to 6%~8%, thus obtaining pre-treated peanut kernels.

[0008] Preferably, step S51 specifically includes: heating the crude peanut oil to 35℃~40℃ through a constant temperature preheater, stabilizing it in a buffer tank, and then feeding it into a disc centrifuge at a flow rate of 0.8 L / min~1.2 L / min; separating suspended impurities at a speed of 3000 rpm~4000 rpm; the length-to-diameter ratio of the centrifuge drum is 4:1~5:1; six sets of vertical baffles are evenly distributed circumferentially on the inner wall of the drum; the height of the baffles is 1 / 3~1 / 2 of the height of the drum; the thickness of the baffles is 3 mm~5 mm; the top of the baffles is inclined at 15°~20° towards the direction of drum rotation; and reinforcing ribs with a width of 8 mm~10 mm are provided at the weld between the baffles and the inner wall of the drum. During centrifugation, the change in the conductivity of the oil phase is monitored in real time. The initial conductivity of the crude peanut oil when it enters the centrifuge is used as a benchmark. Separation is terminated when the real-time conductivity decreases by less than 50 μS / cm for 3 consecutive minutes. The separated heavy phase impurities are discharged through the slag discharge port at the bottom of the drum. The slag discharge interval is 15 to 20 minutes. During slag discharge, the pressure fluctuation inside the drum is maintained at ≤0.05 MPa by a pressure compensation device. Before the light phase oil is introduced into the temporary storage tank through the oil outlet pipe, the oil temperature is first reduced to 25℃ to 30℃ by a serpentine tube cooler. The cooling water flow rate in the serpentine tube is 2 m / s to 3 m / s, and the oil flow and cooling water are in countercurrent contact.

[0009] Preferably, step 1) specifically includes: S11. A near-infrared hyperspectral imaging system with a wavelength range of 900 nm to 1700 nm is used, with a line scanning speed of 200 kernels / second to 300 kernels / second. Each peanut kernel is subjected to 3 pulse scans with an interval of 20 ms. S12. The obtained spectral data is processed by a partial least squares discriminant analysis model, which is trained based on the spectral characteristics of more than 3,000 aflatoxin contaminated samples and outputs the mold probability value in real time. When the mold probability value of a single peanut kernel is ≥0.90, the pneumatic nozzle is triggered, and compressed air with a pressure of 0.6 MPa to 0.8 MPa is sprayed to blow the moldy particles into the waste channel. S13. Peanut kernels that have been initially screened by the near-infrared hyperspectral imaging system enter a manual sorting table equipped with a 450 nm to 550 nm dual-band LED light source. The surface illuminance of the sorting table conveyor belt is maintained at 8000 lux to 12000 lux, and the conveying speed is 0.3 m / s to 0.5 m / s. The sorting operator wears 500 nm to 600 nm band-pass filter glasses for secondary sorting. The sorting operation table is equipped with a spring buffer device. Removed particles enter a sealed recycling container through a slide with an inclination angle of 30° to 35°. The hyperspectral imaging system has a spatial resolution of 0.2 mm / pixel to 0.3 mm / pixel and a spectral resolution of ≤8 nm. After each batch of tests is completed, whiteboard calibration is automatically performed with a calibration interval of ≤2 h.

[0010] Preferably, step S52 is followed by the following steps: Filtered peanut oil is fed into a closed plate heat exchanger, with the oil inlet temperature controlled at 28℃~32℃. The heat exchanger uses a propylene glycol solution at -5℃~0℃ as the refrigerant, causing the oil temperature to drop to 18℃~22℃ within 15 seconds. The cooled peanut oil is then fed into a molecular sieve dehydration tower via a nitrogen-sealed pipeline for dehydration. The molecular sieve is type 3A zeolite, with a packing height of 2.5 to 3 times the tower diameter. The oil flows through the bed at a linear velocity of 0.6 m / h~0.8 m / h, and the nitrogen pressure inside the tower is maintained at 0.12 MPa~0.15 MPa. The dehydrated peanut oil then enters a vacuum degassing tank for degassing to obtain degassed oil. The absolute pressure inside the tank is controlled at 8 kPa~12 kPa. The peanut oil flows down the surface of stainless steel corrugated packing in a film-like form, with a specific surface area of ​​350 m². 2 / m 3 ~400 m 2 / m 3 The degassing time is 25 min to 35 min. The degassed oil is tested by an online dissolved oxygen meter. When the dissolved oxygen content is ≤0.8 mg / L, it is transported to the storage tank through a double-layer 316L stainless steel sleeve. Nitrogen gas is introduced into the sleeve jacket and maintained at a flow rate of 1.5 m / s to 2 m / s to form a gas barrier. The storage tank is equipped with a conical bottom agitator with a blade tip linear velocity ≤0.3 m / s. The breather valve on the top of the tank is equipped with a two-stage silica gel-activated carbon composite filter. The inner wall of the storage tank is coated with a parylene coating with a thickness of 8 μm to 12 μm.

[0011] Preferably, step 3) specifically includes: S31. The apex angle of the triangular tooth pattern is 55°~65°, the tooth height is 0.5 mm~0.8 mm, the tooth spacing is 2 mm~3 mm, and the tooth pattern is arranged obliquely at 45° along the axial direction of the roller surface. The center line distance between two adjacent rows of tooth patterns is 4 mm~6 mm. The feed inlet of the roller crusher is equipped with a double spiral guide plate with a spiral pitch of 50 mm~80 mm, a gap between the end of the guide plate and the roller surface of 1.5 mm~2 mm, and a guide plate rotation speed of 10 r / min~15 r / min. The fine powder rate of the crushed particles is monitored online by a laser particle size analyzer. When the proportion of particles with a diameter <0.5 mm exceeds 5%, an alarm is triggered and the gap of the double spiral guide plate is adjusted to 1.8 mm~2.2 mm simultaneously. S32. During the crushing process, the roller surface temperature is monitored in real time. When the temperature exceeds 40℃, the built-in spray device is activated to spray atomized water droplets onto the roller surface. The water droplet size is 50 μm to 80 μm, and the spray flow rate is 20 mL / min to 30 mL / min. The spray water is food-grade reverse osmosis pure water. Due to the extremely small spray volume (the water volume sprayed per unit time is only 0.02% to 0.05% of the crushed material), and the roller surface temperature in the crushing zone is about 40℃, most of the water evaporates rapidly after contacting the roller surface. The small amount of water absorbed by the material will be discharged with the oil vapor during the subsequent pressing and heating process, and will not have a significant impact on the final moisture content of the peanut kernels. S33. The crushed peanut particles are graded by a double-layer vibrating screen. The upper screen mesh size is 3.5 mm, the lower screen mesh size is 2 mm, the screen surface inclination angle is adjusted to 12°~15°, and the screen body amplitude is set to 3 mm~5 mm. Among them, the drive motor of the roller crusher is equipped with a dynamic load compensation device. When the current fluctuation exceeds the rated value ±15%, the roller spacing is automatically adjusted to compensate by 0.1 mm to 0.3 mm, and the crushing energy consumption is controlled at 0.8 kWh / t to 1.2 kWh / t.

[0012] Preferably, in step 4), the circulating water system adopts a dual-loop circulation. The first loop is connected to the compression section jacket of the pressing chamber, with a water temperature of 45℃~48℃ and a flow rate of 1.5 m / s~2 m / s. The second loop is connected to the cake outlet jacket, with a water temperature of 48℃~50℃ and a flow rate of 0.8m / s~1 m / s. The temperature difference between the two loops is controlled by a proportional valve to be ≤3℃.

[0013] Preferably, in step 4), nitrogen injection is provided at three inlet points: the first inlet point is at the end of the feeding section, with an injection volume of 40% to 50%; the second inlet point is in the middle of the compression section, with an injection volume of 30% to 40%; and the third inlet point is at the front of the cake discharge section, with an injection volume of 10% to 20%. The injection pressure is 0.1 MPa to 0.3 MPa higher than the real-time pressure inside the pressing chamber at each injection point.

[0014] Preferably, step 4) further includes: real-time acquisition of the pressing chamber pressure curve during the pressing process; when the pressure fluctuation exceeds the set value ±5 MPa, the taper of the cake section is dynamically adjusted by a hydraulic servo mechanism at an adjustment rate of 0.5 mm / min to 1 mm / min; the screw shaft is equipped with a torque compensation device; when the torque value reaches 1800 N·m to 2200 N·m, the rotation speed is automatically reduced by 1 r / min to 2 r / min to maintain the power stable at 22 kW to 25 kW; the residual oil rate of the pressed cake is detected by a near-infrared online analyzer; when the residual oil rate is >8%, an alarm is triggered and the pressure at the end of the compression section is simultaneously increased by 3 MPa to 5 MPa for a duration of 30 s to 60 s; an oxygen content sensor is installed at the pressing chamber outlet; when the detected value is >0.6%, the total nitrogen flow rate is increased by 20% to 30% until the oxygen content recovers to 0.3% to 0.5%.

[0015] The present invention has at least the following beneficial effects: 1. The present invention provides a method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing. By combining near-infrared spectroscopy detection with manual sorting, the method significantly improves the screening accuracy of moldy particles and reduces the risk of aflatoxin residue. Gradient baking and citric acid soaking work synergistically to effectively inhibit lipoxygenase activity and reduce the generation of harmful substances. A triangular toothed roller crusher combined with dynamic roller spacing adjustment optimizes particle size distribution and avoids excessive fine powder causing oil passage blockage. Segmented temperature-controlled pressing combined with nitrogen injection inhibits oil oxidation reaction, retains natural flavor components, and reduces acid value and peroxide value. Centrifugation and ceramic membrane filtration are used together to improve impurity removal efficiency and ensure the transparency and stability of the finished oil.

[0016] 2. The present invention's method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing utilizes a centrifugal separator with inclined baffles and reinforcing ribs to enhance centrifugal field stability and improve the separation efficiency of suspended impurities. A 0.1–0.5 μm ceramic membrane filtration precisely traps colloidal particles, reducing turbidity in the finished oil. Coordinated control of operating pressure and membrane flow rate prevents membrane fouling and clogging, extending the service life of the filtration device. Preheating before centrifugation reduces oil viscosity, increases impurity settling speed, and shortens the separation cycle. Gradient heating combined with transverse airflow distribution achieves uniform dehydration of peanut kernels, preventing localized charring. Ultrasonic-assisted citric acid soaking enhances penetration, thoroughly removing surface contaminants and microorganisms. Three-channel countercurrent cleaning combined with ozone treatment effectively eliminates residual solvents and microbial contamination. The finned heat dissipation tube design of the conveyor-type dryer accelerates moisture evaporation, precisely controlling the final moisture content and preventing subsequent breakage and adhesion.

[0017] 3. The present invention's method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing enhances centrifugal separation efficiency and reduces oil entrainment loss through optimized drum baffle tilt angle; real-time conductivity monitoring accurately determines the separation endpoint, avoiding excessive centrifugation that leads to oil emulsification; a pressure compensation device maintains stable drum pressure, ensuring smooth system operation during continuous slag discharge; counter-current cooling via a serpentine tube rapidly reduces oil temperature, inhibiting the degradation of heat-sensitive components and improving oil quality; a near-infrared hyperspectral imaging system enables non-destructive internal mold detection, compensating for blind spots in manual sorting; a combination of dual-band LED light source and filter glasses improves the accuracy of surface defect identification; a spring buffer device absorbs slight vibrations from the sorting table conveyor belt movement, reducing misoperations caused by sorting operator hand tremors and improving sorting accuracy; a sealed recycling channel prevents cross-contamination; and automatic whiteboard calibration ensures the reliability of spectral data and guarantees long-term detection stability.

[0018] 4. The present invention's method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing utilizes a plate heat exchanger for rapid cooling to inhibit the initial stage of oil oxidation; a 3A molecular sieve dehydration tower precisely removes trace amounts of moisture, preventing hydrolysis and rancidity; a vacuum degassing process efficiently removes dissolved oxygen, extending shelf life; a nitrogen barrier conveying system and a storage tank coating design provide dual protection, blocking oxidation and metal ion catalytic reactions during storage; a triangular toothed pattern enhances crushing shear force, reducing fine powder generation; a double-helix guide plate dynamically adjusts the uniformity of feeding, preventing localized overload on the roller surface; a spray cooling device controls the roller temperature in real time, preventing pre-oxidation of the oil due to high temperatures; and a vibrating screen for precise separation of qualified particles, reducing fluctuations in subsequent processes.

[0019] 5. The present invention's method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing utilizes a dual-loop circulating water system for zoned temperature control, ensuring that the temperature gradient in the pressing chamber meets the pressing kinetics requirements; proportional valve linkage controls the temperature difference to ≤3℃, avoiding local overheating that could lead to oil denaturation; high-temperature water flow in the compression section accelerates oilseed plasticization, while low-temperature water flow in the cake discharge section solidifies the cake structure, synergistically improving oil yield; multi-stage nitrogen injection point distribution matches the pressing chamber pressure gradient, forming a continuous inert gas barrier; dynamic adjustment of injection pressure inhibits oxygen penetration, blocking free radical chain reactions; segmented allocation of air intake optimizes gas utilization and reduces nitrogen consumption costs.

[0020] 6. The present invention's method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing uses a hydraulic servo mechanism to adjust the taper of the cake section in real time, maintaining a stable pressing chamber pressure curve and reducing mechanical impact losses; a torque compensation device dynamically balances load fluctuations, ensuring long-term equipment reliability; near-infrared online residual oil rate detection is linked to pressure regulation, enabling accurate determination of the pressing endpoint; and an oxygen content sensor provides feedback to adjust nitrogen flow, constructing a closed-loop antioxidant control system.

[0021] Other advantages, objectives and features of the present invention will be apparent in part from the following description, and in part from the understanding of those skilled in the art through study and practice of the invention. Detailed Implementation

[0022] The present invention will now be described in further detail so that those skilled in the art can implement it based on the description.

[0023] It should be understood that terms such as “having,” “comprising,” and “including” as used herein do not exclude the presence or addition of one or more other elements or combinations thereof.

[0024] Example 1 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing includes the following steps: Raw materials: Commercially available peanuts in shells (origin: Shandong), 1000 kg of peanut kernels after shelling.

[0025] The main equipment used in this embodiment includes: a near-infrared hyperspectral imaging system (Jiangsu Tianrui Instruments, Hyperion 3000); a pneumatic nozzle (Festo, VPPM series); a dual-band LED light source (Osram, Duris S5); bandpass filter glasses (HOYA, B-390); a spring buffer device (MISUMI standard parts); a whiteboard (Labsphere, Spectralon); a hot air circulating oven (BINDER, ED series); an ultrasonic generator (Binersin, B3500); a water pump (Grundfos, CR series); an ozone generator (Mitsubishi Electric, ODO-3A); a vibrating screen (Lowell, RW-SF); a tracked dryer (Bühler, CTRA series); a near-infrared moisture meter (PerkinElmer, DA 7250); a double-roll crusher (Bühler, MDDR); and a laser particle size analyzer (Malvin, Mastersizer). 3000 series); Infrared temperature sensor (Omron, E5CN); Spray nozzle (Spray Systems, 1 / 4J series); PLC control system (Siemens, S7-1200); Screw press (Andritz, SPX series); High temperature pump (Wilo, MHI series); Proportional valve (Siemens, SKD series); Nitrogen generator (Linde Group); Oxygen content sensor (Mettler-Toledo, InPro 6850i); Pressure sensor (WIKA, S-10 and A-10); Hydraulic servo mechanism (Bosch Rexroth, SYDFE series); Frequency converter (ABB, ACS880); Near-infrared online analyzer (Brook, MATRIX-F); Thermostatic preheater and plate heat exchanger (Alfa Laval, TS and M6 series); Disc centrifuge (GEA Westfalia Separator, SA1); Conductivity meter (Mettler-Toledo, InPro... The following equipment is included: 7100 type; serpentine tube cooler (SPX FLOW plate heat exchanger); ceramic membrane filter (Pall Corporation, Membralox X19 type); flow meter (Klöhn, OPTISWIRL 4200 type); molecular sieve (UOP, 3A type zeolite); stainless steel corrugated packing (Kokgerridge, Flexipac HC type); vacuum pump (Busch, BUSCH R5 type); parylene coating (SCS, Parylene C type); composite filter (Pall, Ultipor V Grade type); stirrer (SPX FLOW Lightnin, L series). All of the above equipment are commercially available products and can be directly selected and assembled for use according to the parameter range disclosed in this invention.

[0026] 1) Near-infrared hyperspectral mold identification: A near-infrared hyperspectral imaging system is used, with a wavelength range of 900–1700 nm, a spatial resolution of 0.25 mm / pixel, and a spectral resolution ≤8 nm. The linear scanning speed is 250 kernels / second, and each peanut kernel undergoes three pulse scans at 20 ms intervals. The spectral data is processed by a partial least squares discriminant analysis model (trained based on 3500 sets of aflatoxin-contaminated samples) to output the mold probability value in real time. When the mold probability value of a single peanut kernel is ≥0.90, a pneumatic nozzle is triggered with a spray pressure of 0.7 MPa and a nozzle diameter of 2.5 mm to blow the moldy particles into the waste channel. The waste channel has an inclination angle of 32°, and the inner wall of the chute is coated with Teflon. The peanut kernels after initial screening enter a manual sorting table equipped with a dual-band LED light source, with an illumination of 10000 lux and a transmission speed of 0.4 m / s. Sorting operators wear 550 nm bandpass filter glasses for secondary sorting. The sorting platform is equipped with a spring-loaded buffer device, and rejected particles enter a sealed recycling container via a 32° inclined chute. Automatic whiteboard calibration is performed after each batch is tested, with a calibration interval of 1.5 hours. In this embodiment, the mold probability threshold is set to ≥0.90. To verify the rationality of this threshold, the inventors conducted comparative tests at four thresholds: 0.80, 0.85, 0.90, and 0.95 (using the same sorting system as in Example 1). The results showed that at a threshold of 0.80, the false rejection rate for healthy particles reached 3.2%; at 0.85, the false rejection rate was 1.8% but the missed detection rate was 1.5%; at 0.90, the false rejection rate was 0.8% and the missed detection rate was 0.7%; and at 0.95, the missed detection rate increased to 2.1%. Considering all factors, ≥0.90 is the optimal choice, ensuring a low missed detection rate while avoiding excessive rejection of healthy particles.

[0027] 2) Pretreatment: Selected peanut kernels are spread evenly on the metal mesh belt of a hot air circulating oven, with a thickness of 4 cm. The transverse airflow distribution device is turned on to maintain an air velocity of 1.0 m / s. Gradient heating: The temperature is maintained at 55℃ for the first 10 minutes to evaporate surface moisture, then increased to 60℃ for the middle 10 minutes to promote internal moisture diffusion, and finally stabilized at 65℃ for the last 10 minutes to complete dehydration. The total roasting time is 30 minutes. The roasted peanut kernels are immediately transferred to a soaking tank with a citric acid solution of 0.5% mass fraction, pH 2.5, and a liquid-to-material ratio of 6.5 L:1 kg. The soaking tank is equipped with an ultrasonic generator, applying mechanical vibration at a frequency of 35 kHz with an ultrasonic power density of 0.4 W / cm³. 2 Soak for 15 minutes. After soaking, feed the peanuts into a three-channel countercurrent washing machine: the first channel uses 42℃ warm water for rinsing at a flow rate of 1.8 m / s. 3 / h; the second channel is purged with cold water containing 2.0 mg / L ozone; the third channel is purged with compressed air (pressure 0.5 MPa) to remove residual moisture from the surface. After washing, the peanut kernels are drained through a vibrating screen (amplitude 4 mm, draining time 6 min) and then fed into a tracked dryer with finned heat dissipation pipes. The drying temperature is 38℃ and the track conveying speed is 0.65 m / min, until the moisture content of the peanut kernels drops to 7% (monitored online by a near-infrared moisture meter).

[0028] 3) Directional Crushing and Particle Control: A double-roll crusher is used. The roll surface is decorated with triangular teeth with a 60° apex angle, a tooth height of 0.65 mm, and a tooth spacing of 2.5 mm. The teeth are arranged obliquely at 45° along the axial direction of the roll surface, with a 5 mm center-to-center distance between adjacent rows of teeth. The roll spacing is 1.0 mm, and crushing time is 25 s to a particle size of 2-3 mm. A double-spiral guide plate is installed at the feed inlet with a spiral pitch of 65 mm and a 1.8 mm gap between the end and the roll surface. The rotation speed is 12 r / min. A laser particle size analyzer is used to monitor the fine powder ratio online. When the proportion of particles <0.5 mm exceeds 5%, an alarm is triggered, and the gap between the double-spiral guide plate is adjusted to 2.0 mm simultaneously. The roll surface temperature is monitored in real time during crushing (infrared temperature sensor). When the temperature exceeds 40℃, the built-in spray device is activated to spray atomized water droplets onto the roll surface. The water droplet diameter is 65 μm, and the spray flow rate is 25 mL / min. The cooling water is reverse osmosis pure water (conductivity ≤10 μS / cm). The crushed peanut kernels are graded using a double-layer vibrating screen. The upper screen has a mesh size of 3.5 mm, the lower screen has a mesh size of 2 mm, the screen inclination angle is 13°, and the screen body amplitude is 4 mm. The drive motor is equipped with a dynamic load compensation device (PLC control system). When the current fluctuation exceeds the rated value ±15%, the roller spacing is automatically adjusted to compensate by 0.2 mm. The crushing energy consumption is controlled at 1.0 kWh / t.

[0029] 4) Low-oxygen, low-temperature pressing: A screw press with a length-to-diameter ratio of 9:1 and a screw shaft speed of 16 r / min is used. The pressing chamber is divided into three sections: the feed section has a screw lead of 135 mm, the compression section has a screw lead decreasing to 70 mm, and the cake discharge section has a fixed screw lead of 45 mm. During the pressing process, circulating water is introduced into the pressing chamber jacket to control the pressing temperature at 52℃ and maintain the pressing chamber pressure at 70 MPa. The circulating water system adopts a dual-loop circulation: the first loop connects to the compression section jacket of the pressing chamber, with a water temperature of 46.5℃ and a flow rate of 1.8 m / s; the second loop connects to the cake discharge section jacket, with a water temperature of 49℃ and a flow rate of 0.9 m / s; the two loops are linked by a proportional valve to control the temperature difference to ≤3℃. Nitrogen injection is implemented with three inlet points: the first inlet is at the end of the feeding section, accounting for 45% of the injection volume; the second inlet is in the middle of the compression section, accounting for 35% of the injection volume; and the third inlet is at the front of the cake discharge section, accounting for 20% of the injection volume. The injection pressure is 0.2 MPa higher than the internal pressure of the pressing chamber at each injection point, and the nitrogen purity is ≥99.5% (nitrogen generator). An oxygen content sensor is installed at the pressing chamber outlet to adjust the total nitrogen flow rate in real time, keeping the oxygen content in the pressing chamber controlled at 0.4%. During the pressing process, the pressing chamber pressure curve is collected in real time (pressure sensor). When the pressure fluctuation exceeds the set value ±5 MPa, the taper of the cake discharge section is dynamically adjusted by a hydraulic servo mechanism at an adjustment rate of 0.8 mm / min. The screw press spindle is equipped with a torque compensation device (frequency converter). When the torque value reaches 2000 N·m, the speed is automatically reduced by 1.5 r / min to maintain a stable power of 23.5 kW. The residual oil content of the pressed cake is detected by a near-infrared online analyzer. When the residual oil content is >8%, an alarm is triggered and the pressure at the end of the compression section is increased by 4 MPa simultaneously for 45 seconds. An oxygen content sensor is installed at the outlet of the pressing chamber. When the detected value is >0.6%, the total nitrogen flow rate is increased by 25% until the oxygen content returns to 0.3%–0.5%.

[0030] 5) Membrane filtration stabilization: Crude peanut oil is combined and heated to 38°C using a constant-temperature preheater. After pressure stabilization in a buffer tank, it is fed into a disc centrifuge at a flow rate of 1.0 L / min. Suspended impurities are separated at 3500 rpm for 18 minutes. The centrifuge drum has a length-to-diameter ratio of 4.5:1. Six sets of vertical baffles are evenly distributed circumferentially on the inner wall of the drum. The height of the baffles is 40% of the height of the drum, and the thickness of the baffles is 4 mm. The top of the baffles is inclined at 18° towards the direction of drum rotation. A 9 mm wide reinforcing rib is installed at the weld between the baffles and the inner wall of the drum.

[0031] During centrifugation, a conductivity meter was used to monitor the changes in the conductivity of the oil phase in real time. The initial conductivity of the preheated peanut oil before entering the centrifuge was used as the baseline. Separation was terminated when the real-time conductivity decreased by less than 50 μS / cm from the initial value for 3 consecutive minutes. This slowing decrease indicated that the removal of conductive impurities (mainly water and free fatty acids) in the oil phase had reached equilibrium. Further extending the centrifugation time would have limited effect on improving oil purity and might increase the risk of emulsification. The separated heavy phase impurities were discharged through the slag discharge port at the bottom of the drum at 18-minute intervals. During slag discharge, a pressure compensation device (pneumatic nozzle) was used to maintain pressure fluctuations within the drum ≤0.05 MPa. Before being introduced into the temporary storage tank through the oil outlet pipe, the light phase oil was cooled to 28°C using a serpentine tube cooler with a cooling water flow rate of 2.5 m / s, and the oil flow contacted the cooling water counter-currently. The centrifuged peanut oil was then filtered through a ceramic membrane filter. The ceramic membrane filter is a multi-channel filter made of alumina material with a membrane pore size of 0.3 μm, an operating pressure of 0.3 MPa, and a membrane surface flow velocity of 1.2 m / s. Before filtration, a protective layer is formed by pre-coating with diatomaceous earth (12 μm particle size), with a pre-coating amount of 0.65 kg / m³. 2 The filtration system is equipped with a pressure sensor and flow meter to adjust the transmembrane pressure difference to ≤0.15 MPa in real time. Filtered peanut oil is fed into a closed plate heat exchanger, with the oil inlet temperature controlled at 30℃. The heat exchanger uses a -3℃ propylene glycol solution (30% volume concentration) as the refrigerant, causing the oil temperature to drop to 20℃ within 15 seconds. The cooled peanut oil is then fed into a molecular sieve dehydration tower via a nitrogen-sealed pipeline for dehydration. The molecular sieve is type 3A zeolite (2.0 mm particle size), with a packing height 2.8 times the tower diameter. The oil flows through the bed at a linear velocity of 0.7 m / h, and the nitrogen pressure inside the tower is maintained at 0.13 MPa. The dehydrated peanut oil then enters a vacuum degassing tank for degassing. The absolute pressure inside the tank is controlled at 10 kPa. The peanut oil flows down the surface of stainless steel corrugated packing in a film-like form. The packing has a specific surface area of ​​375 m². 2 / m 3 The degassing time was 30 min. The degassed oil was tested using an online dissolved oxygen meter (oxygen content sensor). When the dissolved oxygen content was ≤0.8 mg / L, it was transported to the storage tank through a double-layered 316L stainless steel sleeve. Nitrogen gas was introduced into the sleeve jacket and maintained at a flow rate of 1.8 m / s to form a gas barrier. The storage tank was equipped with a conical-bottom agitator with a blade tip linear velocity ≤0.3 m / s. A two-stage silica gel-activated carbon composite filter was installed at the tank top breather valve. The inner wall of the storage tank was coated with a 10 μm thick poly(p-xylene) coating. This yielded virgin peanut oil.

[0032] The goal of centrifugal separation is to remove suspended impurities, colloidal particles, and residual moisture from crude oil. Among these impurities, moisture and free fatty acids have a certain degree of conductivity, and their content changes can be reflected in fluctuations in the conductivity of the oil phase. While suspended solids do not directly contribute to conductivity, they usually form complex systems with moisture and colloids. As centrifugation proceeds, the removal of moisture and colloids occurs simultaneously with the sedimentation of solids; therefore, the decreasing trend of conductivity can indirectly characterize the overall purification process. To ensure the reliability of endpoint determination, an online turbidity meter (detection wavelength 860 nm) is installed on the oil outlet pipeline while monitoring conductivity. When the turbidity value remains below 5 NTU for 3 minutes and the conductivity decrease simultaneously meets the above conditions, both the turbidity value and the conductivity value are used to confirm the separation endpoint, thus avoiding misjudgment through double verification.

[0033] Example 2 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing, which differs from Example 1 in that: in step 4), maintaining the pressing chamber pressure at 70 MPa is replaced by maintaining the pressing chamber pressure at 85 MPa.

[0034] Example 3 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing, which differs from Example 1 in that: in step 4), maintaining the pressing chamber pressure at 70 MPa is replaced by maintaining the pressing chamber pressure at 50 MPa.

[0035] Example 4 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing, which differs from Example 1 in that: 1) Near-infrared hyperspectral mold identification: A near-infrared hyperspectral imaging system is used, with a wavelength range of 900–1700 nm, a spatial resolution of 0.2 mm / pixel, and a spectral resolution ≤8 nm. The line scan speed is 200 kernels / second, and each peanut kernel undergoes three pulse scans at 20 ms intervals. The spectral data is processed by a partial least squares discriminant analysis model (trained based on 3500 sets of aflatoxin-contaminated samples) to output the mold probability value in real time. When the mold probability value of a single peanut kernel is ≥0.90, a pneumatic nozzle is triggered with a spray pressure of 0.6 MPa and a nozzle diameter of 2.5 mm to blow the moldy particles into the waste channel. The waste channel has an inclination angle of 32°, and the inner wall of the chute is coated with Teflon. The peanut kernels after initial screening enter a manual sorting table equipped with a dual-band LED light source, with an illumination of 8000 lux and a transmission speed of 0.3 m / s. Sorting operators wear 550 nm bandpass filter glasses for secondary sorting. The sorting table is equipped with a spring buffer device, and rejected particles enter a sealed recycling container through a 32° inclined chute. After each batch is tested, a whiteboard calibration is automatically performed, with a calibration interval of 2 hours.

[0036] Example 5 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing, which differs from Example 1 in that: 2) Pretreatment: Selected peanut kernels are spread evenly on the metal mesh belt of a hot air circulating oven, with a thickness of 3 cm. The transverse airflow distribution device is turned on to maintain an air velocity of 0.8 m / s. Gradient heating: The temperature is maintained at 55℃ for the first 10 minutes to evaporate surface moisture, then increased to 60℃ for the middle 10 minutes to promote internal moisture diffusion, and finally stabilized at 65℃ for the last 10 minutes to complete dehydration. The total roasting time is 30 minutes. The roasted peanut kernels are immediately transferred to a soaking tank with a citric acid solution of 0.5% mass fraction, pH 2.3, and a liquid-to-material ratio of 5 L:1 kg. The soaking tank is equipped with an ultrasonic generator, applying mechanical vibration at a frequency of 28 kHz with an ultrasonic power density of 0.3 W / cm³. 2 Soak for 10 minutes. After soaking, feed the peanuts into a three-channel countercurrent washing machine: the first channel uses 40℃ warm water for rinsing at a flow rate of 1.8 m / s. 3 The second channel is purged with cold water containing 1.5 mg / L ozone; the third channel is purged with compressed air (0.5 MPa pressure) to remove residual moisture from the surface. After washing, the peanut kernels are drained through a vibrating screen (4 mm amplitude, 6 min draining time) and then fed into a tracked dryer with finned heat dissipation tubes. The drying temperature is 35℃ and the track conveyor speed is 0.5 m / min, until the moisture content of the peanut kernels drops to 6% (monitored online by a near-infrared moisture meter).

[0037] Example 6 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing, which differs from Example 1 in that: 5) Membrane filtration stabilization: Crude peanut oil is combined and heated to 35°C using a constant-temperature preheater. After pressure stabilization in a buffer tank, it is fed into a disc centrifuge at a flow rate of 0.8 L / min. Suspended impurities are separated at 3000 rpm for 18 minutes. The centrifuge drum has a length-to-diameter ratio of 4:1. Six sets of vertical baffles are evenly distributed circumferentially on the inner wall of the drum. The height of the baffles is 1 / 3 of the drum height, and the baffle thickness is 3 mm. The top of the baffles is inclined at 15° towards the direction of drum rotation. Reinforcing ribs with a width of 8 mm are installed at the weld between the baffles and the inner wall of the drum.

[0038] During centrifugation, a conductivity meter was used to monitor the changes in the conductivity of the oil phase in real time. The initial conductivity of the preheated peanut oil before entering the centrifuge was used as the baseline. Separation was terminated when the real-time conductivity decreased by less than 50 μS / cm from the initial value for 3 consecutive minutes. The separated heavy phase impurities were discharged through the slag discharge port at the bottom of the drum at 15-minute intervals. During slag discharge, a pressure compensation device (pneumatic nozzle) was used to maintain the pressure fluctuation within the drum within ≤0.05 MPa. Before being introduced into the temporary storage tank through the oil outlet pipe, the light phase oil was cooled to 25°C using a serpentine tube cooler with a cooling water flow rate of 2 m / s within the serpentine tube, and the oil flow and cooling water were in counter-current contact. The centrifuged peanut oil was then filtered through a ceramic membrane filter. The ceramic membrane filter was a multi-channel filter made of alumina material with a membrane pore size of 0.3 μm, an operating pressure of 0.3 MPa, and a membrane surface velocity of 1.2 m / s. Before filtration, a protective layer is formed by pre-coating with diatomaceous earth (particle size 12 μm), with a pre-coating amount of 0.65 kg / m³. 2 The filtration system is equipped with a pressure sensor and flow meter to adjust the transmembrane pressure difference to ≤0.15 MPa in real time. Filtered peanut oil is fed into a closed plate heat exchanger, with the oil inlet temperature controlled at 30℃. The heat exchanger uses a -3℃ propylene glycol solution (30% volume concentration) as the refrigerant, causing the oil temperature to drop to 20℃ within 15 seconds. The cooled peanut oil is then fed into a molecular sieve dehydration tower via a nitrogen-sealed pipeline for dehydration. The molecular sieve is type 3A zeolite (2.0 mm particle size), with a packing height 2.8 times the tower diameter. The oil flows through the bed at a linear velocity of 0.7 m / h, and the nitrogen pressure inside the tower is maintained at 0.13 MPa. The dehydrated peanut oil then enters a vacuum degassing tank for degassing. The absolute pressure inside the tank is controlled at 10 kPa. The peanut oil flows down the surface of stainless steel corrugated packing in a film-like form. The packing has a specific surface area of ​​375 m². 2 / m 3 The degassing time was 30 min. The degassed oil was tested using an online dissolved oxygen meter (oxygen content sensor). When the dissolved oxygen content was ≤0.8 mg / L, it was transported to the storage tank through a double-layered 316L stainless steel sleeve. Nitrogen gas was introduced into the sleeve jacket and maintained at a flow rate of 1.8 m / s to form a gas barrier. The storage tank was equipped with a conical-bottom agitator with a blade tip linear velocity ≤0.3 m / s. A two-stage silica gel-activated carbon composite filter was installed at the tank top breather valve. The inner wall of the storage tank was coated with a 10 μm thick poly(p-xylene) coating. This yielded virgin peanut oil.

[0039] Example 7 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing, which differs from Example 1 in that: 5) Membrane filtration stabilization: Crude peanut oil is combined and heated to 38°C using a constant-temperature preheater. After pressure stabilization in a buffer tank, it is fed into a disc centrifuge at a flow rate of 1.0 L / min. Suspended impurities are separated at 3500 rpm for 18 minutes. The centrifuge drum has a length-to-diameter ratio of 4.5:1. Six sets of vertical baffles are evenly distributed circumferentially on the inner wall of the drum. The height of the baffles is 40% of the height of the drum, and the thickness of the baffles is 4 mm. The top of the baffles is inclined at 18° towards the direction of drum rotation. A 9 mm wide reinforcing rib is installed at the weld between the baffles and the inner wall of the drum.

[0040] During centrifugation, a conductivity meter was used to monitor the changes in the conductivity of the oil phase in real time. The initial conductivity of the preheated peanut oil before entering the centrifuge was used as the baseline. Separation was terminated when the real-time conductivity decreased by less than 50 μS / cm from the initial value for 3 consecutive minutes. The separated heavy phase impurities were discharged through the slag discharge port at the bottom of the drum at 18-minute intervals. During slag discharge, a pressure compensation device (pneumatic nozzle) was used to maintain the pressure fluctuation within the drum within ≤0.05 MPa. Before being introduced into the temporary storage tank through the oil outlet pipe, the light phase oil was cooled to 28℃ using a serpentine tube cooler. The cooling water flow rate inside the serpentine tube was 2.5 m / s, with the oil flow contacting the cooling water counter-currently. The centrifuged peanut oil was then filtered through a ceramic membrane filter. The ceramic membrane filter was a multi-channel alumina filter with a membrane pore size of 0.3 μm, an operating pressure of 0.3 MPa, and a membrane surface velocity of 1.2 m / s. Before filtration, a protective layer is formed by pre-coating with diatomaceous earth (particle size 12 μm), with a pre-coating amount of 0.65 kg / m³. 2 The filtration system is equipped with a pressure sensor and flow meter to adjust the transmembrane pressure difference to ≤0.15 MPa in real time. Filtered peanut oil is fed into a closed plate heat exchanger, with the oil inlet temperature controlled at 28℃. The heat exchanger uses a -5℃ propylene glycol solution (30% volume concentration) as the refrigerant, causing the oil temperature to drop to 18℃ within 15 seconds. The cooled peanut oil is then fed into a molecular sieve dehydration tower via a nitrogen-sealed pipeline for dehydration. The molecular sieve is type 3A zeolite (2.0 mm particle size), with a packing height 2.5 times the tower diameter. The oil flows through the bed at a linear velocity of 0.6 m / h, and the nitrogen pressure inside the tower is maintained at 0.12 MPa. The dehydrated peanut oil then enters a vacuum degassing tank for degassing. The absolute pressure inside the tank is controlled at 8 kPa. The peanut oil flows down the surface of stainless steel corrugated packing in a film-like form. The packing has a specific surface area of ​​350 m². 2 / m 3The degassing time was 25 minutes. The degassed oil was tested using an online dissolved oxygen meter (oxygen content sensor). When the dissolved oxygen content was ≤0.8 mg / L, it was transported to the storage tank through a double-layered 316L stainless steel sleeve. Nitrogen gas was introduced into the sleeve jacket and maintained at a flow rate of 1.5 m / s to form a gas barrier. The storage tank was equipped with a conical-bottom agitator with a blade tip linear velocity ≤0.3 m / s. A two-stage silica gel-activated carbon composite filter was installed at the tank top breather valve. The inner wall of the storage tank was coated with a parylene coating with a thickness of 8 μm. This yielded virgin peanut oil.

[0041] Example 8 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing, which differs from Example 1 in that: 3) Directional Crushing and Particle Control: A double-roll crusher is used. The roll surface is decorated with triangular teeth with a 55° apex angle, a tooth height of 0.5 mm, and a tooth spacing of 2 mm. The teeth are arranged obliquely at 45° along the axial direction of the roll surface, with a center-to-center distance of 4 mm between adjacent rows of teeth. The roll spacing is 1.0 mm, and crushing takes 25 seconds to achieve a particle size of 2-3 mm. A double-spiral guide plate is installed at the feed inlet with a spiral pitch of 50 mm and a 1.5 mm gap between the end and the roll surface. The rotation speed is 10 r / min. A laser particle size analyzer is used to monitor the fine powder ratio online. When the proportion of particles <0.5 mm exceeds 5%, an alarm is triggered, and the gap between the double-spiral guide plate is adjusted to 1.8 mm simultaneously. The roll surface temperature is monitored in real time during crushing (infrared temperature sensor). When the temperature exceeds 40℃, the built-in spray device is activated to spray atomized water droplets onto the roll surface. The water droplet diameter is 50 μm, and the spray flow rate is 20 mL / min. The cooling water is reverse osmosis pure water (conductivity ≤10 μS / cm). The crushed peanut particles are graded by a double-layer vibrating screen. The upper screen has a mesh size of 3.5 mm, the lower screen has a mesh size of 2 mm, the screen inclination angle is 12°, and the screen body amplitude is 3 mm. The drive motor is equipped with a dynamic load compensation device (PLC control system). When the current fluctuation exceeds the rated value ±15%, the roller spacing is automatically adjusted to compensate by 0.1 mm. The crushing energy consumption is controlled at 1.0 kWh / t.

[0042] Example 9 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing, which differs from Example 1 in that: 4) Low-oxygen, low-temperature pressing: A screw press with a length-to-diameter ratio of 9:1 and a screw shaft speed of 16 r / min is used. The pressing chamber is divided into three sections: the screw lead in the feeding section is 135 mm, the screw lead in the compression section decreases to 70 mm, and the screw lead in the cake discharge section is fixed at 45 mm. During pressing, circulating water is introduced into the pressing chamber jacket to control the pressing temperature at 52℃ and maintain the pressing chamber pressure at 70 MPa. The circulating water system adopts a dual-loop circulation: the first loop connects to the compression section jacket of the pressing chamber, with a water temperature of 45℃ and a flow rate of 1.5 m / s; the second loop connects to the cake discharge section jacket, with a water temperature of 48℃ and a flow rate of 0.8 m / s; the two loops are linked by a proportional valve to control the temperature difference ≤3℃. Nitrogen injection is set at three inlet points: the first inlet point is at the end of the feeding section, accounting for 40% of the injection volume; the second inlet point is in the middle of the compression section, accounting for 40% of the injection volume; and the third inlet point is at the front end of the cake discharge section, accounting for 20% of the injection volume. The injection pressure is 0.1 MPa higher than the real-time pressure inside the pressing chamber at the injection point, and the nitrogen purity is ≥99.5% (nitrogen generator). An oxygen content sensor is installed at the pressing chamber outlet to adjust the total nitrogen flow rate in real time, keeping the oxygen content inside the pressing chamber at 0.4%. The pressing chamber pressure curve is collected in real time during the pressing process (pressure sensor). When the pressure fluctuation exceeds the set value ±5 MPa, the taper of the cake section is dynamically adjusted via a hydraulic servo mechanism at an adjustment rate of 0.8 mm / min. The screw press spindle is equipped with a torque compensation device (frequency converter). When the torque value reaches 2000 N·m, the speed is automatically reduced by 1.5 r / min to maintain a stable power of 23.5 kW. The residual oil content of the pressed cake is detected by a near-infrared online analyzer. When the residual oil content is >8%, an alarm is triggered, and the pressure at the end of the compression section is simultaneously increased by 4 MPa for 45 seconds. An oxygen content sensor is installed at the outlet of the pressing chamber. When the detected value is >0.6%, the total nitrogen flow rate is increased by 25% until the oxygen content is restored to 0.3% to 0.5%.

[0043] Example 10 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing, which differs from Example 1 in that: 4) Low-oxygen, low-temperature pressing: A screw press with a length-to-diameter ratio of 9:1 and a screw shaft speed of 16 r / min is used. The pressing chamber is divided into three sections: the feed section has a screw lead of 135 mm, the compression section has a screw lead decreasing to 70 mm, and the cake discharge section has a fixed screw lead of 45 mm. During the pressing process, circulating water is introduced into the pressing chamber jacket to control the pressing temperature at 52℃ and maintain the pressing chamber pressure at 70 MPa. The circulating water system adopts a dual-loop circulation: the first loop connects to the compression section jacket of the pressing chamber, with a water temperature of 46.5℃ and a flow rate of 1.8 m / s; the second loop connects to the cake discharge section jacket, with a water temperature of 49℃ and a flow rate of 0.9 m / s; the two loops are linked by a proportional valve to control the temperature difference to ≤3℃. Nitrogen injection is implemented with three inlet points: the first inlet is at the end of the feeding section, accounting for 45% of the injection volume; the second inlet is in the middle of the compression section, accounting for 35% of the injection volume; and the third inlet is at the front of the cake discharge section, accounting for 20% of the injection volume. The injection pressure is 0.2 MPa higher than the real-time pressure inside the pressing chamber at each injection point, and the nitrogen purity is ≥99.5% (nitrogen generator). An oxygen content sensor is installed at the pressing chamber outlet to adjust the total nitrogen flow rate in real time, keeping the oxygen content inside the pressing chamber at 0.4%. During the pressing process, the pressing chamber pressure curve is collected in real time (pressure sensor). When the pressure fluctuation exceeds the set value ±5 MPa, the taper of the cake discharge section is dynamically adjusted by a hydraulic servo mechanism at an adjustment rate of 0.5 mm / min. The screw press spindle is equipped with a torque compensation device (frequency converter). When the torque value reaches 1800 N·m, the speed is automatically reduced by 1 r / min to maintain a stable power of 22 kW. The residual oil content of the pressed cake is detected by a near-infrared online analyzer. When the residual oil content is >8%, an alarm is triggered and the pressure at the end of the compression section is increased by 3 MPa simultaneously for 30 seconds. An oxygen content sensor is installed at the outlet of the pressing chamber. When the detected value is >0.6%, the total nitrogen flow rate is increased by 20% until the oxygen content recovers to 0.3%–0.5%.

[0044] Comparative Example 1 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing, compared with Example 1, omits the automatic sorting and pneumatic nozzle removal of the near-infrared hyperspectral imaging system in step 1), and relies only on manual visual sorting (same number of people, same working time) to screen mold particles, while the remaining steps (pretreatment, crushing, pressing, membrane filtration) are the same as in Example 1.

[0045] Comparative Example 2 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing, compared with Example 1, omits the citric acid solution soaking and ultrasonic-assisted treatment in step 2), and directly proceeds to the cleaning and drying process after baking, while the remaining steps are the same as in Example 1.

[0046] Comparative Example 3 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing, compared with Example 1, step 3) adopts a conventional smooth roller crusher (without triangular tooth patterns, the same roller spacing of 1.0 mm, and the same crushing time of 25 s), and the remaining steps are the same as in Example 1.

[0047] Comparative Example 4 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing. Compared with Example 1, in step 4), nitrogen injection is changed to only one injection point at the end of the feeding section (with the same total injection volume), without injection points in the middle of the compression section and the front of the cake discharge section. The remaining steps are the same as in Example 1.

[0048] Comparative Example 5 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing, which differs from Example 1 in that nitrogen is not injected in step 4) (pressing is carried out in an open air environment), while the remaining steps are the same as in Example 1.

[0049] Comparative Example 6 A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing, compared with Example 1, step 4) does not use circulating water for temperature control, the pressing temperature naturally rises to 100-120℃ (typical working temperature of traditional screw press), and nitrogen is not injected, the remaining steps are the same as in Example 1.

[0050] Experimental data: (1) Comparison of sorting effects for moldy plants: Detection methods: The false negative rate of moldy particles was determined by testing aflatoxin residue according to GB 5009.22-2016. The residue level ≥5 ppb was used as the criterion for positive mold detection. The proportion of particles with aflatoxin residue ≥5 ppb after sorting was counted out of the total theoretical number of moldy particles. The false rejection rate of healthy particles was determined by retesting rejected particles using the same method. The proportion of particles with aflatoxin residue <5 ppb was counted out of the total number of rejected particles. The sorting speed was determined by counting the number of peanut kernels processed per unit time using an online counting system.

[0051] Test subjects: The mold sorting effect of Example 1, Example 4 (using different combinations of near-infrared sorting parameters) and Comparative Example 1 (pure manual sorting) was tested, and the results are shown in Table 1.

[0052] Table 1. Comparison of mold sorting effects As shown in Table 1, Examples 1 and 4, employing a combination of automatic identification using a near-infrared hyperspectral imaging system and pneumatic nozzle rejection, supplemented by manual secondary sorting, achieved a mold particle false negative rate of only 0.7% and 0.8%, respectively. This represents a reduction of approximately 90%–91% compared to the purely manual sorting method in Comparative Example 1 (false negative rate of 7.8%), with aflatoxin residue levels controlled below 0.9 ppb in both examples. Specifically, Example 1, using a scanning speed of 250 particles / second, a sorting illumination of 10,000 lux, and a spatial resolution of 0.25 mm / pixel, achieved a higher processing speed (approximately 900,000 particles / hour) while maintaining a low false negative rate (0.7%), resulting in the optimal overall sorting efficiency. Example 4, using a scanning speed of 200 particles / second, an illumination of 8,000 lux, and a spatial resolution of 0.2 mm / pixel, had a slightly lower false rejection rate (0.7%), but a slightly lower processing speed than Example 1. The rate of incorrect rejection of healthy particles in both embodiments was controlled below 0.8%, indicating that at a mold probability threshold of ≥0.90, the synergistic combination of automatic sorting and manual secondary sorting can effectively remove moldy particles while avoiding excessive loss of healthy particles.

[0053] (2) Comparison of pretreatment effects Detection methods: Lipoxygenase activity was determined by linoleic acid substrate-ultraviolet spectrophotometry, with the lipoxygenase activity in untreated raw peanut kernels as a 100% control. The enzyme activity inhibition rate of each example and comparative example was calculated. The surface charring rate after roasting was determined by randomly sampling 100 peanut kernels and counting the proportion of particles with obvious charring spots on the surface. The microbial contamination rate was determined by the total number of colonies according to GB 4789.2. The initial acid value of the crude oil after subsequent pressing was tested according to GB / T 5530.

[0054] Test subjects: The pretreatment effects of Example 1 (complete pretreatment process), Example 5 (using different combinations of pretreatment parameters) and Comparative Example 2 (omitting citric acid solution soaking and ultrasonic-assisted treatment) were tested, and the results are shown in Table 2.

[0055] Table 2. Comparison of Pretreatment Effects Table 2 shows that the gradient heating baking combined with citric acid solution immersion and ultrasonic-assisted treatment in Examples 1 and 5 achieved lipoxygenase activity inhibition rates of 93.6% and 91.2%, respectively, which is more than 115% higher than that of Comparative Example 2 (which omitted citric acid immersion and had an inhibition rate of only 42.3%). Regarding the surface charring rate after baking, Examples 1 and 5 were 1.5% and 1.8%, respectively, which is more than 75% lower than that of Comparative Example 2 (7.2%), indicating that gradient heating combined with lateral airflow distribution can effectively avoid charring caused by local overheating. In terms of microbial contamination rate, the combination of citric acid immersion and three-channel countercurrent cleaning (including ozone water treatment) reduced the microbial contamination rate of Examples 1 and 5 to 2.1 × 10⁻⁶. 2 CFU / g and 3.5×10 2 The CFU / g of the solution was significantly better than that of Comparative Example 2 (1.2 × 10³ CFU / g), which omitted citric acid soaking and ultrasonic treatment, with reductions of approximately 82.5% and 70.8%, respectively. This indicates that citric acid soaking and ultrasonic treatment not only directly kill bacteria, but more importantly, they can disrupt the biofilm structure of microorganisms on the surface of peanut kernels, thereby significantly enhancing the penetration and sterilization effect of subsequent ozone water cleaning. Example 1 used pH 2.5, a liquid-to-solid ratio of 6.5 L:1 kg, an ultrasonic frequency of 35 kHz, and a power density of 0.4 W / cm³. 2 The pretreatment conditions of Example 1 were slightly better than those of Example 5 in all indicators (pH 2.3, liquid-to-solid ratio 5 L:1 kg, ultrasonic frequency 28 kHz, power density 0.3 W / cm²), indicating that the combination of pretreatment parameters in Example 1 achieved a better balance in inhibiting enzyme activity, reducing coking rate and removing microorganisms.

[0056] (3) Comparison of fragmentation effects Detection methods: Particle size distribution was detected using a laser particle size analyzer (Malvin Mastersizer 3000), and the proportion of particles with a diameter of 2-3 mm and the proportion of fine powder with a diameter of <0.5 mm were statistically analyzed; crushing energy consumption was recorded by the power monitoring system of the drive motor, which recorded the power consumption per unit processing volume (tons); subsequent pressing oil extraction efficiency was calculated as the ratio of actual oil yield to theoretical oil content (total oil content determined by Soxhlet extraction).

[0057] Test subjects: The crushing effects of Example 1 (complete crushing process), Example 8 (using different combinations of crushing parameters) and Comparative Example 3 (using a smooth roller crusher) were tested, and the results are shown in Table 3.

[0058] Table 3. Comparison of Crushing Effects As shown in Table 3, in Examples 1 and 8, the roller crushers with 45° obliquely staggered triangular teeth on the roller surface achieved a qualified particle size of 2-3 mm of 93.8% and 91.5%, respectively, which is 34%-37% higher than that of the smooth roller crusher in Comparative Example 3 (68.4%). The fine powder rate (particle size <0.5 mm) was 4.3% and 4.8%, respectively, which is 48%-53% lower than that of Comparative Example 3 (9.2%). Example 1 uses tooth profile parameters of 60° apex angle, 0.65 mm tooth height, 2.5 mm tooth spacing, and 5 mm center-line distance between adjacent rows of tooth patterns. Combined with a double-helix guide plate (65 mm pitch, 1.8 mm gap) and online monitoring and feedback adjustment using a laser particle size analyzer, the particle size qualification rate reaches 93.8%, and the fine powder rate is only 4.3%, making it the best crushing effect among the three schemes. Example 8 uses tooth profile parameters of 55° apex angle, 0.5 mm tooth height, and 2 mm tooth spacing. Its effect is slightly inferior to Example 1 but still significantly better than Comparative Example 3. Regarding crushing energy consumption, Examples 1 and 8 are 1.0 kWh / t and 1.1 kWh / t, respectively, which is 31%–38% lower than the 1.6 kWh / t of Comparative Example 3. This is related to the triangular tooth pattern enhancing shear crushing efficiency and reducing repeated crushing. Regarding subsequent pressing oil extraction efficiency, the better particle size distribution (high proportion of 2-3mm and low fine powder) enabled the oil extraction efficiencies of Example 1 and Example 8 to reach 92.7% and 91.3% respectively, which is 5-6 percentage points higher than the 86.1% of Comparative Example 3. This indicates that a suitable crushed particle size and low fine powder ratio are conducive to smooth oil passage and effective oil extraction during the pressing process.

[0059] (4) Comparison of the effects of low-oxygen pressing Detection methods: The oxygen content in the pressing chamber was monitored and recorded in real time by an oxygen content sensor (Mettler-Toledo InPro 6850i) installed at the outlet of the pressing chamber; the acid value of crude oil was tested according to GB / T 5530 method; the peroxide value of crude oil was tested according to GB / T5538 method; the oil oxidation rate was calculated based on the peroxide value of Comparative Example 6 (conventional high temperature nitrogen-free pressing), and the percentage reduction of peroxide value for each scheme was calculated; the nitrogen consumption was recorded by the mass flow meter at the outlet of the nitrogen generator, showing the volume of nitrogen consumed per unit mass of peanut oil product.

[0060] Test subjects: The low-oxygen pressing effects of Example 1 (complete low-oxygen pressing scheme), Example 2 (pressing chamber pressure 85 MPa), Example 3 (pressing chamber pressure 50 MPa), Example 9 (different combinations of temperature control and nitrogen injection parameters), Example 10 (different combinations of real-time pressing control parameters), as well as Comparative Example 4 (single-point nitrogen injection), Comparative Example 5 (no nitrogen protection), and Comparative Example 6 (conventional high-temperature nitrogen-free pressing) were tested. The results are shown in Table 4.

[0061] Table 4. Comparison of Low-Oxygen Pressing Effects Table 4 shows that Examples 1, 2, 3, 9, and 10 all employed a multi-point nitrogen injection scheme with three air inlets (end of the feeding section, middle of the compression section, and front of the cake discharge section), and used a sensor at the pressing chamber outlet to adjust the total nitrogen flow rate in real time. The oxygen content in the pressing chamber of each example was stably controlled within the target range of 0.3%–0.5%. In contrast, Comparative Example 4 used a single-point nitrogen injection scheme, and the oxygen content could only be controlled between 0.8% and 1.2%, indicating that multi-point injection combined with real-time feedback control is key to achieving a deep hypoxic environment. Comparative Examples 5 and 6 operated in an open air environment, and the measured oxygen content in the pressing chamber was approximately 19.5% (close to the composition of ambient air), a stark contrast to the 0.3%–0.5% hypoxic environment of the examples.

[0062] Crude oleic acid value and peroxide value are core indicators for measuring the degree of oil oxidation. The crude oleic acid value of each example was 0.48–0.56 mgKOH / g, significantly lower than 0.78 mgKOH / g in Comparative Example 4, 1.35 mgKOH / g in Comparative Example 5, and 1.80 mgKOH / g in Comparative Example 6. Regarding peroxide value, the values ​​of each example were 1.6–2.0 mmol / kg, significantly lower than 2.9 mmol / kg in Comparative Example 4, 4.6 mmol / kg in Comparative Example 5, and 5.3 mmol / kg in Comparative Example 6. Compared to Comparative Example 6 (conventional high-temperature nitrogen-free pressing), the reduction in oil oxidation rate in each example reached 62%–70%, while the reduction in oxidation rate in Comparative Example 4 (single-point nitrogen injection, oxygen content 0.8%–1.2%) was only 45%, and in Comparative Example 5 (low-temperature control only, no nitrogen protection) it was only 13%. This gradient difference fully demonstrates that reducing the oxygen content from 19.5% to 0.8%–1.2% (Comparative Example 4) achieves a 45% oxidation inhibition effect, while further reducing it to 0.3%–0.5% (in all examples) can increase the oxidation inhibition effect to 62%–70%. This proves that the present invention, through multi-point precise nitrogen injection, controls the oxygen content at an extremely low level of 0.3%–0.5%, which plays a key role in maximizing the inhibition of oil oxidation. Among them, Example 2 (pressing chamber pressure 85 MPa) showed the largest reduction in oxidation rate (70%), which may be due to the higher pressure causing the oil to be extruded faster and shortening the residence time of the material in the pressing chamber; Example 3 (pressing chamber pressure 50 MPa) showed a slightly smaller reduction in oxidation rate (62%), but it was still significantly better than Comparative Example 4 (45%) and Comparative Example 5 (13%).

[0063] Regarding nitrogen consumption, the examples all consumed 8.0–8.8 L / kg oil, a reduction of 23%–30% compared to the single-point nitrogen injection scheme in Comparative Example 4 (11.5 L / kg oil). This demonstrates that multi-point precise nitrogen injection improves nitrogen utilization efficiency and avoids the waste caused by single-point large-volume nitrogen injection. Example 1, with oxygen content stably controlled at 0.3%–0.5%, achieved the best overall performance with a nitrogen consumption of only 8.2 L / kg oil. Differences among the examples (pressing chamber pressure 50–85 MPa, temperature control parameters, nitrogen injection ratio, and real-time control parameters) all effectively maintained oxygen content within the target range and obtained excellent oxidation indicators, indicating that the various process parameter ranges covered by this invention can achieve good low-oxygen pressing results.

[0064] (5) Summary of main product quality indicators for each embodiment Test methods: The acid value of the finished oil is tested according to GB / T 5530; the peroxide value of the finished oil is tested according to GB / T 5538; the aflatoxin residue is tested according to GB 5009.22-2016; and the transmittance of the finished oil is determined at a wavelength of 660 nm according to GB / T5525.

[0065] Test subjects: The main quality indicators of all products in Examples 1 to 10 were tested, and the results are shown in Table 5.

[0066] Table 5. Summary of key product indicators for Examples 1-10 As shown in Table 5, all products from Examples 1 to 10 maintained excellent and stable quality levels in terms of acid value (0.32–0.38 mgKOH / g), peroxide value (1.3–1.7 mmol / kg), aflatoxin residue (≤0.9 ppb), and transmittance (97.5%–97.9%). Among them, Example 2 (pressing chamber pressure 85 MPa) had the lowest peroxide value (1.3 mmol / kg), indicating that higher pressing pressure helps to shorten the material residence time and reduce the chance of oxidation. Example 3 (pressing chamber pressure 50 MPa) had slightly higher acid value and peroxide value than other examples, but still at an excellent level (acid value 0.38 mgKOH / g, peroxide value 1.7 mmol / kg), proving that the present invention can still guarantee good product quality under the lower limit pressure of 50 MPa. Example 7 (optimized post-membrane filtration parameters) had the highest transmittance (97.9%), indicating that its combination of rapid cooling, molecular sieve dehydration and vacuum degassing had the most significant effect on improving the clarity and purity of the oil. The product indicators among the examples were small and all were significantly better than those of the comparative examples, indicating that the core process parameter range defined by independent claim 1 and the different parameter combinations covered by the dependent claims can stably produce high-quality virgin peanut oil, and the scope of protection of the present invention is supported by sufficient experimental data.

[0067] Although the technical solutions of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and embodiments shown and described herein.

Claims

1. A method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing, characterized in that, Includes the following steps: 1) The raw peanut kernels are scanned multiple times using a near-infrared hyperspectral imaging system of 900 nm to 1700 nm to obtain spectral data. The data is then processed by a partial least squares discriminant analysis model to output the mold probability value in real time. When the mold probability value of a single peanut kernel is ≥0.90, it is removed by a pneumatic nozzle and then manually sorted to obtain selected peanut kernels. 2) After roasting the selected peanut kernels at 55-65℃ for 20-40 min, soak them in a 0.4-0.6% citric acid solution for 10-20 min, rinse them, and then air dry them to obtain pre-treated peanut kernels; 3) Input the pre-treated peanut kernels into a roller crusher with a roller spacing of 0.8 mm to 1.2 mm and crush for 20 to 30 seconds until the peanut kernels have a particle size of 2 mm to 3 mm. The roller crusher has staggered triangular teeth on the roller surface with a tooth height of 0.5 mm to 0.8 mm and a tooth spacing of 2 mm to 3 mm. 4) Peanut kernels are fed into a screw press for pressing to obtain crude peanut oil. The screw press has a length-to-diameter ratio of 8:1 to 10:1, a screw shaft speed of 15 r / min to 18 r / min, and a pressing chamber divided into three sections: the feeding section with a screw lead of 120 mm to 150 mm, the compression section with a screw lead of 60 mm to 80 mm, and the cake discharge section with a screw lead of 40 mm to 50 mm. During the pressing process, circulating water is introduced into the pressing chamber jacket to control the pressing temperature at 50℃ to 55℃ and maintain the pressing chamber pressure at 50 MPa to 85 MPa. Multiple nitrogen injection points are set in the pressing chamber during the pressing process to continuously inject nitrogen with a purity of ≥99.5%. The total nitrogen flow rate is adjusted in real time by an oxygen content sensor at the pressing chamber outlet to keep the oxygen content in the pressing chamber at 0.3% to 0.5%. 5) After combining the crude peanut oils and separating them by centrifugation, the crude peanut oils are filtered through a ceramic membrane filter with a pore size of 0.1 μm to 0.5 μm to obtain virgin peanut oil.

2. The method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing as described in claim 1, characterized in that, Step 5) specifically includes: S51. The crude peanut oil is fed into a centrifugal separator and suspended impurities are separated at a speed of 3000 rpm to 4000 rpm for 15 min to 20 min. The inner wall of the centrifugal separator drum is equipped with six sets of vertical baffles, the height of which is 1 / 3 to 1 / 2 of the height of the drum, and the thickness of which is 3 mm to 5 mm. S52. The centrifuged peanut oil is filtered through a ceramic membrane filter with a pore size of 0.1 μm to 0.5 μm to obtain filtered peanut oil; wherein, the operating pressure of the ceramic membrane filter is 0.2 MPa to 0.4 MPa, and the membrane surface flow velocity is 0.8 m / s to 1.5 m / s.

3. The method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing as described in claim 1, characterized in that, Step 2) specifically includes: S21. Spread the selected peanut kernels evenly on the metal mesh belt of the hot air circulating oven, with a thickness of 3 cm to 5 cm. Turn on the transverse airflow distribution device to keep the wind speed at 0.8 m / s to 1.2 m / s. Increase the temperature in a gradient of 55℃ to 65℃: maintain 55℃ for the first 10 minutes to evaporate the surface moisture of the selected peanut kernels, raise the temperature to 60℃ for the middle 10 minutes to promote the diffusion of internal moisture, and stabilize at 65℃ for the last 10 minutes to complete the dehydration. S22. The selected roasted peanut kernels are immediately transferred to a soaking tank. The pH value of the citric acid solution in the soaking tank is 2.3-2.8, and the liquid-to-solid ratio is 5-8 L:1 kg. During the soaking process, an ultrasonic generator is used to apply mechanical vibration at a frequency of 28 kHz-40 kHz, with an ultrasonic power density of 0.3 W / cm³. 2 ~0.5 W / cm 2 ; S23. After soaking, feed the selected peanut kernels into a three-channel countercurrent washing machine. The first channel uses warm water at 40℃~45℃ with a water flow rate of 1.5m. 3 / h~2m 3 / h, the second channel is filled with cold water with an ozone concentration of 1.5 mg / L to 3.0 mg / L, and the third channel is purged with compressed air to remove residual moisture from the surface; S24. After being cleaned and drained by a vibrating screen, the selected peanut kernels are fed into a track dryer with finned heat dissipation tubes. The drying temperature is 35℃~40℃ and the track conveying speed is 0.5 m / min~0.8 m / min until the moisture content of the peanut kernels drops to 6%~8%, thus obtaining pre-treated peanut kernels.

4. The method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing as described in claim 2, characterized in that, Step S51 specifically includes: heating the crude peanut oil to 35℃~40℃ through a constant temperature preheater, and after being pressure-stabilized by a buffer tank, feeding it into a disc centrifuge at a flow rate of 0.8 L / min~1.2 L / min; separating suspended impurities at a speed of 3000 rpm~4000 rpm; the length-to-diameter ratio of the centrifuge drum is 4:1~5:1; six sets of vertical baffles are evenly distributed circumferentially on the inner wall of the drum; the height of the baffles is 1 / 3~1 / 2 of the height of the drum; the thickness of the baffles is 3 mm~5 mm; the top of the baffles is inclined at 15°~20° towards the direction of drum rotation; and reinforcing ribs with a width of 8 mm~10 mm are provided at the weld between the baffles and the inner wall of the drum. During centrifugation, the change in the conductivity of the oil phase is monitored in real time. The initial conductivity of the crude peanut oil when it enters the centrifuge is used as a benchmark. Separation is terminated when the real-time conductivity decreases by less than 50 μS / cm for 3 consecutive minutes. The separated heavy phase impurities are discharged through the slag discharge port at the bottom of the drum. The slag discharge interval is 15 to 20 minutes. During slag discharge, the pressure fluctuation inside the drum is maintained at ≤0.05 MPa by a pressure compensation device. Before the light phase oil is introduced into the temporary storage tank through the oil outlet pipe, the oil temperature is first reduced to 25℃ to 30℃ by a serpentine tube cooler. The cooling water flow rate in the serpentine tube is 2 m / s to 3 m / s, and the oil flow and cooling water are in countercurrent contact.

5. The method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing as described in claim 1, characterized in that, Step 1) specifically includes: S11. A near-infrared hyperspectral imaging system with a wavelength range of 900 nm to 1700 nm is used, with a line scanning speed of 200 kernels / second to 300 kernels / second. Each peanut kernel is subjected to 3 pulse scans with an interval of 20 ms. S12. The obtained spectral data is processed by a partial least squares discriminant analysis model, which is trained based on the spectral characteristics of more than 3,000 aflatoxin contaminated samples and outputs the mold probability value in real time. When the mold probability value of a single peanut kernel is ≥0.90, the pneumatic nozzle is triggered, and compressed air with a pressure of 0.6 MPa to 0.8 MPa is sprayed to blow the moldy particles into the waste channel. S13. Peanut kernels that have been initially screened by the near-infrared hyperspectral imaging system enter a manual sorting table equipped with a 450 nm to 550 nm dual-band LED light source. The surface illuminance of the sorting table conveyor belt is maintained at 8000 lux to 12000 lux, and the conveying speed is 0.3 m / s to 0.5 m / s. The sorting operator wears 500 nm to 600 nm band-pass filter glasses for secondary sorting. The sorting operation table is equipped with a spring buffer device. Removed particles enter a sealed recycling container through a slide with an inclination angle of 30° to 35°. The hyperspectral imaging system has a spatial resolution of 0.2 mm / pixel to 0.3 mm / pixel and a spectral resolution of ≤8 nm. After each batch of tests is completed, whiteboard calibration is automatically performed with a calibration interval of ≤2 hours.

6. The method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing as described in claim 2, characterized in that, Following step S52, the process includes the following steps: Filtered peanut oil is fed into a closed-loop plate heat exchanger, with the oil inlet temperature controlled at 28℃~32℃. The heat exchanger uses a propylene glycol solution at -5℃~0℃ as the refrigerant, causing the oil temperature to drop to 18℃~22℃ within 15 seconds. The cooled peanut oil is then fed into a molecular sieve dehydration tower via a nitrogen-sealed pipeline for dehydration. The molecular sieve is type 3A zeolite, with a packing height of 2.5 to 3 times the tower diameter. The oil flows through the bed at a linear velocity of 0.6 m / h~0.8 m / h, and the nitrogen pressure inside the tower is maintained at 0.12 MPa~0.15 MPa. The dehydrated peanut oil then enters a vacuum degassing tank for degassing to obtain degassed oil. The absolute pressure inside the tank is controlled at 8 kPa~12 kPa. The peanut oil flows down the surface of stainless steel corrugated packing in a film-like form, with a specific surface area of ​​350 m². 2 / m 3 ~400 m 2 / m 3 The degassing time is 25 min to 35 min. The degassed oil is tested by an online dissolved oxygen meter. When the dissolved oxygen content is ≤0.8 mg / L, it is transported to the storage tank through a double-layer 316L stainless steel sleeve. Nitrogen gas is introduced into the sleeve jacket and maintained at a flow rate of 1.5 m / s to 2 m / s to form a gas barrier. The storage tank is equipped with a conical bottom agitator with a blade tip linear velocity ≤0.3 m / s. The breather valve on the top of the tank is equipped with a two-stage silica gel-activated carbon composite filter. The inner wall of the storage tank is coated with a parylene coating with a thickness of 8 μm to 12 μm.

7. The method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing as described in claim 1, characterized in that, Step 3) specifically includes: S31. The apex angle of the triangular tooth pattern is 55°~65°, the tooth height is 0.5 mm~0.8 mm, the tooth spacing is 2 mm~3 mm, and the tooth pattern is arranged obliquely at 45° along the axial direction of the roller surface. The center line distance between two adjacent rows of tooth patterns is 4 mm~6 mm. The feed inlet of the roller crusher is equipped with a double spiral guide plate with a spiral pitch of 50 mm~80 mm, a gap between the end of the guide plate and the roller surface of 1.5 mm~2 mm, and a guide plate rotation speed of 10 r / min~15 r / min. The fine powder rate of the crushed particles is monitored online by a laser particle size analyzer. When the proportion of particles with a diameter <0.5 mm exceeds 5%, an alarm is triggered and the gap of the double spiral guide plate is adjusted to 1.8 mm~2.2 mm simultaneously. S32. During the crushing process, the roller surface temperature is monitored in real time. When the temperature exceeds 40℃, the built-in spray device is activated to spray atomized water droplets onto the roller surface. The water droplet size is 50 μm to 80 μm and the spray flow rate is 20 mL / min to 30 mL / min. S33. The crushed peanut particles are graded by a double-layer vibrating screen. The upper screen mesh size is 3.5 mm, the lower screen mesh size is 2 mm, the screen surface inclination angle is adjusted to 12°~15°, and the screen body amplitude is set to 3 mm~5 mm. Among them, the drive motor of the roller crusher is equipped with a dynamic load compensation device. When the current fluctuation exceeds the rated value ±15%, the roller spacing is automatically adjusted to compensate by 0.1 mm to 0.3 mm, and the crushing energy consumption is controlled at 0.8 kWh / t to 1.2 kWh / t.

8. The method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing as described in claim 1, characterized in that, In step 4), the circulating water system adopts a dual-loop circulation. The first loop is connected to the compression section jacket of the pressing chamber, with a water temperature of 45℃~48℃ and a flow rate of 1.5 m / s~2 m / s. The second loop is connected to the cake outlet jacket, with a water temperature of 48℃~50℃ and a flow rate of 0.8 m / s~1 m / s. The temperature difference between the two loops is controlled by a proportional valve to be ≤3℃.

9. The method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing as described in claim 1, characterized in that, In step 4), nitrogen injection is set at three inlet points. The first inlet point is at the end of the feeding section, and its injection volume accounts for 40% to 50%. The second inlet point is in the middle of the compression section, and its injection volume accounts for 30% to 40%. The third inlet point is at the front of the cake discharge section, and its injection volume accounts for 10% to 20%. The injection pressure is 0.1 MPa to 0.3 MPa higher than the real-time pressure inside the pressing chamber at each injection point.

10. The method for preparing virgin peanut oil based on near-infrared mold sorting and low-oxygen temperature-controlled pressing as described in claim 1, characterized in that, Step 4) also includes: real-time acquisition of the pressing chamber pressure curve during the pressing process; when the pressure fluctuation exceeds the set value ±5 MPa, the taper of the cake section is dynamically adjusted by the hydraulic servo mechanism at an adjustment rate of 0.5 mm / min to 1 mm / min; the screw shaft is equipped with a torque compensation device, which automatically reduces the rotation speed by 1 r / min to 2 r / min when the torque value reaches 1800 N·m to 2200 N·m, maintaining the power stable at 22 kW to 25 kW; the residual oil rate of the pressed cake is detected by a near-infrared online analyzer, and when the residual oil rate is >8%, an alarm is triggered and the pressure at the end of the compression section is simultaneously increased by 3 MPa to 5 MPa for a duration of 30 s to 60 s; an oxygen content sensor is installed at the pressing chamber outlet, and when the detected value is >0.6%, the total nitrogen flow rate is increased by 20% to 30% until the oxygen content recovers to 0.3% to 0.5%.

Images