Pretreatment method of sample for detecting nutritional ingredients of whole poultry body

Abstract

The invention discloses a pretreatment method of a sample for detecting nutritional ingredients of a whole poultry body, which comprises the following steps: detecting the temperature of different parts of a frozen whole poultry, identifying substandard parts, performing supplementary freezing, and performing slitting according to partition freezing degree distribution; in the drying process, a temperature sequence is monitored to recognize and regulate abnormal points, the surface layer and center moisture content is detected to calculate the gradient, low-temperature delay treatment is conducted on a sample with the gradient exceeding the standard, and the drying end point is judged by combining moisture content data and the moisture content gradient; carrying out stratified sampling on the coarsely crushed sample, detecting the component proportion and particle size distribution, determining superfine crushing process parameters, and carrying out liquid carbon dioxide assisted superfine crushing and recycling; and performing multi-point sampling on the mixed sample, evaluating the uniformity, detecting the oxygen concentration of each layer, determining residual oxygen concentration data, determining filling parameters of recovered carbon dioxide, and sealing. According to the method, accurate matching of process parameters and cyclic utilization of carbon dioxide are realized, a pretreatment sample with uniform components and stable oxidation is obtained, and the requirements of analysis and detection on sample representativeness and stability are met.

Description

Technical Field

[0001] This invention relates to the field of food testing sample preparation technology, and in particular to a pretreatment method for samples used in the detection of nutritional components in whole poultry. Background Technology

[0002] As an important source of animal protein, accurate testing of the overall nutritional composition of poultry is fundamental to feed evaluation, food quality control, and nutritional research. Testing the nutritional composition of whole poultry requires preparing the entire poultry carcass into a uniform powder sample. This preparation process involves steps such as freeze pretreatment, dehydration and drying, mechanical pulverization, and mixing and sealing. The quality control of each step directly affects the representativeness of the final sample's components and the accuracy of the analytical results.

[0003] Current whole poultry body sample pretreatment techniques typically use uniform process parameters to process whole raw materials, neglecting the characteristic differences of different tissue components. Insufficient freezing in some areas during the freezing pretreatment stage leads to difficulties in segmentation; temperature fluctuations during the drying stage cause uneven moisture removal; differences in hardness among components during the pulverization stage result in uneven particle size distribution, and the failure to recover the cooling medium leads to waste; residual oxygen during the sealing stage accelerates lipid oxidation. These problems result in low component homogeneity and poor storage stability of the pretreated samples, making it difficult to meet the requirements of sample representativeness and consistency for nutritional component testing. Summary of the Invention

[0004] This invention discloses a pretreatment method for samples used in the analysis of nutritional components of whole poultry. It implements zoned freezing state monitoring and control, achieves precise drying through temperature anomaly identification and compensation and moisture content gradient monitoring and control, determines differentiated pulverization process parameters based on component stratification characteristics, employs liquid carbon dioxide-assisted cooling and recovery filling, and optimizes filling by combining uniformity assessment and residual oxygen monitoring. This yields pretreated samples with narrow particle size distribution, uniform composition, and stable oxidation, ensuring good representativeness and repeatability of nutritional component detection values ​​from each sampling location, meeting the requirements of sample uniformity and stability for analytical testing.

[0005] This invention proposes a sample pretreatment method for detecting the nutritional components of whole poultry, comprising the following steps: Frozen whole poultry is subjected to zoned temperature detection and freezing control to generate a zoned freezing degree distribution. Based on the zoned freezing degree distribution, the frozen whole poultry is cut into small pieces. Temperature-time monitoring is performed on the drying process of the poultry carcass pieces to generate a drying process curve. The surface and core moisture content of the poultry carcass pieces are detected to obtain moisture content data and moisture content gradient. Based on the moisture content data, the moisture content gradient and the drying process curve, the drying endpoint is determined to obtain a qualified dried sample. The dried qualified sample is coarsely pulverized to generate a coarse powder sample. The coarse powder sample is then subjected to stratified sampling and particle size distribution detection to obtain component distribution data and particle size distribution data. The ultrafine pulverization process parameters are determined based on the component distribution data and particle size data. The coarse powder sample is then subjected to liquid carbon dioxide-assisted low-temperature ultrafine pulverization treatment according to the ultrafine pulverization process parameters to obtain an ultrafine powder sample and the carbon dioxide is recovered. The ultrafine powder sample is homogenized to obtain a mixed sample. The oxygen content and uniformity of the mixed sample are detected to obtain residual oxygen concentration data and uniformity data. The carbon dioxide recovery filling parameters are determined based on the residual oxygen concentration data and the uniformity data. The mixed sample is then sealed according to the carbon dioxide recovery filling parameters to obtain a pre-treated sample.

[0006] The beneficial effects of this invention are reflected in the following points: First, the temperature of the breast, leg, and wing parts of the frozen whole poultry is detected separately to identify substandard parts and supplement freezing. The cutting path is planned according to the distribution of freezing degree in each area to ensure that each part is cut under a uniform freezing state, ensuring that the cut surface is regular and the tissue is not deformed. During the drying process, the temperature sequence is monitored in real time to identify abnormal points. Time compensation is performed for low temperature abnormalities and temperature adjustment is performed for high temperature abnormalities. The gradient is calculated by detecting the surface and core moisture content. The samples with excessive gradients are subjected to low temperature delay treatment to promote uniform moisture migration. The drying endpoint is determined by combining the moisture content data and the moisture content gradient to ensure that the moisture content meets the standard and is uniform inside and out, avoids protein thermal denaturation, and maintains the true representativeness of the nutritional components of the original poultry in the test samples. Secondly, the coarse powder sample was sampled in layers. Based on the density data of each layer, the boundaries of the component layers were identified and the influence of the thickness of the boundary mixing layer was corrected. The component ratio of each layer was calculated to generate component distribution data. The difficult-to-crush components were identified by combining the particle size distribution data and the particle size characteristics of the components. The recommended crushing time was queried according to the ratio of the three components. The temperature monitoring threshold was determined by combining the difficult-to-crush marker and intermittent cooling was triggered to obtain ultrafine powder samples with narrow particle size distribution and uniform morphology, so that any sampling point can represent the original component ratio of the whole poultry. Liquid carbon dioxide was used to assist in low-temperature crushing and was recovered for subsequent filling, realizing the recycling of carbon dioxide in the sample preparation process. Finally, the homogeneity of the mixed sample was assessed by taking multiple samples from different points to detect differences in components. The oxygen content of each layer was measured, and the maximum value was taken as the residual oxygen concentration data. The initial filling amount was determined by comparing it with the target threshold. Based on the homogeneity data, low homogeneity areas were identified and supplementary filling marks were generated. The secondary filling amount was determined based on the residual oxygen value after filling and the supplementary filling marks after filling effect test. Low oxygen sealing was achieved by using recycled carbon dioxide for filling and sealing, which inhibited lipid oxidation and protein degradation, ensuring that the nutritional composition of the pretreated sample remained stable during storage and meeting the consistency requirements of repeated testing.

[0007] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0008] The accompanying drawings illustrate specific examples of the technical solutions described in this invention and, together with the detailed embodiments, form part of the specification, serving to explain the technical solutions, principles, and effects of this invention.

[0009] Unless otherwise specified, the same reference numerals in different figures represent the same or similar technical features, and different reference numerals may be used to represent the same or similar technical features.

[0010] Figure 1 This is a schematic flowchart of a sample pretreatment method for detecting the nutritional components of whole poultry, according to the present invention.

[0011] Figure 2 This is a schematic diagram of the liquid carbon dioxide-assisted ultrafine pulverization and recycling system of the present invention.

[0012] Wherein: 1-Airflow pulverizer; 2-Pulverizing chamber; 3-Counter-jet nozzle; 4-Cooling jacket; 5-Liquid carbon dioxide storage tank; 6-First flow control valve; 7-Temperature sensor; 8-Inlet; 9-Outlet; 10-Gas recovery pipeline; 11-Dehumidifier filter; 12-Compressor unit; 13-Recovery storage tank; 14-Filling container; 15-Mixed sample; 16-Air inlet; 17-Exhaust outlet; 18-Adjustable nozzle; 19-Oxygen concentration probe; 20-Gas supply pipeline; 21-Second flow control valve. Detailed Implementation

[0013] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0014] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0015] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0016] The technical solutions of the embodiments of this application will be described below.

[0017] like Figure 1 As shown, this embodiment of the invention provides a sample pretreatment method for detecting the nutritional components of whole poultry, including the following steps S110-S140: Step S110: Perform zoned temperature detection and freezing control on the frozen whole poultry to generate a zoned freezing degree distribution. Based on the zoned freezing degree distribution, perform a cutting process on the frozen whole poultry to form small pieces of poultry.

[0018] In some embodiments, the step of performing zoned temperature detection and freezing control on frozen whole poultry to generate a zoned freezing degree distribution includes: separately detecting the temperature of the breast, leg, and wing parts of the frozen whole poultry to obtain the part temperature; comparing the part temperature with a freezing temperature threshold to identify parts that do not meet the threshold and generating a mark for the parts that do not meet the threshold; performing supplementary freezing treatment on the parts marked with the marks for the parts that do not meet the threshold to determine the degree of freezing of the parts; and constructing a zoned freezing degree distribution based on the degree of freezing of the parts.

[0019] Temperature measurements were taken from the breast, leg, and wing areas of frozen whole poultry. Before temperature testing, the frozen whole poultry was confirmed to be intact and undamaged. The surface temperature of the frozen whole poultry was quickly scanned using an infrared thermometer to confirm a frozen state below -10℃. For the breast area, the probe was inserted perpendicularly to the body surface into the thickest part of the pectoralis major muscle on both sides of the keel, to a depth of 2.5 cm, corresponding to the midpoint of a 5-6 cm thickness of the pectoral muscle. If the insertion depth was too shallow, only the surface temperature was measured, which would not represent the internal frozen state; if it was too deep, approaching the bone, the high thermal conductivity of the bone would affect the accuracy of the reading. For the leg area, two independent measurement points were used: the thigh and the lower leg. The thigh probe was inserted 1.8 cm into the middle of the lateral femoral muscle, and the lower leg probe was inserted 1.2 cm into the intermuscular space between the tibia and fibula. These two measurement points covered the main muscle distribution area of ​​the leg. For the wing area, the probe was inserted 1 cm into the junction of the wing root and wing center, where the muscle tissue is relatively thick and the bones are small. After probe insertion, it is allowed to stand for 15 seconds to reach thermal equilibrium. The heat generated by the friction between the probe and muscle tissue during insertion is approximately 0.2 to 0.5°C. Reading the temperature directly without allowing it to stand will overestimate the actual temperature, leading to a false judgment of failure to meet the standard. After 15 seconds, the temperature reading stability error is less than 0.05°C. The PT100 platinum resistance temperature sensor has a response time of less than 5 seconds and a resolution of 0.1°C. The sensor's output resistance value is converted to a Celsius temperature value via a temperature conversion module. The temperature at the chest area is -13.2°C, the thigh area is -16.1°C, the lower leg area is -16.9°C, the wing root area is -17.5°C, and the wing midsection area is -18.1°C, with a temperature difference of 2 to 5°C between the five detection points.

[0020] The system compares the temperature of each part of the meat with a freezing temperature threshold to identify and mark substandard parts. The freezing temperature threshold of -15℃ is determined based on the freezing characteristics of poultry meat tissue. Below -15℃, the internal ice crystal structure of muscle tissue is stable, and the mechanical strength of the tissue cells is sufficient to support subsequent cutting operations. The automatic comparison program compares the temperature of each part with the freezing temperature threshold of -15℃ one by one. The comparison algorithm uses a simple threshold judgment: parts with a temperature higher than the threshold (higher temperature, lower degree of freezing) are marked TRUE, indicating substandard, while parts with a temperature lower than or equal to the threshold are marked FALSE, indicating that they meet the standard. The chest region has a temperature of -13.2℃, which is higher than the freezing temperature threshold of -15℃, with a difference of 1.8℃; the comparison result returns TRUE, indicating a substandard part. The thigh region has a temperature of -16.1℃, which is lower than the freezing temperature threshold, with a temperature difference of -1.1℃; the comparison result returns FALSE, indicating that it meets the standard. Temperatures of -16.9℃ in the lower leg area, -17.5℃ in the wing root area, and -18.1℃ in the middle wing area were all below the freezing temperature threshold, and the comparison results for these areas were all FALSE, indicating a good freezing condition. The substandard area markers recorded information about the substandard areas in structured data format. The chest region was assigned the substandard area marker code "01-CHEST," where "01" indicates that this area is the first priority area requiring supplemental freezing, and "CHEST" identifies the specific area name. The data structure for the substandard area markers includes two fields: area code and temperature difference. The temperature difference of 1.8℃ in the chest region was recorded in the "Temperature Deviation" field of the substandard area marker. The substandard area markers provide precise location and degree information for subsequent supplemental freezing treatment.

[0021] For areas marked as substandard, supplementary freezing is performed to determine the degree of freezing. The breast region marked with these substandard areas is automatically conveyed to a rapid freezing unit. The rapid freezing unit operates at -30°C, 15°C lower than the standard freezing threshold, allowing the temperature of the marked areas to drop to the required range in a short time. Supplementary freezing employs forced convection cooling, with a refrigeration fan blowing cold air at a speed of 2 m / s directly onto the surface of the breast region marked with substandard areas to accelerate heat dissipation. The supplementary freezing time is determined by referring to a table based on the temperature difference recorded in the substandard area markings and the poultry meat's thermophysical parameters. A temperature difference of 1.8°C in the breast region corresponds to approximately 45 minutes of supplementary freezing time. If the supplementary freezing time is insufficient (e.g., only 20 minutes), the breast region temperature may only drop to -14°C, resulting in a semi-frozen muscle tissue during subsequent slicing. This causes deformation rather than brittle fracture when the blade enters, affecting the regularity of the cut surface and the consistency of sample preparation. After supplemental freezing, the chest area marked with substandard areas was retested for verification. During the retest, the temperature sensor probe insertion position was exactly the same as the initial test to ensure comparability of the measurement results. The retest temperature of the chest area was -16.8℃, successfully dropping below the freezing temperature threshold of -15℃. The difference between this temperature and the threshold changed from 1.8℃ in the initial test to -1.8℃, indicating that the area had reached a fully frozen state. The degree of freezing of a region is defined as the difference between the final measured temperature and the freezing temperature threshold of -15℃, expressed in degrees Celsius. A positive difference indicates insufficient freezing, while a negative difference indicates sufficient freezing. The freezing degree of the chest area was -1.8℃, the thigh area -1.1℃, the lower leg area -1.9℃, the wing root area -2.5℃, and the wing midsection area -3.1℃. The negative freezing degree values ​​for all areas indicate that each area has been fully frozen. This zoned freezing control enabled precise control of the freezing state of different areas.

[0022] A regional distribution of freezing degree was constructed based on the freezing degree of different body parts. The freezing degree of different body parts was arranged in anatomical spatial order: chest, legs, and wings. The arrangement resulted in a one-dimensional array of length 5, with each element representing the freezing degree of one detection point on the chest, two on the legs, and two on the wings. This one-dimensional freezing degree array was mapped to the three-dimensional spatial coordinate system of the frozen whole poultry. The detection point in the chest region was located in the front middle of the poultry, corresponding to coordinates (0,0,0); the detection point in the thigh region was located in the lower rear of the poultry, corresponding to coordinates (-5,-8,0); the detection point in the lower leg region was located in the distal end of the poultry, corresponding to coordinates (-5,-15,0); the detection point at the wing root was located in the upper side of the poultry, corresponding to coordinates (6,-3,0); and the detection point in the middle of the wing was located in the outer side of the poultry, corresponding to coordinates (10,-3,0). The coordinate unit was centimeters. A correspondence was established between the freezing degree of the five detection points and the spatial coordinates. The freezing degree values ​​of undetected locations were estimated by weighting their spatial distances from the detection points. The distribution of freezing degree in different zones is stored in the form of a three-dimensional data matrix, covering an area of ​​approximately 20 cm × 18 cm × 10 cm on the surface of the frozen whole poultry. Each voxel in the matrix stores the freezing degree value corresponding to its spatial location. In the distribution of freezing degree in different zones, areas with negative freezing degree values ​​indicate that the freezing is complete, and the larger the absolute value of the freezing degree, the deeper the freezing. Areas with large differences in freezing degree values ​​between adjacent locations correspond to the boundaries of different tissues.

[0023] Frozen whole poultry is slit into smaller pieces based on the distribution of freezing degree in different zones. The slitting blade path is planned along the location of the maximum freezing degree gradient in the zone freezing degree distribution. The location of the maximum gradient corresponds to the boundary between different areas of the frozen whole poultry. At this location, the tissue properties change abruptly, such as at the junction of bone and muscle, where the mechanical strength difference is significant in the frozen state, allowing the blade to selectively cut into the side with lower hardness. At the junction of the breast and leg, the freezing degree gradient is 0.7℃ / cm according to the zone freezing degree distribution. The slitting blade cuts along the hip joint space. In the frozen state, the articular cartilage is brittle and fractures, reducing the slitting resistance by about 60%. At the junction of the leg and wing, the slitting is performed at the shoulder joint. The zone freezing degree distribution shows a freezing degree gradient of 1.4℃ / cm at this location, making collagen fibers prone to breakage. A band saw is used as the slitting tool to cut the frozen whole poultry. The saw blade width is 3 mm, the saw tooth spacing is 5 mm, and the running speed is 15 m / s to ensure a smooth cut and avoid frictional heat generation. The breast region is longitudinally cut along the keel midline into two small pieces, each weighing approximately 200 grams. The leg region is transversely cut at the junction of the thigh and lower leg knee joints, resulting in a thigh piece weighing 150 grams and a lower leg piece weighing 80 grams. The wing region is cut at the junction of the wing root and wing mid-section elbow joints, resulting in a wing root piece weighing 60 grams and a wing mid-section piece weighing 40 grams. These pieces are conveyed to the next process via a stainless steel conveyor belt maintained at -5°C to prevent overheating. The pieces remain on the conveyor belt for no more than 2 minutes to ensure they remain frozen before entering the drying process. After being cut, the frozen whole poultry is divided into six pieces: left breast piece, right breast piece, thigh piece, lower leg piece, wing root piece, and wing mid-section piece.

[0024] Step S120: Temperature-time monitoring is performed on the drying process of poultry carcass pieces to generate a drying process curve. The surface and core moisture content of the poultry carcass pieces are detected to obtain moisture content data and moisture content gradient. Based on the moisture content data, moisture content gradient and drying process curve, the drying endpoint is determined to obtain a qualified dried sample.

[0025] In some embodiments, the step of monitoring the temperature and time during the drying process of the poultry carcass pieces to generate a drying process curve includes: real-time monitoring of the temperature and time during the drying process to obtain a temperature sequence and drying time; performing anomaly detection on the temperature sequence to identify temperature anomaly points and generate temperature anomaly markers; compensating the drying time based on the temperature anomaly markers to determine a compensating drying time; and correspondingly associating the temperature sequence with the compensating drying time to generate a drying process curve.

[0026] Real-time temperature and time monitoring is performed during the drying process to obtain temperature sequences and drying times. The drying process takes place inside a drying chamber, and temperature control directly affects the drying efficiency and quality of the poultry carcass pieces. Temperature sensors (K-type thermocouples) continuously measure the temperature inside the drying chamber at 30-second sampling intervals. The sampling interval is determined based on the rate of temperature change; the initial temperature change rate for the poultry carcass pieces is approximately 2°C / minute. A 30-second sampling interval captures key temperature change points. If the sampling interval is too long (e.g., 5 minutes), short-term temperature fluctuations may be missed, leading to undetected anomalies. If it is too short (e.g., 5 seconds), data redundancy occurs, increasing storage burden. The thermoelectric potential signal output by the sensor is converted into a standard voltage signal of 0 to 10 volts by a temperature transmitter. After analog-to-digital conversion, the voltage signal is stored in a temperature array. The temperature array records the temperature measurement values ​​of each sampling point in chronological order, forming a temperature sequence. A standard drying time of 120 minutes corresponds to 240 sampling points. The initial length of the temperature sequence is 240, and the array indices from 0 to 239 correspond to drying times from 0 to 120 minutes. Time monitoring is achieved using a timer with a resolution of 1 second. The timer starts when the poultry pieces enter the drying chamber and stops when the drying process ends, recording the total drying time. A correspondence is established between the drying time and the temperature sequence using an array index; the nth element in the temperature sequence corresponds to a drying time of n × 0.5 minutes.

[0027] Anomaly detection is performed on the temperature sequence to identify temperature anomalies and generate temperature anomaly markers. The standard operating temperature range of the drying oven is set to 55 to 65°C. This range is determined based on a combination of the thermal stability of poultry protein and the efficiency of moisture evaporation. Below 55°C, the moisture evaporation rate is insufficient; at only 50°C, the evaporation rate decreases by approximately 40%, and the moisture content after 120 minutes of drying may still be higher than 12%, failing to meet the requirements for subsequent pulverization. Temperatures above 65°C may cause partial protein denaturation; if the temperature reaches 70°C and remains above 10 minutes, irreversible denaturation of muscle protein occurs, altering the protein composition of the sample and causing the test results to deviate from the true nutritional composition values ​​of the original poultry carcass. Furthermore, the sample will develop a burnt odor, affecting subsequent component identification. Each temperature value in the temperature sequence is compared one by one with the standard temperature range. Temperature values ​​falling within the 55 to 65°C range are considered normal, while values ​​exceeding this range are considered abnormal. The algorithm for identifying temperature anomalies employs a threshold detection method. The detection program iterates through the temperature sequence, and when a temperature value T is detected that is less than 55℃ or greater than 65℃, the index of that sampling point is recorded as a temperature anomaly. The temperature value of the 45th sampling point in the temperature sequence is 52℃, which is lower than the standard lower limit of 55℃, and this sampling point is marked as a temperature anomaly, with a corresponding drying time of 22.5 minutes. The temperature value of the 78th sampling point in the temperature sequence is 67℃, which is higher than the standard upper limit of 65℃, and this sampling point is also marked as a temperature anomaly, with a corresponding drying time of 39 minutes. The temperature anomaly markers record detailed information about the temperature anomalies in structured data format, with the data structure containing two fields: anomaly temperature value and anomaly type. The anomaly type of the 45th temperature anomaly is marked as "low temperature anomaly," and the anomaly type of the 78th temperature anomaly is marked as "high temperature anomaly." The anomaly type of the temperature anomaly markers is used for differentiation in subsequent compensation calculations. In this drying process, one low temperature anomaly and one high temperature anomaly were detected, providing a basis for drying time compensation.

[0028] The compensation drying time is determined based on temperature anomaly markers. Low-temperature anomaly markers reduce moisture evaporation efficiency, requiring extended drying time for compensation. The compensation time is determined by the difference between the abnormal temperature value in the anomaly marker and the standard lower limit, with a compensation coefficient of 2 minutes / ℃, meaning 2 minutes of drying time is compensated for every 1℃ temperature difference. The temperature anomaly marker at the 45th temperature anomaly point shows a temperature of 52℃, with a difference of 3℃ from the standard lower limit of 55℃. The corresponding compensation time for this anomaly marker is 6 minutes. If the compensation time is insufficient (e.g., only 2 minutes), the low-temperature anomaly is not fully corrected, resulting in higher moisture content in some areas of the sample, causing adhesion during subsequent pulverization. The compensation time is accumulated to the standard drying time of 120 minutes, resulting in a new total drying time of 126 minutes. High-temperature anomaly markers, while accelerating moisture evaporation, may affect sample quality. High-temperature anomaly markers do not require drying time compensation, but the drying oven setting temperature needs to be lowered after the anomaly marker is generated. The temperature anomaly marker at the 78th temperature anomaly point indicated a temperature of 67℃, higher than the standard upper limit of 65℃. This anomaly triggered a temperature control command, reducing the drying oven's set temperature from 60℃ to 58℃. The calculation of the compensatory drying time only considered the impact of the low-temperature anomaly marker. In this drying process, one low-temperature anomaly marker corresponded to a compensation time of 6 minutes, resulting in a total compensatory drying time of 126 minutes. After the compensatory drying time calculation was completed, the control endpoint of the drying time was updated. A drying end signal was triggered when the drying time reached the 126-minute compensatory drying time, ensuring the adequacy of the drying process.

[0029] A drying process curve is generated by correlating the temperature sequence with the compensated drying time. The sampling point index of the temperature sequence is converted into the corresponding drying time coordinate by multiplying by the sampling interval of 0.5 minutes. The converted time coordinate serves as the horizontal axis data of the drying process curve. The compensated drying time of 126 minutes is used as the maximum value of the horizontal axis, with the horizontal axis ranging from 0 to 126 minutes. The vertical axis represents the drying oven temperature, ranging from 0 to 70℃. For the compensation stage corresponding to the compensated drying time, 12 additional sampling points are added to record the temperature changes during that stage, extending the total length of the temperature sequence to 252 sampling points. The 252 temperature values ​​in the temperature sequence serve as the vertical axis data of the drying process curve. Each temperature value and its corresponding time coordinate form a coordinate point, and the 252 coordinate points are connected in chronological order to form a continuous drying process curve. The drying process curve is plotted as a line graph, with a scale mark every 10 minutes on the horizontal axis and every 10℃ on the vertical axis. Temperature anomalies are marked on the drying process curve for easy identification. The 45th anomaly point is located at (22.5, 52) on the drying process curve. This point is marked in red to indicate a low-temperature anomaly. The 78th anomaly point is located at (39, 67) on the drying process curve. This point is marked in orange to indicate a high-temperature anomaly. The drying process curve is saved as a data file and linked to the corresponding batch of poultry carcass records. The drying process curve records the temperature change history throughout the entire drying process.

[0030] Moisture content data and gradients were obtained by measuring the surface and core moisture content of poultry carcasses. After the drying process, the carcasses were removed from the drying oven, and the surface temperature was confirmed to be below 30°C using an infrared thermometer before moisture content testing. The moisture content was measured using an oven drying method. Samples were taken from both the surface and core of each carcass piece. The surface sampling depth was 0 to 2 mm, and the core sampling depth was the center of the carcass piece's thickness. For example, for the left breast piece (20 mm thick), a 5-gram sample was taken 1 mm below the surface, and a 5-gram sample was taken 10 mm below the core. The samples were placed in a pre-weighed aluminum box, and the box and sample were weighed together on an analytical balance to record the wet weight with an accuracy of 0.001 g. The aluminum box containing the sample was then placed in a 105°C oven for 4 hours, cooled to room temperature, and weighed to record the dry weight. The moisture content gradient is calculated using the following formula: ΔM = Mcenter - Msurface, where ΔM is the moisture content gradient (%), Mcenter is the center moisture content (%), and Msurface is the surface moisture content (%). The surface moisture content of the left breast piece is 6.8%, and the center moisture content is 9.5%, so the moisture content gradient ΔM = 9.5% - 6.8% = 2.7%. The moisture content data for small pieces of the poultry body is taken as the average of the surface and center moisture contents, which is 8.15%. The test results for six small pieces of the poultry body show: right breast piece moisture content 8.05%, moisture content gradient 3.1%; thigh piece moisture content 7.85%, moisture content gradient 3.3%; lower leg piece moisture content 6.60%, moisture content gradient 3.0%; wing root piece moisture content 5.75%, moisture content gradient 1.9%; and mid-wing piece moisture content 4.85%, moisture content gradient 1.3%. The moisture content gradient is positively correlated with the thickness of the small pieces of poultry, with thicker samples exhibiting a larger gradient due to the longer internal moisture migration path.

[0031] Determine the drying end point based on the moisture content data, moisture content gradient and drying process curve to obtain a qualified dried sample. The drying qualification criteria include three items: first, the moisture content data is not higher than 10%; second, the moisture content gradient is not higher than the gradient threshold of 3.0%; third, there are no severe high-temperature abnormal points in the drying process curve. This standard is determined based on the requirements of the subsequent crushing preparation process of the sample and the need to maintain the authenticity of the nutritional components of the test sample. The determination of the drying end point first conducts a threshold test on the moisture content data, and the moisture content data of 6 small pieces of poultry bodies are all lower than the 10% threshold, meeting the requirements. Secondly, the moisture content gradient is tested. The gradient threshold is set at 3.0%. This threshold is determined based on the poultry meat crushing process. When the gradient exceeds 3.0%, the surface layer is over-dried to form a hardened layer, and when crushing, the surface layer fragments are mixed with the internal soft particles, resulting in uneven particle size distribution. The test results show that the gradient of the thigh piece is 3.3% and the gradient of the right chest piece is 3.1%, exceeding the threshold, and it is determined that the gradient is unqualified. Supplementary treatment is carried out on the samples with excessive gradient: return to the drying oven, reduce the drying temperature to 50°C, and extend the drying time by 30 minutes. The low-temperature extension promotes the slow migration of internal moisture to the surface layer. After supplementary treatment, the gradient of the thigh piece drops to 2.1% and the gradient of the right chest piece drops to 2.2%, both meeting the requirements. The inspection of abnormal points in the drying process curve is achieved by querying the temperature abnormal mark. In this test, 1 low-temperature abnormal point (already processed by time compensation) and 1 high-temperature abnormal point (67°C for 30 seconds, determined to be an acceptable abnormal) are detected. After all three determination criteria are met, the drying end point determination result is qualified, and a qualified dried sample is generated. The qualified dried sample is transferred to an aluminum foil composite material sealed container, filled with high-purity nitrogen to reduce the oxygen concentration to less than 0.5%, and left standing for 24 hours to make the internal moisture distribution uniform. Samples with a low moisture content gradient do not agglomerate during crushing, and the particle size uniformity is better.

[0032] Step S130, perform coarse crushing on the qualified dried sample to generate a coarse powder sample, conduct layered sampling detection and particle size distribution detection on the coarse powder sample to obtain component distribution data and particle size distribution data, determine the ultra-fine crushing process parameters based on the component distribution data and particle size distribution data, and perform liquid carbon dioxide-assisted low-temperature ultra-fine crushing on the coarse powder sample according to the ultra-fine crushing process parameters to obtain an ultra-fine powder sample and recover carbon dioxide.

[0033] Specifically, the dried and qualified samples were coarsely pulverized to produce coarse powder. Six small pieces of poultry carcasses from the dried and qualified samples were sequentially fed into a hammer mill for primary pulverization. Before pulverization, the dried and qualified samples were confirmed to have no abnormal mold or discoloration on the surface. The total mass of the dried and qualified samples was approximately 730 grams. The hammer mill was equipped with 12 alloy steel hammers, rotating at 2800 rpm. The high-speed rotating hammers generated a strong impact and tearing effect on the small pieces of poultry carcasses. The left breast piece, weighing 200 grams, was first fed into the pulverizer's inlet. The impact of the hammers tore the left breast piece into multiple fragments. These fragments were further pulverized by repeated collisions with the sieve plate in the pulverizing chamber. The pulverizer's sieve plate had a 5 mm aperture. Particles smaller than 5 mm passed through the sieve plate and particles larger than 5 mm continued to be pulverized in the pulverizing chamber until they reached the discharge particle size. The coarse pulverization time for the left breast piece was 2 minutes. The heat generated during the pulverization process was dissipated through the water-cooled jacket of the pulverizing chamber. The cooling water temperature was 5°C, and the circulation flow rate was 8 liters / minute, keeping the temperature of the pulverizing chamber below 25°C. After coarse grinding, the sample was discharged from the grinder outlet and collected in a stainless steel tray. The mass of the coarse powder sample from the left breast piece was 198 grams, with a grinding loss rate of 1%, mainly due to dust dispersion. The particle morphology of the coarse powder sample was irregular polygonal, feather residue was fibrous, bone fragments were flaky, and muscle tissue was granular. The three components showed significantly different morphologies in the coarse powder sample due to differences in physical properties. Six small pieces of poultry were sequentially coarsely ground, with a total coarse powder sample mass of 728 grams. 198 grams of the coarse powder sample from the left breast piece was used for stratified sampling and particle size distribution analysis, while the remaining coarse powder sample was stored for later use.

[0034] In some embodiments, the step of performing stratified sampling and particle size distribution detection on the coarse powder sample to obtain component distribution data and particle size distribution data includes: performing stratified sampling on the coarse powder sample to obtain samples of each layer; performing component analysis on each layer of samples to obtain the component ratio of each layer; extracting the distribution differences of feathers, bones and muscles from the component ratio of each layer to generate component distribution data; and performing particle size sieving and statistical analysis on the coarse powder sample to obtain particle size distribution data.

[0035] The coarse powder sample was sampled in three layers (top, middle, and bottom) to obtain samples for each layer. 198 grams of coarse powder sample were placed into a cylindrical sampling container 20 cm high and 8 cm in diameter. The container's inner wall was smooth and free of electrostatic coating to prevent sample adhesion. A free-stacking method was used during filling; the coarse powder sample was slowly poured in from the top of the container at a rate controlled at 50 g / min to avoid stratification disturbance. After filling, the coarse powder sample reached a height of approximately 16 cm within the container. During the stratification process, different density components of the coarse powder sample naturally stratified: the less dense feather fibers floated to the top, the denser bone fragments sank to the bottom, and the moderately dense muscle particles were distributed in the middle layer. Stratified sampling divided the coarse powder sample in the container into three equal layers (top, middle, and bottom), each approximately 5.3 cm high. The top layer corresponds to the top of the container (5.3 cm), the middle layer to 5.3–10.6 cm, and the bottom layer to 10.6–16 cm. The sampling tool used is a thin-walled stainless steel sampling spoon, 0.5 mm thick. During sampling, the spoon is inserted horizontally into the designated layer, with the insertion depth controlled at the center of the layer. If the sampling spoon is inserted too deeply into the lower layer, it will cause the upper layer sample to mix with the lower layer components, distorting the density data. Moving the sampling spoon horizontally to cover multiple locations within the layer ensures the representativeness of each layer's sample. The sample weight for each layer is 10 grams: 10 grams for each layer in the upper, middle, and lower layers. Each layer's sample is placed in a sample bag labeled with a layer number. The sampling location, time, and weight for each layer's sample are recorded in the sampling record sheet.

[0036] For example, the step of performing component analysis on each layer of the sample to obtain the component ratio of each layer includes: performing density detection on each layer of the sample to generate density data for each layer; identifying the boundary positions of the feather layer, bone layer and muscle layer based on the density data of each layer; measuring the thickness of the boundary positions to obtain the thickness of the boundary mixed layer; correcting the mass of each layer according to the thickness of the boundary mixed layer and calculating the component ratio of each layer.

[0037] Density data for each layer of the sample was generated by density testing. For the top layer, 10 grams of sample were placed in a graduated cylinder with a known volume (50 mL, 0.5 mL graduation). After loading, the sample was gently vibrated to allow it to settle naturally until it reached a stable state. The vibration frequency was 30 Hz, and the vibration time was 30 seconds. If the vibration time was too short (e.g., 10 seconds), the sample stacking would be unstable, resulting in large fluctuations in the volume reading. If the vibration time was too long (e.g., 2 minutes), fine particles would settle, leading to higher density data for each layer. After vibration, the volume occupied by the sample was determined by reading the graduated cylinder graduation. The volume reading for each top layer was 18 mL. Dividing the sample mass (10 g) by the volume (18 mL) yielded a density of 0.56 g / mL for the top layer. The middle layer, also 10 grams of sample, was measured using the same method, resulting in a volume reading of 15 mL and a density of 0.67 g / mL. The bottom layer, also 10 grams of sample, had a volume reading of 12 mL and a density of 0.83 g / mL. The density data for each layer shows an increasing trend from top to bottom. The differences in density data between layers reflect the stratified distribution of different components. The density of the feather component is approximately 0.3 g / mL, the density of the muscle component is approximately 0.9 g / mL, and the density of the bone component is approximately 1.8 g / mL. The density data of the upper layer (0.56 g / mL) is close to the mixed density of feathers and muscle, the density data of the middle layer (0.67 g / mL) is close to the uniform mixed density of the three components, and the density data of the lower layer (0.83 g / mL) is close to the mixed density of bone and muscle.

[0038] The boundaries between the feather layer, bone layer, and muscle layer are identified based on density data from each layer. The boundary between the upper and middle layers is determined by density thresholds for each layer. The feather layer is defined as a region with a density less than 0.60 g / mL. If the threshold is set too high (e.g., 0.70), some areas mixed with muscle may be misidentified as feather layers; if it is set too low (e.g., 0.50), the mixed layer of feathers and muscle cannot be identified. The muscle layer is defined as a region with a density between 0.60 and 0.75 g / mL, and the bone layer is defined as a region with a density greater than 0.75 g / mL. The upper layer, with an average density of 0.56 g / mL below the 0.60 g / mL threshold, is identified as the feather layer; the middle layer, with an average density of 0.67 g / mL falling within the 0.60 to 0.75 g / mL range, is identified as the muscle layer; and the lower layer, with an average density of 0.83 g / mL above the 0.75 g / mL threshold, is identified as the bone layer. The boundary between the feather layer and the muscle layer was located at a height of 5.3 cm in the container. At this boundary, the density data for each layer transitioned from 0.56 g / mL to 0.67 g / mL. The boundary between the muscle layer and the bone layer was located at a height of 10.6 cm in the container. At this boundary, the density data for each layer transitioned from 0.67 g / mL to 0.83 g / mL. Precise boundary location was achieved through subdivided sampling. Density was measured every 0.2 cm within a 1 cm range above and below the 5.3 cm height. The measurement results showed that the boundary location was confirmed as having the largest density gradient among the layers at the 5.3 cm boundary.

[0039] The thickness of the boundary mixed layer was obtained by measuring the thickness at the boundary location. At 5.3 cm from the boundary between the feather layer and the muscle layer, the component mixing state of the sample at this boundary location was observed under a microscope. The microscope magnification was 40x, and the observation field diameter was 2 mm. The simultaneous presence of feather fibers and muscle particles in the field of view indicated that this boundary location was a mixed layer. The thickness of the boundary mixed layer was determined by observing each layer progressively. Starting from the boundary location of 5.3 cm, observation sections were taken at intervals of 0.1 cm upwards. When the proportion of feather fibers in the field of view exceeded 90% at a height of 0.2 cm to 5.1 cm, it was determined to be a pure feather layer. When the proportion of muscle particles in the field of view exceeded 90% at a height of 0.2 cm to 5.5 cm downwards from the boundary location of 5.3 cm, it was determined to be a pure muscle layer. The thickness of the boundary mixing layer between the feather layer and the muscle layer ranged from 5.1 to 5.5 cm in height, with a thickness of 0.4 cm. Within this range, the ratio of feather fibers to muscle particles gradually transitioned from 90:10 to 10:90. The presence of the boundary mixing layer thickness indicates that the mixing of components of different densities was inevitable during the stratified stacking process. The thickness of the boundary mixing layer was directly proportional to the particle size; the larger the average particle size of the coarse powder sample, the thicker the boundary mixing layer. Measurements were taken at the boundary between the muscle layer and the bone layer at 10.6 cm using the same method. At this boundary location, both muscle particles and bone fragments were observed simultaneously within the field of view. The thickness of the boundary mixing layer at this location ranged from 10.4 to 10.8 cm in height, with a thickness of 0.4 cm as well.

[0040] The mass of each layer was corrected based on the thickness of the boundary mixing layer, and the component ratio of each layer was calculated. The sampling range of the upper layer was 0 to 5.3 cm. Considering the thickness of the boundary mixing layer of 0.4 cm, the boundary mixing layer occupied a height of 0.2 cm (5.1 to 5.3 cm) within the upper layer sampling range. The upper pure feather layer ranged from 0 to 5.1 cm, and the mixing layer ranged from 5.1 to 5.3 cm. From the 10-gram sample of the upper layer, 4.2 grams of feathers, 0.8 grams of bone, and 5.0 grams of muscle were actually sorted. Before correction, the component ratio of each layer in the upper layer was 42% feathers, 8% bone, and 50% muscle. The correction amount was calculated using the formula Δm=(d / H)×M×R, where d is the height of the boundary mixing layer within the sampling range of this layer, H is the total sampling height of this layer, M is the mass of the sample in this layer, and R is the proportion coefficient of the dominant component of the adjacent layer in the mixing layer. The correction amount Δm was deducted from the mass of the component mixed in from the adjacent layer and compensated for in the mass of the dominant component of this layer. In the upper layer correction calculation, d is 0.2 cm, H is 5.3 cm, M is 10 g, and R is 0.5 (feathers and muscle each account for 50% in the mixed layer). Δm = (0.2 / 5.3) × 10 × 0.5 is approximately 0.19 g. After correction, the upper layer muscle mass is 5.0 minus Δm equals 4.81 g, and the proportion of each muscle layer component is adjusted to 48%. After correction, the upper layer feather mass is 4.2 plus Δm equals 4.39 g, and the proportion of each feather layer component is adjusted to 44%. The skeleton does not participate in the mixing process within the thickness range of the boundary mixed layer, and the proportion of each skeleton layer component remains at 8%. After correction, the sum of the masses of the three components, 4.39 plus 0.8 plus 4.81 equals 10 g, which is consistent with the total sample mass. The middle layer is affected by the thickness of the boundary mixing layer at both the upper and lower boundaries. The upper boundary of the middle layer is mixed with feather components, and the lower boundary is mixed with bone components. After correction, the proportion of feather components in each layer of the middle layer is 16%, the proportion of bone components in each layer of the middle layer is 24%, and the proportion of muscle components in each layer of the middle layer is 60%. The lower layer is corrected according to the same formula based on the thickness of the boundary mixing layer. After correction, the proportion of feather components in each layer of the lower layer is 2%, the proportion of bone components in each layer of the lower layer is 58%, and the proportion of muscle components in each layer of the lower layer is 40%.

[0041] The distribution differences of feathers, bones, and muscles in each layer were extracted to generate component distribution data. The proportions of feathers in each layer (44%) were significantly higher than those in the middle layer (16%) and the bottom layer (2%), indicating a clear enrichment of feather components in the upper layer. The difference in component proportions between the upper and lower layers reached 42 percentage points. Within each layer, the proportion of bones increased from 8% in the upper layer to 24% in the middle layer and then to 58% in the lower layer, showing a lower layer enrichment of bone components. The difference in component proportions between the lower and upper layers reached 50 percentage points. The proportion of muscles was relatively evenly distributed across the three layers: 48% in the upper layer, 60% in the middle layer, and 40% in the lower layer, with interlayer differences in muscle components less than 20 percentage points. Component distribution data were obtained by calculating the concentration gradient of each component at different layers. The concentration gradient was defined as the difference between the proportions of components in the upper and lower layers divided by the total stacking height of the sample, which was 16 cm. The component distribution data for feathers was (44-2) / 16, approximately 2.6% / cm; for bones, approximately (58-8) / 16, approximately 3.1% / cm; and for muscles, approximately (60-40) / 16, approximately 1.25% / cm. The highest value for the bone component in the component distribution data indicates that the bones were most clearly separated during the stratification process, while the lowest value for the muscle component indicates that the muscles were relatively evenly distributed across the layers.

[0042] Particle size distribution data were obtained by sieving coarse powder samples. 50 grams of coarse powder sample were fed into a vibrating sieve. Before sieving, it was confirmed that there were no obvious agglomerates in the coarse powder sample, as the particle size range of the coarse powder sample directly affects the determination of ultrafine grinding process parameters. The sieve was equipped with four layers of sieves: 10 mesh, 20 mesh, 40 mesh, and 80 mesh, corresponding to apertures of 2 mm, 0.85 mm, 0.425 mm, and 0.18 mm, respectively. The vibration frequency was 50 Hz, the amplitude was 5 mm, and the sieving time was 15 minutes. After sieving, the samples retained on each sieve layer were weighed. 12 grams of sample were retained on the 10 mesh sieve, and the particle size distribution data showed that particles larger than 2 mm accounted for 24%, mainly consisting of bone fragments and insufficiently ground muscle particles. 15 grams of sample were retained on the 20 mesh sieve, and the particle size distribution data showed that particles between 0.85 and 2 mm accounted for 30%, a mixture of fine bone fragments and muscle particles. 13 grams of sample were retained on a 40-mesh sieve, with particle size distribution data showing that particles ranging from 0.425 to 0.85 mm accounted for 26%, primarily consisting of muscle particles and feather fiber segments. 7 grams of sample were retained on an 80-mesh sieve, with particle size distribution data showing that particles ranging from 0.18 to 0.425 mm accounted for 14%. 3 grams of sample passed through the 80-mesh sieve, with particle size distribution data showing that particles smaller than 0.18 mm accounted for 6%. The particle size distribution data indicates that the coarse powder sample has a wide particle size range, spanning nearly two orders of magnitude from 0.1 mm to 5 mm. The fact that 24% of the particles were larger than 2 mm suggests that even after coarse grinding, a significant number of large particles still require further grinding.

[0043] In some embodiments, determining the ultrafine grinding process parameters based on the component distribution data and the particle size distribution data includes: extracting the feather ratio, bone ratio, and muscle ratio from the component distribution data; extracting the particle size differences of each component from the particle size distribution data to generate component particle size characteristics; identifying difficult-to-grind components through the particle size distribution data and the component particle size characteristics to generate difficult-to-grind markers; and determining the ultrafine grinding process parameters based on the feather ratio, bone ratio, and muscle ratio combined with the difficult-to-grind markers.

[0044] The proportions of feathers, bones, and muscles were extracted from the component distribution data. The component distribution data included the proportions of each component in the upper, middle, and lower layers. The overall component proportions of the coarse powder sample were calculated using a weighted average, with the weighting factor representing the mass percentage of each layer. The upper layer (66 grams) accounted for 33.3% of the 198 grams of the coarse powder sample; the middle layer (66 grams) accounted for 33.3%; and the lower layer (66 grams) accounted for 33.3%. Since the three layers had equal masses, the weighted average was simplified to an arithmetic average. The feather proportion was obtained by the arithmetic average of the proportions of the three feather layers: upper feathers (44%) + middle feathers (16%) + lower feathers (2%), divided by 3, approximately 20.7%. The bone proportion was obtained by dividing the upper bone (8%) + middle bone (24%) + lower bone (58%) by 3, approximately 30%. The muscle proportion was obtained by dividing the upper muscle (48%) + middle muscle (60%) + lower muscle (40%) by 3, approximately 49.3%. The component distribution data revealed that feathers accounted for approximately 20.7%, bone 30%, and muscle 49.3%. In the coarse powder sample, muscle had the highest proportion, approaching 50%, followed by bone at 30%, and feathers the lowest at approximately 21%. The sum of the three component proportions (20.7%, 30%, and 49.3%) equals 100%, validating the consistency of component proportions across the three layers in the component distribution data. The component identification results reflected in the component distribution data correspond to the anatomical structure of small bird body pieces. Muscle tissue constitutes the majority of the dry weight of these pieces, while bone and feather residues account for a smaller proportion. The feather proportion of approximately 20.7% is higher than the expected value of 15% based on bird anatomy, indicating that the feather component was not fully pulverized during coarse grinding, retaining larger fiber segments. The bone proportion of 30% is close to the expected value, while the muscle proportion of 49.3% is slightly lower than the expected value of 52%.

[0045] The particle size differences of each component were extracted from the particle size distribution data to generate component particle size characteristics. In the component distribution data, the dominant component of the upper layer sample was feathers. Sieving 10 grams of the upper layer sample through an 80-mesh sieve revealed that fine particles smaller than 0.18 mm accounted for only 0.5 grams (5%), while coarse particles larger than 2 mm accounted for 4 grams (40%), indicating a predominantly coarse particle size distribution in the upper layer. In the component distribution data, the dominant component of the middle layer sample was muscle. Sieving results showed that particles smaller than 0.18 mm accounted for 12%, and particles larger than 2 mm accounted for 18%, indicating a relatively uniform particle size distribution in the middle layer. In the component distribution data, the dominant component of the lower layer sample was bone. Sieving results showed that particles smaller than 0.18 mm accounted for 3%, and particles larger than 2 mm accounted for 55%, indicating a predominantly coarse particle size distribution in the lower layer. The particle size characteristics of the components were obtained by calculating the proportion of coarse and fine particles in each layer. The upper layer, with 40% coarse particles, corresponds to the particle size characteristics of the feather component; the middle layer, with 18% coarse particles, corresponds to the particle size characteristics of the muscle component; and the lower layer, with 55% coarse particles, corresponds to the particle size characteristics of the bone component. The particle size characteristics of the feather component are fibrous with a high proportion of coarse particles, the bone component is flaky with an extremely high proportion of coarse particles, and the muscle component is granular with a uniform particle size distribution. The differences in the particle size characteristics of the three components provide a basis for subsequent identification of difficult-to-grind components.

[0046] Difficult-to-grind components were identified and labeled using particle size distribution data and component particle size characteristics. Particle size distribution data showed that coarse particles larger than 2 mm accounted for 24%, primarily bone fragments. Component particle size characteristics revealed that the proportion of coarse particles in the bone component (55%) was significantly higher than that in the feather component (40%) and muscle component (18%). Comparative analysis of particle size distribution data and component particle size characteristics indicated that the bone component, due to its high hardness and dense structure, was difficult to fully pulverize during coarse grinding. The Mohs hardness of hydroxyapatite crystals in bone is approximately 5, far exceeding the soft tissue hardness of muscle tissue. This causes the impact energy of the hammer mill to be preferentially consumed in bone deformation rather than breakage. A large number of bone fragments remained in the particle size range of 2 mm or larger in the particle size distribution data, thus the bone component was identified as a difficult-to-grind component. Although the feather component showed a relatively high proportion of coarse particles (40%) in its component particle size characteristics, the flexibility of feather fibers made them easily sheared and torn during ultrafine grinding, therefore the feather component was not identified as a difficult-to-grind component. Among the particle size characteristics of the muscle component, coarse particles accounted for the lowest proportion (18%) and had a uniform particle size distribution. The muscle tissue had already achieved a good degree of pulverization after coarse grinding, and the muscle component was not identified as a difficult-to-grind component. The difficult-to-grind marker was recorded using Boolean labels. The difficult-to-grind marker for the bone component was TRUE, while the difficult-to-grind markers for the feather and muscle components were both FALSE. The difficult-to-grind marker recorded information about the component type that required special treatment.

[0047] For example, determining the ultrafine grinding process parameters based on the feather ratio, bone ratio, and muscle ratio in conjunction with the difficult-to-grind marker includes: querying the recommended grinding time for each component based on the feather ratio, bone ratio, and muscle ratio; determining a temperature monitoring threshold based on the difficult-to-grind marker; performing real-time temperature monitoring during the grinding process based on the temperature monitoring threshold to identify over-temperature points and generate intermittent cooling markers; and comprehensively determining the ultrafine grinding process parameters based on the recommended grinding time for each component, the intermittent cooling markers, and the difficult-to-grind markers.

[0048] Recommended grinding times for each component were queried separately for feather, bone, and muscle proportions. For a feather proportion of approximately 20.7%, the recommended grinding time for each component in the component processing database showed a linear relationship with its proportion: 20% feather corresponded to a recommended grinding time of 8 minutes, 21% to 8.5 minutes, and approximately 20.7% feather yielded a recommended grinding time of approximately 8.35 minutes through linear interpolation. The component processing database stores recommended grinding times for different components at various proportions. The database was built based on extensive experimental test results, covering feather proportions from 5% to 40%, bone proportions from 10% to 50%, and muscle proportions from 30% to 70%. For a bone proportion of 30%, the database showed a recommended grinding time of 15 minutes. This timeframe takes into account the high hardness of the bone component, requiring longer repeated impacts. Due to its dense inorganic structure, the bone component breaks down more slowly under airflow impact; therefore, the recommended grinding time for the bone proportion is the longest among the three components. With a muscle content of 49.3%, the recommended grinding time for each component was 10 minutes. A muscle content close to 50% means nearly half the sample is muscle. The softness and porous structure of muscle tissue result in a relatively short grinding time. The recommended grinding time for each component with the muscle content is the shortest among the three components. Muscle fibers are easily torn along their fiber direction under high-speed airflow, forming fine particles. The recommended grinding time for each component reflects the shortest time required for that component to reach the target particle size, which is set at over 80% of particles smaller than 50 micrometers.

[0049] The temperature monitoring threshold is determined based on the "difficult-to-grind" marker. The "TRUE" marker indicates the presence of difficult-to-grind bone components in the sample. Bone components require higher airflow impact energy during ultrafine grinding, leading to increased heat generation. The frictional and deformation heat generated by the high-speed collision of bone particles with the grinding chamber wall is significantly higher than that of muscle and feather components. The temperature monitoring threshold corresponding to the "difficult-to-grind" marker is determined based on the proportion of the difficult-to-grind component. A 30% bone component proportion corresponds to a temperature monitoring threshold reduction factor of 0.85. Multiplying the standard temperature monitoring threshold of 30℃ by the reduction factor yields an adjusted temperature monitoring threshold of 25.5℃. If the "difficult-to-grind" marker is "FALSE," the temperature monitoring threshold remains at the standard value of 30℃ without adjustment. Lowering the temperature monitoring threshold ensures that the sample temperature does not become excessively high during bone component grinding, thus preventing changes in nutritional composition. Unsaturated fatty acids and heat-sensitive vitamins in poultry meat are prone to oxidative degradation when exposed to temperatures above 30℃ for extended periods, causing the fatty acid profile and vitamin content of the tested sample to deviate from the original poultry levels. The temperature monitoring threshold of 25.5℃ serves as the upper limit for temperature control during the grinding process. When the grinding chamber temperature exceeds this threshold, an intermittent cooling procedure is triggered. Temperature monitoring uses an infrared temperature sensor, which is installed on the side wall of the grinding chamber. The measurement range is -20 to 100℃, the response time is 1 second, and the sampling frequency is 1 Hz. The sensor monitors the airflow temperature in the grinding chamber in real time. The temperature signal is transmitted to the temperature controller, which compares the measured temperature with the temperature monitoring threshold of 25.5℃. When the measured temperature exceeds the temperature monitoring threshold, an over-temperature alarm signal is output.

[0050] Based on a temperature monitoring threshold, real-time temperature monitoring is performed during the pulverization process to identify over-temperature points and generate intermittent cooling markers. At the start of the pulverization process, the pulverization chamber temperature is room temperature (20°C). The high-speed impact of the pulverizing airflow on the sample generates frictional and deformation heat, causing the pulverization chamber temperature to gradually rise. At the 3rd minute, the temperature reaches 23°C; at the 5th minute, it reaches 25°C; and at the 7th minute, it reaches 27°C, exceeding the temperature monitoring threshold of 25.5°C. The temperature controller identifies this as an over-temperature point and triggers the intermittent cooling program. The intermittent cooling marker records the time and temperature value of this over-temperature point. After the intermittent cooling marker is triggered, the pulverizer automatically stops operating and enters cooling mode. In cooling mode, liquid carbon dioxide is introduced into the pulverization chamber for cooling. The liquid carbon dioxide enters the cooling jacket on the outer wall of the pulverization chamber from a cryogenic storage tank via a flow control valve. The storage tank temperature is -20°C and the pressure is 2.0 MPa. The liquid carbon dioxide absorbs heat and vaporizes within the jacket. The latent heat of vaporization of liquid carbon dioxide is approximately 571 kJ / kg, enabling efficient cooling even at a low flow rate. The cooling time is 3 minutes. After 3 minutes, the pulverization chamber temperature drops to 18°C, below the temperature monitoring threshold. After cooling was completed, the pulverizer resumed operation. At the 10th minute, the temperature rose again to 26℃, exceeding the temperature monitoring threshold, and an intermittent cooling marker was generated again, initiating intermittent cooling. A total of four over-temperature points were identified during the pulverizing process, at the 7th, 10th, 13th, and 16th minutes. The intermittent cooling marker recorded a total of four cooling events, with each cooling event lasting 3 minutes. The total cooling time corresponding to the four intermittent cooling markers was 12 minutes.

[0051] The ultrafine grinding process parameters were determined based on the recommended grinding time for each component, intermittent cooling markers, and difficult-to-grind markers. The recommended grinding time for each component was calculated as a weighted average to obtain the overall grinding time, with the weights being the proportions of each component. The overall grinding time T = T_feather × P_feather + T_skeleton × P_skeleton + T_muscle × P_muscle. Substituting the recommended grinding time and proportions for each component, we get T = 8.35 × 0.207 + 15 × 0.30 + 10 × 0.493, approximately 11.2 minutes. Among the recommended grinding times for each component, the skeletal component (15 minutes) contributed the most to the overall grinding time, at 4.5 minutes. The intermittent cooling markers indicated four intermittent cooling events during the grinding process, each lasting 3 minutes. The total cooling time corresponding to the intermittent cooling markers was 12 minutes. The total process time was the overall grinding time plus the cooling time corresponding to the intermittent cooling markers, totaling 23.2 minutes. The intermittent cooling markers extended the actual process time by approximately 107%. The TRUE marking of difficulty in pulverization necessitates an increase in pulverization pressure. Multiplying the standard pressure of 0.7 MPa by the pressure enhancement factor of 1.286 corresponding to the difficulty-to-pulverize marking yields an adjusted pulverization pressure of 0.9 MPa. This pressure corresponds to higher impact energy, effectively pulverizing bone components. The airflow velocity corresponding to the difficulty-to-pulverize marking is increased from the standard 150 m / s to 200 m / s, resulting in an increase of approximately 78% in particle collision kinetic energy, which is proportional to the square of the velocity. The final ultrafine pulverization process parameters are determined as follows: pulverization pressure 0.9 MPa, pulverization run time 11.2 minutes, 4 intermittent cooling cycles, total process time 23.2 minutes, and temperature monitoring threshold 25.5℃. These five parameters constitute a complete set of ultrafine pulverization process parameters, determined based on the recommended pulverization time for each component, the intermittent cooling marking, and the difficulty-to-pulverize marking.

[0052] Based on the ultrafine grinding process parameters, coarse powder samples were subjected to low-temperature ultrafine grinding with liquid carbon dioxide assistance to obtain ultrafine powder samples, and carbon dioxide was recovered. For example... Figure 2As shown. A 198g coarse powder sample was pre-cooled before ultrafine grinding. The sample was placed in a -20℃ freezer for 30 minutes to lower the temperature to -5℃, increasing its brittleness. The pre-cooled sample was then fed into the air jet mill 1 through inlet 8. The working air pressure of mill 1 was set to 0.9 MPa according to the ultrafine grinding process parameters. This pressure was precisely controlled by an air compressor and a pressure regulating valve, with a pressure fluctuation range of ±0.02 MPa. The air jet mill 1 uses a counter-jet grinding structure. Two opposing counter-jet nozzles 3 eject high-speed airflow that converges at the center of the grinding chamber 2, causing sample particles to collide and pulverize in the convergence area. A cooling jacket 4 is installed on the outer wall of the grinding chamber 2. Liquid carbon dioxide enters the cooling jacket 4 from the liquid carbon dioxide storage tank 5 through the first flow control valve 6 to achieve a low-temperature grinding environment. Under standard operating conditions, the liquid carbon dioxide flow rate is 0.5 liters / minute. The pulverization process was executed according to the ultrafine pulverization process parameters. After 2 minutes of operation, temperature sensor 7 detected a temperature of 23°C in pulverization chamber 2. At the 7-minute mark, the temperature rose to 27°C, exceeding the monitoring threshold of 25.5°C in the ultrafine pulverization process parameters, at which point pulverizer 1 automatically stopped and entered intermittent cooling. During intermittent cooling, the liquid carbon dioxide flow rate was increased to 1.2 liters / minute for rapid cooling. After 3 minutes, the temperature dropped to 18°C, and pulverizer 1 resumed operation. The second, third, and fourth intermittent cooling cycles occurred at the 10th, 13th, and 16th minutes, respectively. The total pulverization operation time, calculated according to the ultrafine pulverization process parameters, was 11.2 minutes, plus 12 minutes for the four intermittent cooling cycles, for a total process time of 23.2 minutes. During the pulverization process, the carbon dioxide gas vaporized in the cooling jacket 4 is collected through the gas recovery pipe 10. The gas recovery pipe 10 transports the vaporized gas to the dehumidification filter 11 to remove moisture and dust particles from the gas. The gas treated by the dehumidification filter 11 enters the compressor unit 12 for compression. The compressed carbon dioxide is stored in the recovery storage tank 13. The pulverization process consumes approximately 8 liters of liquid carbon dioxide, with a recovery rate of 92%. The carbon dioxide purity in the recovery storage tank 13 is above 98.5%. After the ultrafine pulverization is completed, the ultrafine powder sample is collected from the discharge port 9. The collected ultrafine powder sample weighs 190 grams, with a pulverization loss rate of 4%, mainly from the fine dust collected by the bag filter. The particle size distribution of the ultrafine powder sample is determined by a laser particle size analyzer. The average particle size is 35 micrometers, and the proportion of particles with a particle size of less than 50 micrometers is 85%, meeting the expected target. The ultrafine powder sample is a light brown, fine powder. The powder has no grainy feel when rubbed between the fingers. The ultrafine powder sample is placed in a sealed bottle for later use.

[0053] Step S140: Homogenize the ultrafine powder sample to obtain a mixed sample. Detect the oxygen content and uniformity of the mixed sample to obtain residual oxygen concentration data and uniformity data. Determine the carbon dioxide recovery filling parameters based on the residual oxygen concentration data and uniformity data. Seal the mixed sample according to the carbon dioxide recovery filling parameters to obtain a pre-treated sample.

[0054] Specifically, the ultrafine powder sample is homogenized to obtain a mixed sample. 190 grams of the ultrafine powder sample is transferred from a sealed bottle to a V-type mixer for homogenization. Before transfer, the ultrafine powder sample must be confirmed to be in good condition with no obvious agglomeration. The average particle size of the ultrafine powder sample is 35 micrometers, suitable for homogenization by the V-type mixer. The V-type mixer has a mixing cylinder volume of 500 ml, and the cylinder consists of two cylindrical cylinders with a 60-degree angle. The V-shaped structure causes the sample to undergo splitting, merging, and shearing during rotation. The mixer speed is set to 25 rpm. At this speed, the sample can tumble sufficiently within the cylinder without centrifugal stratification. If the speed is too high (e.g., 50 rpm), centrifugal force causes fine particles to concentrate towards the cylinder wall and coarse particles to concentrate in the center, resulting in false mixing and causing deviations in the component ratio at different locations during sampling and testing. If the speed is too low (e.g., 10 rpm), the sample will not tumble sufficiently, and the mixing time needs to be extended to more than 60 minutes. The mixing time was set to 20 minutes, determined based on the average particle size of the ultrafine powder sample (35 micrometers) and the mixing drum filling rate (38%). If the filling rate was too high (e.g., 80%), the sample would not tumble sufficiently within the drum, reducing mixing efficiency by more than 50%. If the filling rate was too low (e.g., 10%), the sample would form a thin layer on the drum wall, resulting in rapid mixing but low batch throughput. The mixing process was conducted under a carbon dioxide protective environment. Before loading, carbon dioxide (99.9% purity) was introduced into the mixing drum to replace the internal air, and the replacement process lasted for 3 minutes to ensure the oxygen content inside the drum dropped below 1%. After mixing, the mixed sample was discharged from the mixer's discharge port and collected in a stainless steel tray. The mixed sample weighed 189 grams, with a mixing loss rate of 0.5%, mainly due to adhesion to the drum wall and discharge residue. The mixed sample particles exhibited a uniform light brown color without significant color variations; color uniformity is a direct indicator of mixing effectiveness. The mixed sample was then sieved through a 100-mesh sieve to remove any potential agglomerated particles. After sieving, the mixed sample weighed 188 grams.

[0055] In some embodiments, the step of detecting oxygen content and uniformity of the mixed sample to obtain residual oxygen concentration data and uniformity data includes: detecting oxygen content in the upper, middle and lower layers of the mixed sample to obtain oxygen concentration values ​​for each layer; sampling from multiple points in the mixed sample to detect component differences and generate component difference values; determining the maximum value of the oxygen concentration values ​​of each layer as residual oxygen concentration data; and generating uniformity data based on the component difference values.

[0056] Oxygen content was measured in the top, middle, and bottom layers of the mixed sample to obtain the oxygen concentration values ​​for each layer. 188 grams of the mixed sample were placed into a cylindrical detection container 15 cm high and 10 cm in diameter. The container was made of transparent acrylic glass to facilitate observation of the layered state of the mixed sample. After filling, the mixed sample piled up to a height of approximately 12 cm in the container, with a smooth surface without significant unevenness. Oxygen content detection divided the mixed sample into three layers based on height: the top layer corresponds to a height range of 0-4 cm, the middle layer to 4-8 cm, and the bottom layer to 8-12 cm. The equal height of the three layers ensured the comparability of the detection data. The oxygen concentration probe used an electrochemical sensor. The sensor's measurement principle is based on the current signal generated by the oxidation-reduction reaction of oxygen on the electrode surface; the current intensity is directly proportional to the oxygen concentration. The probe is 5 mm in diameter and 8 cm in length. The probe tip is equipped with a porous, breathable membrane that allows oxygen to diffuse to the electrode surface but prevents sample particles from entering. For the upper layer detection, the probe is inserted vertically into the center of the container to a depth of 2 cm. After insertion, it is left to stand for 10 seconds until the sensor response stabilizes. If the reading is taken directly without allowing it to stand, the mechanical disturbance to the sample during probe insertion will create a local hypoxic zone around the probe, potentially underestimating the measured value by 0.2 to 0.3 percentage points. After stabilization, the oxygen concentration values ​​for each layer are read. The oxygen concentration value for each layer in the upper layer is 0.8%. For the middle layer detection, the probe is inserted to a depth of 6 cm, and the oxygen concentration value for each layer in the middle layer is 0.6%. For the lower layer detection, the probe is inserted to a depth of 10 cm, and the oxygen concentration value for each layer in the lower layer is 0.4%. The oxygen concentration values ​​show a decreasing trend from top to bottom. This is because the density of carbon dioxide (1.98 kg / m³) is greater than the density of air (1.29 kg / m³), and carbon dioxide settles to the bottom under gravity, forming an oxygen concentration gradient.

[0057] Component difference values ​​were generated by sampling multiple points from a mixed sample to detect component differences. The surface of the mixed sample was divided into nine sampling points using a 3×3 grid with a grid spacing of 3 cm. The sampling points were numbered P1 to P9 from top left to bottom right. Multi-point sampling of the mixed sample comprehensively reflects the spatial differences in component distribution. A micro-sampling spoon with a volume of 0.5 ml was used as the sampling tool. Each sampling point extracted approximately 0.3 g of sample from the mixed sample, and the sampling depth was controlled within the sample surface range of 0 to 1 cm. The samples from the nine sampling points were placed into numbered sample tubes, which were then sealed with carbon dioxide for component analysis. Component analysis was performed using a near-infrared spectroscopy analyzer with a wavelength range of 1000 to 2500 nm. Components such as proteins, fats, and water in the sample exhibit characteristic absorption peaks at different wavelengths. The mixed sample in each sample tube was transported to the analyzer's measurement chamber via an autosampler. The measurement chamber temperature was controlled at a constant 25℃ to eliminate the influence of temperature on the spectrum. If the temperature fluctuation exceeds ±2℃, the absorption peak position of moisture will shift by approximately 5 nanometers, resulting in a moisture content measurement error of ±0.5%. The results from sampling point P1 showed a protein content of 62.3%, a fat content of 18.5%, and a moisture content of 8.2%. Sampling points P2 through P9 were measured sequentially to obtain their respective component data. The component difference value was obtained by calculating the ranges of protein content, fat content, and moisture content across the nine sampling points. The maximum protein content was 62.8%, the minimum was 61.9%, and the range was 0.9 percentage points. The ranges for fat content and moisture content were 0.7 and 0.5 percentage points respectively. The component difference value was defined as the weighted average of the ranges of the three components, with a weight of 0.5 for protein, 0.3 for fat, and 0.2 for moisture, resulting in a component difference value of approximately 0.76 percentage points.

[0058] The maximum oxygen concentration value for each layer was determined as the residual oxygen concentration data. Each layer's oxygen concentration value includes three values: 0.8% for the upper layer, 0.6% for the middle layer, and 0.4% for the lower layer. The upper layer value of 0.8% is the highest, while the middle and lower layer values ​​are 0.2 and 0.4 percentage points lower than the upper layer, respectively. The residual oxygen concentration data was determined using the maximum value principle, based on the "barrel effect" of oxidation reactions. The part with the highest oxygen concentration in the mixed sample determines the overall oxidation risk. Even if the oxygen concentration value of the lower layers is only 0.4%, far below the target threshold, the high oxygen concentration of 0.8% in the upper layer can still cause oxidation of unsaturated fatty acids in that region of the sample. The residual oxygen concentration data is taken as the maximum value of each layer, i.e., the upper layer's 0.8%. The residual oxygen concentration data of 0.8% is compared with the target residual oxygen threshold of 0.5%. If the residual oxygen concentration data exceeds the target threshold by 0.3 percentage points, it indicates that the mixed sample needs to be filled with carbon dioxide to reduce the oxygen content. The degree to which the residual oxygen concentration exceeds the standard determines the amount of carbon dioxide required for filling; the greater the exceedance of the residual oxygen concentration, the higher the required filling and replacement ratio. The residual oxygen concentration also reflects the oxygen adsorption of the mixed sample during homogenization. Ultrafine powder samples have a strong gas adsorption capacity due to their large specific surface area. Oxygen adsorbed on the sample surface is more difficult to replace with carbon dioxide than free oxygen. The decreasing trend of oxygen concentration from top to bottom indicates that carbon dioxide, with a density of 1.98 kg / m³ greater than air density of 1.29 kg / m³, settles to the bottom under gravity. The lower carbon dioxide concentration in the upper region leads to a higher residual oxygen concentration in that region.

[0059] Uniformity data is generated based on the component difference value. A component difference value of 0.77 percentage points is used to calculate the uniformity data using the uniformity conversion formula: Uniformity data = 100% - Component difference value × Coefficient, with the coefficient set to 80 (unit: % / percentage point). This coefficient is based on extensive experimental calibration; a component difference value of 1.25 percentage points corresponds to 0% uniformity data (completely non-uniform), and a component difference value of 0% corresponds to 100% uniformity data (ideal uniformity). The uniformity data is approximately 38.4%, indicating that the component distribution uniformity of the mixed sample reaches the ideal level of 38.4%. The uniformity data is below the target uniformity threshold of 95%, with a difference of 56.6 percentage points, indicating that the homogenization effect of the mixed sample has not met the expected requirements. The reasons for the low uniformity data may include insufficient mixing time, improper mixing speed, or excessively wide particle size distribution of the sample leading to stratification. The uniformity data is used to determine whether secondary mixing is required. When the uniformity data is below 90%, the mixed sample must be returned to the mixer for supplementary mixing; when the uniformity data is below 95%, supplementary mixing is recommended. The homogeneity of the mixed sample was 38.4%, which is below 90%, triggering a second mixing procedure. The mixed sample was reloaded into the V-type mixer, and the mixing speed was increased to 30 rpm, with the mixing time extended to 30 minutes. After the second mixing, multi-point sampling and testing were performed again. The new component difference value decreased to 0.18 percentage points, and the homogeneity improved to 85.6%. Since the homogeneity of 85.6% was still below the target threshold of 95%, a third mixing was performed. After the third mixing, the homogeneity reached 96.2%, meeting the requirements.

[0060] In some embodiments, determining the carbon dioxide recovery filling parameters based on the residual oxygen concentration data and the uniformity data includes: comparing the residual oxygen concentration data with a target residual oxygen threshold to determine the initial filling amount; performing threshold comparison on the uniformity data to identify low-uniformity areas and generating supplementary filling markers; conducting a filling effect test based on the initial filling amount to determine the residual oxygen value after filling; and determining the secondary filling amount based on the residual oxygen value after filling and the supplementary filling markers to generate the carbon dioxide recovery filling parameters.

[0061] The initial filling volume was determined by comparing the residual oxygen concentration data with the target residual oxygen threshold. The difference between the residual oxygen concentration data of 0.8% and the target residual oxygen threshold of 0.5% is 0.3 percentage points, indicating the extent to which the oxygen concentration in the mixed sample needs to be reduced. The initial filling volume was determined based on the difference between the residual oxygen concentration data and the target threshold, as well as the mass of the mixed sample. A 188g mixed sample occupies approximately 300mL of space within the container, with a total container volume of 1200mL and a free space volume of 900mL above the mixed sample. The oxygen in the free space comes from two sources: air introduced during filling and oxygen desorbed from the surface of the mixed sample. The goal of the initial filling was to reduce the oxygen concentration in the free space from the current 21% (oxygen concentration in the air) to below the target concentration of 0.5%, while simultaneously replacing adsorbed oxygen on the surface of the mixed sample. The filling gas uses carbon dioxide recovered through the S130 pulverization process, with a purity of 98.5%. Carbon dioxide, being denser than air, facilitates sedimentation and displaces residual oxygen at the bottom, while also possessing antibacterial properties, extending the effective detection period of the sample. The initial filling volume is calculated using the replacement ratio, defined as the ratio of the volume of carbon dioxide added to the volume of free space. Complete replacement requires a replacement ratio of at least 5 times. If the replacement ratio is only 3 times, the residual oxygen concentration can only be reduced to about 2%. If the replacement ratio reaches 8 times, the residual oxygen concentration can be reduced to 0.3%, but gas consumption increases by 60%, resulting in poor economic efficiency. The initial filling volume is determined to be 4500 ml, corresponding to a replacement ratio of 5 times, a filling pressure of 0.05 MPa, and a filling flow rate of 200 ml / min. During filling, carbon dioxide is introduced through the inlet at the top of the container, equipped with a diffuser plate to ensure uniform carbon dioxide distribution and prevent localized sample dispersion. The filling process lasts 22.5 minutes. After filling is completed, the air inlet and outlet are immediately sealed to create a slightly positive pressure environment inside the container to prevent external air from flowing back.

[0062] Threshold comparison was performed on the homogeneity data to identify low-uniformity areas and generate supplementary filling markers. The homogeneity data of 96.2% was compared with the target homogeneity threshold of 95%. The homogeneity data exceeding the threshold by 1.2 percentage points indicates that the overall homogeneity of the mixed sample meets the requirements. Although the overall homogeneity data meets the standard, it is still necessary to identify the existence of local low-uniformity areas, which refer to component differences at certain specific locations exceeding the overall average level. Low-uniformity area identification was achieved by analyzing the component data of 9 sampling points. The protein content of the 9 sampling points was compared with the average value. Sampling points with deviations exceeding ±0.5% of the average value were marked as low-uniformity areas. The protein content of sampling point P3 (62.9%) was 0.5 percentage points higher than the average of 62.4%, while the protein content of sampling point P7 (61.8%) was 0.6 percentage points lower than the average value. Sampling points P3 and P7 were identified as low-uniformity areas. The supplementary filling markers record the location coordinates and deviation range of the low uniformity area. The supplementary filling marker for sampling point P3 is "P3:+0.5%", and the supplementary filling marker for sampling point P7 is "P7:-0.6%". The supplementary filling markers are used to guide local treatment during the secondary filling process. The low uniformity area adopts a directional filling method during the secondary filling. The location information recorded in the supplementary filling markers guides the adjustable nozzles to blow carbon dioxide directionally towards the low uniformity area to promote local convection mixing.

[0063] The residual oxygen value after filling was determined by testing the filling effect based on the initial filling volume. After the initial filling volume of 4500 ml, the mixture was allowed to stand for 5 minutes to allow the gas distribution within the container to reach equilibrium. During this standing period, carbon dioxide diffused into the mixed sample driven by the concentration gradient, and the oxygen adsorbed on the sample surface gradually desorbed into the gas phase. The filling effect test used an oxygen concentration probe to re-detect the oxygen content of the upper, middle, and lower layers of the mixed sample. The testing method was the same as the initial test, with the probe insertion position and depth remaining consistent. The residual oxygen value after filling the upper layer was 0.9%, which was higher than the 0.8% before the initial filling. The temporary increase in the residual oxygen value after filling was due to the airflow disturbance corresponding to the initial filling volume promoting the desorption of oxygen adsorbed on the sample surface, resulting in a temporary increase in the gas phase oxygen concentration. The pulse mode of the secondary filling can effectively remove the desorbed oxygen. The residual oxygen value after filling the middle layer was 0.5%, a decrease of 0.1 percentage points from the 0.6% before the initial filling. The replacement effect of the initial filling volume was limited in the middle layer region. The residual oxygen value after the lower layer filling was 0.3%, a decrease of 0.1 percentage points from 0.4% before the initial filling. The maximum residual oxygen value after filling was 0.9% in the upper layer, which was 0.4 percentage points higher than the target residual oxygen threshold of 0.5%, indicating that the initial filling volume failed to achieve the expected oxygen concentration reduction effect.

[0064] The secondary filling volume and recycled carbon dioxide filling parameters were determined based on the residual oxygen value after filling and the supplementary filling marker. The difference between the residual oxygen value after filling (0.9%) and the target residual oxygen threshold (0.5%) was 0.4 percentage points. The widening difference between the residual oxygen value after filling and the target threshold indicates that adsorbed oxygen desorption is the main oxygen source, requiring adjustment of the filling strategy. The determination of the secondary filling volume comprehensively considers both the residual oxygen value after filling and the supplementary filling marker. The residual oxygen value after filling determines the total amount required for secondary filling, while the supplementary filling marker determines the spatial distribution strategy for secondary filling. Pulsed filling is used, where intermittent high-flow-rate carbon dioxide injection generates airflow disturbance, which promotes the desorption of adsorbed oxygen on the sample surface and gas-phase mixing. The secondary filling volume was set at 6000 ml, corresponding to a replacement ratio of approximately 6.7 times, higher than the 5 times ratio of the initial filling. The secondary filling volume is greater than the initial filling volume because the desorption of adsorbed oxygen increases the additional oxygen load. The filling pressure was increased to 0.08 MPa, and the filling flow rate was increased to 500 mL / min. The high flow rate generated gas kinetic energy, which promoted gas-solid mass transfer on the sample surface. Pulse filling adopted an intermittent mode of 3 seconds filling and 2 seconds pausing. If continuous filling and uninterrupted airflow were used to purge, a stable gas film would form on the sample surface, hindering the desorption of adsorbed oxygen. The intermittent mode improved the desorption efficiency by 40% due to the periodic rupture of the gas film. For the low-uniformity areas P3 and P7 indicated by the supplementary filling marks, the secondary filling was equipped with an adjustable nozzle. The nozzle angle was adjusted to point towards the P3 and P7 positions, and the directional airflow generated local convection in the low-uniformity areas, enhancing the carbon dioxide replacement effect in these areas. The final parameters for carbon dioxide recovery filling were determined as follows: initial filling volume 4500 ml (filling pressure 0.05 MPa, filling flow rate 200 ml / min), secondary filling volume 6000 ml (filling pressure 0.08 MPa, filling flow rate 500 ml / min), pulse mode 3 seconds filling 2 seconds stopping, directional filling angle 30 degrees in the P3 direction and 210 degrees in the P7 direction. The carbon dioxide recovery filling parameters constitute a complete set of filling process parameters, which are determined based on the residual oxygen value after filling and the supplementary filling mark.

[0065] Based on the carbon dioxide filling parameters, the mixed sample is sealed to obtain the pre-treated sample, such as... Figure 2As shown. After the mixed sample 15 is loaded into the filling container 14, the carbon dioxide in the recovery storage tank 13 controls the flow rate through the supply gas pipeline 20 and the second flow control valve 21, and enters the filling container 14 from the air inlet 16. The initial filling volume of 4500 ml and the secondary filling volume of 6000 ml in the recovered carbon dioxide filling parameters are executed in sequence. For the initial filling, the filling pressure is set at 0.05 MPa and the filling flow rate is 200 ml / min according to the recovered carbon dioxide filling parameters. The recovered carbon dioxide during the initial filling is evenly distributed in the upper space of the mixed sample 15. After the filling is completed, the oxygen concentration probe 19 is used to detect the oxygen content of the upper, middle and lower layers of the mixed sample 15 respectively to verify the filling effect. The oxygen concentration probe 19 is inserted into different layers of the filling container 14 for measurement. When the residual oxygen value in the upper layer rises to 0.9%, it indicates the desorption of adsorbed oxygen. During the secondary filling, the pulse mode and the directional filling angle in the recovered carbon dioxide filling parameters are applied. The filling pressure is increased to 0.08 MPa and the filling flow rate is increased to 500 ml / min. The angle of the adjustable nozzle 18 is adjusted and pointed to the low uniformity area. The directional filling is achieved by using the pulse mode of filling for 3 seconds and pausing for 2 seconds. The unfilled gas is discharged from the exhaust port 17. After the two fillings, the oxygen concentration probe 19 verifies that the residual oxygen values of the three layers are respectively reduced to 0.4%, 0.3%, and 0.2%, all of which are lower than the target residual oxygen threshold of 0.5%, and the filling effect meets the requirements. Immediately after the filling effect is verified, the mixed sample 15 is sealed. The sealing treatment uses the aluminum foil composite film heat sealing technology. The composite film consists of three layers: an outer polyester film, a middle aluminum foil, and an inner polyethylene film. The aluminum foil provides oxygen and moisture barrier properties, and the polyethylene film provides heat sealing properties. A total of 188 g of the mixed sample 15 is divided into 10 sealed bags, with each bag filled with 18.8 g. The filling process is carried out in a carbon dioxide protection glove box to avoid exposing the mixed sample to the air environment. The filled sealed bags are sealed by a heat sealer. The heat sealing temperature is 180°C, the heat sealing pressure is 0.3 MPa, and the heat sealing time is 3 seconds. Too low temperature results in insufficient sealing strength, and too high temperature causes the polyethylene film to burn through. After heat sealing, a sealing strip with a width of 5 mm is formed. The sealing strength of the sealing strip is verified by a tensile test. A tensile force of 20 N is applied for 10 seconds in the tensile test. If the sealing strip has no breakage or detachment, it is judged as sealed qualified. After the pre-treatment completed sample is taken out of the glove box, it is immediately transferred to a 4°C refrigerated environment. The refrigerated environment reduces the enzyme activity and the chemical reaction rate in the pre-treatment completed sample and extends the effective detection period. The pre-treatment completed sample can be stored for 6 months under refrigerated conditions. During the storage period, the residual oxygen value and the component uniformity are sampled and inspected once a month to ensure the stability of the nutrient composition of the sample during the entire effective detection period and to guarantee the consistency and comparability of the test results when taking samples for testing at different times.

[0066] The purpose of the above embodiments is to reproduce and derive the technical solution of the present invention by way of example, and to fully describe the technical solution, purpose and effect of the present invention. The purpose is to enable the public to have a more thorough and comprehensive understanding of the disclosure of the present invention, and not to limit the scope of protection of the present invention.

[0067] The above embodiments are not an exhaustive list based on the present invention, and there may be many other embodiments not listed. Any substitutions and improvements made without departing from the concept of the present invention are within the protection scope of the present invention.

Claims

1. A method for pretreatment of samples for detecting the nutritional components of whole poultry, characterized in that, include: Frozen whole poultry is subjected to zoned temperature detection and freezing control to generate a zoned freezing degree distribution. Based on the zoned freezing degree distribution, the frozen whole poultry is cut into small pieces. Temperature-time monitoring is performed on the drying process of the poultry carcass pieces to generate a drying process curve. The surface and core moisture content of the poultry carcass pieces are detected to obtain moisture content data and moisture content gradient. Based on the moisture content data, the moisture content gradient and the drying process curve, the drying endpoint is determined to obtain a qualified dried sample. The dried qualified sample is coarsely pulverized to generate a coarse powder sample. The coarse powder sample is then subjected to stratified sampling and particle size distribution detection to obtain component distribution data and particle size distribution data. The ultrafine pulverization process parameters are determined based on the component distribution data and particle size data. The coarse powder sample is then subjected to liquid carbon dioxide-assisted low-temperature ultrafine pulverization treatment according to the ultrafine pulverization process parameters to obtain an ultrafine powder sample and the carbon dioxide is recovered. The ultrafine powder sample is homogenized to obtain a mixed sample. The oxygen content and uniformity of the mixed sample are detected to obtain residual oxygen concentration data and uniformity data. The carbon dioxide recovery filling parameters are determined based on the residual oxygen concentration data and the uniformity data. The mixed sample is then sealed according to the carbon dioxide recovery filling parameters to obtain a pre-treated sample.

2. The sample pretreatment method for detecting the nutritional components of whole poultry according to claim 1, characterized in that, The process of generating a zoned freezing degree distribution for frozen whole poultry through zoned temperature detection and freezing control includes: Temperature measurements were performed on the breast, leg, and wing parts of the frozen whole poultry to obtain the body part temperature. The temperature of the described area is compared with the freezing temperature threshold to identify areas that do not meet the standard and generate a mark for the areas that do not meet the standard. The areas marked as substandard are subjected to additional freezing treatment to determine the degree of freezing of the areas; A zoned freezing degree distribution is constructed based on the freezing degree of the described location.

3. The sample pretreatment method for detecting the nutritional components of whole poultry according to claim 1, characterized in that, The process of generating a drying process curve by monitoring temperature and time during the drying of the poultry carcass pieces includes: The drying process is monitored in real time to obtain the temperature sequence and drying time. Anomaly detection is performed on the temperature sequence to identify abnormal temperature points and generate temperature anomaly markers; The compensation drying time is determined based on the temperature anomaly marker. The temperature sequence is correlated with the compensated drying time to generate a drying process curve.

4. The pretreatment method for samples used in the detection of nutritional components in whole poultry according to claim 1, characterized in that, The step of performing stratified sampling and particle size distribution detection on the coarse powder sample to obtain component distribution data and particle size distribution data includes: The coarse powder sample was sampled in three layers (top, middle, and bottom) to obtain samples from each layer. The component ratios of each layer of the sample were obtained by performing component analysis on each layer. The distribution differences of feathers, bones, and muscles in each layer are extracted to generate component distribution data; The particle size distribution data were obtained by sieving the coarse powder sample.

5. The pretreatment method for samples used in the detection of nutritional components in whole poultry according to claim 1, characterized in that, The step of determining the ultrafine grinding process parameters based on the component distribution data and the particle size distribution data includes: Extract the feather ratio, bone ratio, and muscle ratio from the component distribution data; The particle size differences of each component are extracted from the particle size distribution data to generate component particle size characteristics; The particle size distribution data and the particle size characteristics of the components are used to identify difficult-to-crush components and generate difficult-to-crush labels. The ultrafine grinding process parameters are determined based on the feather ratio, the bone ratio, and the muscle ratio, combined with the difficult-to-grind markers.

6. The pretreatment method for samples used in the detection of nutritional components in whole poultry according to claim 1, characterized in that, The step of detecting the oxygen content and uniformity of the mixed sample to obtain residual oxygen concentration data and uniformity data includes: The oxygen content of the upper, middle and lower layers of the mixed sample was measured to obtain the oxygen concentration value of each layer; Component difference values ​​are generated by sampling from multiple points in the mixed sample to detect component differences. The maximum value of the oxygen concentration in each layer is determined as the residual oxygen concentration data; Uniformity data is generated based on the component difference values.

7. The pretreatment method for samples used in the detection of nutritional components in whole poultry according to claim 1, characterized in that, The step of determining the carbon dioxide recovery filling parameters based on the residual oxygen concentration data and the uniformity data includes: The initial filling volume is determined by comparing the residual oxygen concentration data with the target residual oxygen threshold. Threshold comparison is performed on the uniformity data to identify low-uniformity regions and generate supplementary filling marks; The residual oxygen value after filling is determined by testing the filling effect based on the initial filling volume. The secondary filling amount is determined based on the residual oxygen value after filling and the supplementary filling mark to generate the carbon dioxide recovery filling parameters.

8. The sample pretreatment method for detecting the nutritional components of whole poultry according to claim 4, characterized in that, The step of performing component analysis on each layer of the sample to obtain the component ratio of each layer includes: Density data for each layer of the sample is generated by density detection. Based on the density data of each layer, the boundary positions of the feather layer, bone layer and muscle layer are identified; The thickness of the boundary mixing layer is obtained by measuring the thickness at the boundary location; The mass of each layer is corrected based on the thickness of the boundary mixing layer, and the component ratio of each layer is calculated.

9. The sample pretreatment method for detecting the nutritional components of whole poultry according to claim 5, characterized in that, The determination of ultrafine grinding process parameters based on the feather ratio, bone ratio, and muscle ratio in conjunction with the difficult-to-grind markers includes: For the feather ratio, the bone ratio, and the muscle ratio, respectively, query the recommended crushing time for each component. The temperature monitoring threshold is determined based on the described difficult-to-crush markings; Based on the temperature monitoring threshold, the crushing process is monitored in real time to identify over-temperature points and generate intermittent cooling markers. The ultrafine grinding process parameters are determined based on the recommended grinding time for each component, the intermittent cooling mark, and the difficult-to-grind mark.

Images