A Multi-Parameter Adjustment-Based Energy Consumption Optimization Control System and Method for Precision Acid Removal of Meat

Abstract

This invention discloses a precision aging control system and method for meat based on multi-parameter adjustment, relating to the field of aging energy consumption optimization control. This invention employs a two-stage dynamic control approach, using independent model sets for the preparation and core stages to achieve hierarchical parameter optimization, ensuring that indicators at each stage meet standard requirements. Through multiple rounds of iterative screening to select the lowest energy consumption parameter combination, the cooling energy consumption per ton of product is significantly reduced. A "training-testing-adjustment" mechanism is used to continuously optimize the mapping model, greatly reducing the fluctuation range of mutton tenderness (shear force value). The final generated variable dataset and spacing parameters can be directly connected to aging specifications, achieving seamless integration of process parameters and aging procedures. A phased mapping model is constructed for the preparation and core aging stages, realizing a complete causal chain analysis from the initial carcass state to process intervention variables and final quality indicators, breaking through the limitations of traditional single-stage quality assessment.

Description

Technical Field

[0001] This invention belongs to the field of energy consumption optimization control for deacidification, and more specifically, it relates to a precise deacidification energy consumption optimization control system and method for meat based on multi-parameter adjustment. Background Technology

[0002] Traditional precise aging process parameters for mutton are usually assessed using a single-stage quality evaluation, resulting in poor adaptability of the obtained aging process parameters. This makes it difficult to adapt to the differences between different batches of carcasses. Furthermore, the manual experience-based parameter adjustment method used in the traditional precise aging process for mutton is random and limited, and is greatly influenced by subjective factors. In addition, the traditional precise aging process for mutton suffers from insufficient aging due to blindly reducing consumption. Summary of the Invention

[0003] To address the problems in related technologies, this invention proposes a meat precision acid removal energy consumption optimization control system and method based on multi-parameter adjustment, in order to overcome the aforementioned technical problems existing in the existing related technologies.

[0004] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution: This invention relates to a method for optimizing and controlling the energy consumption of meat during precise acid removal based on multi-parameter adjustment, comprising the following steps: S1. Collect the weight and initial spacing data of the meat to be de-acidified and the corresponding carcasses in the experimental group; S2. Set the variables to be adjusted for the preparation stage and the core aging stage of meat aging, and train the weight and initial arrangement spacing data in S1, and divide the test data. S3. Based on the training data divided in S2, construct and train a well-trained mapping model for the end indicators of the preparation stage and the core acid removal stage. S4. Input the test data divided in S2 into the corresponding end index mapping model in S3 for mapping and adjust the variables to be adjusted in the preparation stage and the core acid removal stage during the test. S5. Apply the adjusted variables from S4 to the actual acid removal operation during the test process; adjust the end index mapping model in S3 based on the operation results to obtain the final end index mapping model. S6. Input the weight, initial spacing data and corresponding aging variable data of the meat to be aged in S1 into the final end index mapping model in S5 for mapping and adjust the variables and spacing of the meat to be aged during the aging process. S7. Repeat S6 multiple times, and select the variable data of the acid removal process with the lowest energy consumption as the current final variable dataset.

[0005] Preferably, step S1 includes the following steps: S11. Obtain multiple pieces of meat to be pickled and measure the initial spacing between the carcasses corresponding to the multiple pieces of meat to be pickled and the average weight data of the carcasses, to obtain the initial carcass spacing set and the carcass weight data set; then set up several carcass experimental groups, each carcass experimental group including several carcasses, to obtain the carcass experimental group set; Several types of quality indicators for meat after aging are defined to obtain a set of quality indicator types after aging; the set of quality indicator types after aging includes carcass pH value, carcass temperature drop rate, carcass meat color, carcass juice loss rate, and carcass odor substance ratio; S12. Collect the average weight data of multiple carcasses in each carcass experimental group in the carcass experimental group set and the initial arrangement spacing data between them to obtain the carcass experimental weight dataset and the carcass experimental initial arrangement spacing dataset; then, according to the existing meat aging standard process, divide the meat aging process into stages to obtain the meat aging stage set; the meat aging stage set includes the preparation stage and the core aging stage. The experimental group's categorized design allows for targeted verification of optimal aging parameters for carcasses of different weights, avoiding meat quality differences caused by a "one-size-fits-all" standard. The staged aging process dynamically adjusts the cooling intensity, employing a rapid cooling mode within the critical two hours after slaughter, followed by a maintenance mode to save energy. Spacing data is linked with the aging stage to control the cold storage fan speed, avoiding energy waste caused by excessive airflow to areas with closely packed carcasses.

[0006] Preferably, step S2 includes the following steps: S21. In conjunction with the meat aging stage set, set the equipment types used in the preparation stage and the core aging stage to obtain a first equipment type set and a second equipment type set; S22. Based on the first equipment type set and the second equipment type set, set the types of variables to be adjusted for the preparation stage and the core acid removal stage, and obtain the first variable type set and the second variable type set; S23. Divide the carcass experimental group set into training and testing groups to obtain carcass experimental training groups and carcass experimental testing groups; in conjunction with the carcass experimental weight dataset and the carcass experimental initial arrangement spacing dataset, obtain the carcass weight and initial arrangement spacing data corresponding to the carcass experimental training group and the carcass experimental testing group to obtain the carcass experimental training weight dataset, the carcass experimental testing weight dataset, the carcass experimental training initial arrangement spacing dataset, and the carcass experimental testing initial arrangement spacing dataset; By setting the first set of variables, subsequent parameter linkage adjustments are made, thereby shortening the precooling time and reducing energy consumption; the second set of variables adopts a gradient airflow organization design to reduce the difference in wind speed between the carcasses and improve the uniformity of acid discharge.

[0007] Preferably, step S3 includes the following steps: S31. Construct a mapping model between the carcass quality index data before the start of the preparation stage and the core acid removal stage, the applied variable data during the process, the carcass experimental training weight data, and the initial arrangement spacing data of the carcass experimental training, and the various index data after the corresponding stage. Train the mapping model based on the carcass experimental training group, the first variable type set, the quality index type set after acid removal, the carcass experimental training weight dataset, and the initial arrangement spacing dataset of the carcass experimental training to obtain the trained preparation end index mapping model set and the trained core acid removal end index mapping model set. By constructing a phased mapping model between the preparation stage and the core deacidification stage, a complete causal chain analysis from the initial carcass state → process intervention variables → final quality indicators is achieved, breaking through the limitations of traditional single-stage quality assessment. During model training, key variable combinations are automatically identified, and process optimization paths that are difficult to capture with human experience are discovered. Differentiated deacidification schemes are matched according to the initial weight dataset to solve the quality fluctuation problem caused by the "one-size-fits-all" approach in traditional production lines. Process fault tolerance is enhanced.

[0008] Preferably, step S4 includes the following steps: S41. Before the start of the preparation stage, collect the quality index data of multiple carcasses in each carcass experimental test group to obtain the initial quality index dataset of carcass test; obtain the acid removal application variable data corresponding to the preparation stage and the core acid removal stage of the carcass experimental test group to obtain the first test initial variable dataset and the second test initial variable dataset. S42. Input the initial index data of each initial index in the initial index dataset of carcass test quality, the weight dataset of carcass experiment test, the initial variable dataset of the first test, and the initial spacing dataset of carcass experiment test into the corresponding mapping model in the trained set of ready-to-end index mapping models for mapping; adjust the initial spacing dataset of carcass experiment test and the initial variable dataset of the first test according to the mapping results to obtain the final variable dataset of the first test. S43. Input the data of each indicator after the preparation stage in S42, the data of carcass arrangement spacing, the dataset of initial variables for the second test, the dataset of carcass experimental test preparation arrangement spacing, and the dataset of carcass experimental test weight into the corresponding mapping model in the set of trained core acid removal end indicator mapping models for mapping, and obtain the dataset of quality indicators for the second test. Based on the second test quality index dataset, the carcass experimental test preparation arrangement spacing dataset and the second test initial variable dataset are adjusted to obtain the second test final variable dataset; By setting two-tiered quality indicator standard ranges for the preparation stage and the core acid removal stage, a quantifiable process target system is formed. The quality data output by the model is compared with the standard range in real time, and the parameter adjustment mechanism is automatically triggered to ensure that the final parameters ensure that the carcass reaches the optimal acid removal state.

[0009] Preferably, step S5 includes the following steps: S51. Use the carcass spacing data and the first test final variable dataset after the preparation stage in S42 in the actual preparation stage of the carcass experimental test group; after the preparation stage is completed, the first actual quality index dataset is obtained. Based on the first actual quality index dataset, determine whether to return to S31 to continue training the trained ready-to-end index mapping model set, and repeat S41, S42, S43 and S51 to obtain the final ready-to-end index mapping model set. S52. Use the carcass spacing data and the second test final variable dataset after the core acid removal stage in S43 in the actual core acid removal stage of the carcass experimental test group; after the core acid removal stage, obtain the second actual quality index dataset. Based on the second actual quality index dataset, determine whether to return to S31 to continue training the trained core acid removal end index mapping model set, and repeat S41, S42, S43, S51 and S52 to obtain the final core acid removal end index mapping model set. By setting independent error thresholds for the preparation stage and the core acid removal stage, the model is automatically retrained when the error exceeds the threshold, ensuring that the model's prediction accuracy is greatly improved. Through the dual verification mechanism, the final product compliance rate is greatly improved compared with the traditional process, and the coefficient of variation of quality indicators between different batches is controlled.

[0010] Preferably, step S6 includes the following steps: S61. Based on the set of quality indicators after aging, collect the quality indicator data of the carcasses corresponding to the current multiple pieces of meat to be aged, and obtain the current initial quality indicator dataset. S62. Input the data of each indicator in the current initial quality indicator dataset, the data of acid removal applied during the preparation stage, the initial carcass spacing set to be acid removed, and the carcass weight dataset to be acid removed into the model corresponding to the final preparation and completion indicator mapping model set for mapping. Adjust the initial carcass spacing set to be acid removed and the data of acid removal applied during the preparation stage according to the mapping results. S63. Input the data of each indicator after the preparation stage in S62, the data of carcass spacing, the variables applied during the core acid removal stage, and the weight dataset of the carcass to be acid removed into the model corresponding to the final core acid removal end indicator mapping model set for mapping. Adjust the carcass spacing set to be acid removed and the data of variables applied during the core acid removal stage according to the mapping results. A progressive quality assurance mechanism is formed through graded control in the preparation and core acid removal stages. Each stage has an independent quality standard range and variable adjustment mechanism to ensure that the parameters of each link accurately match the process requirements. A closed-loop logic of "detection-mapping-adjustment" is adopted to correct parameter combinations that deviate from the standard in real time. The mapping model set automatically matches the optimal parameter combination to replace traditional experience judgment. The quality indicator standard range provides a quantitative evaluation benchmark to avoid subjective bias.

[0011] Preferably, step S7 includes the following steps: S71. Repeat S62 and S63 multiple times to obtain multiple sets of data on the acid removal applied variables and the corresponding carcass spacing data corresponding to the preparation stage and the core acid removal stage in the current acid removal process. These are denoted as the current candidate variable dataset and the current candidate carcass spacing dataset. S72. Evaluate the energy consumption of the acid removal application variable data corresponding to the preparation stage and the core acid removal stage in each group of the current acid removal process in the current candidate variable dataset; select the current candidate variable data corresponding to the minimum energy consumption data as the current final variable dataset. S73. Combine the current final variable dataset and the carcass spacing data in the corresponding current candidate carcass spacing dataset to perform acid removal processing on the current multiple pieces of meat to be acid removed; The system selects the most energy-efficient option from the combination of compliant parameters, significantly reducing the cooling energy consumption per ton of de-acidified meat; the minimum energy consumption selection algorithm breaks through the traditional single quality-oriented model, greatly reducing the overall cost of enterprises; and it fully records the variable data, spacing data, and energy consumption data of each iteration to form a digital twin of the process, providing data support for subsequent optimization.

[0012] The meat precision aging energy consumption optimization control system based on multi-parameter adjustment includes a carcass initial data acquisition module, a carcass experimental variable setting and data division module, an aging stage end index mapping model construction and training module, a carcass experimental test data mapping adjustment module, an aging stage end index mapping model testing module, a current aging variable spacing adjustment module, and a minimum energy consumption aging variable screening module.

[0013] The present invention has the following beneficial effects: 1. This invention employs a two-stage dynamic control approach, using independent model sets for the preparation and core stages to achieve hierarchical parameter optimization, ensuring that indicators at each stage meet standard requirements; through multiple rounds of iteration to screen the lowest energy consumption parameter combination, the cooling energy consumption per ton of product is significantly reduced; a "training-testing-adjustment" mechanism is used to continuously optimize the mapping model, greatly reducing the fluctuation range of mutton tenderness (shear force value); the final generated variable dataset and spacing parameters can be directly connected to the aging specifications, achieving seamless integration of process parameters and aging procedures.

[0014] 2. This invention constructs a phased mapping model between the preparation stage and the core aging stage, achieving a complete causal chain analysis from the initial carcass state to process intervention variables and final quality indicators, overcoming the limitations of traditional single-stage quality assessment. During model training, it automatically identifies key variable combinations and discovers process optimization paths that are difficult to capture through human experience. Based on the trained mapping model, the tenderness / water retention indicators at the aging endpoint can be directly predicted by inputting preparation stage data, allowing for advance adjustment of environmental parameters and significantly reducing the aging failure rate. By replacing some physical experiments with a virtual mapping model, the required sample size for a single batch of sheep carcasses is greatly reduced, while avoiding raw material waste caused by trial and error.

[0015] 3. In this invention, by selecting the most energy-efficient option from the combination of compliant parameters, the cooling energy consumption per ton of aged mutton is greatly reduced; the minimum energy consumption selection algorithm breaks through the traditional single quality-oriented mode, which greatly reduces the overall cost of enterprises.

[0016] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the process for optimizing energy consumption control of precise aging of meat based on multi-parameter adjustment, as described in this invention. Figure 2 This is a schematic diagram of the module of the meat precision acid removal energy consumption optimization control system based on multi-parameter adjustment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the invention, and not all embodiments. Based on the embodiments of the invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the invention.

[0020] Example 1:

[0021] Please see Figure 1 This embodiment takes mutton as an example to illustrate a method for optimizing energy consumption control in the precise aging process of mutton based on multi-parameter adjustment, including the following steps: S1. Collect the weight and initial spacing data of the mutton to be fermented and the corresponding mutton carcasses in the experimental group; S1 includes the following steps: S11. Obtain multiple pieces of mutton to be pickled and measure the initial spacing between the corresponding mutton carcasses and the average weight of the mutton carcasses to obtain the initial spacing set and weight data set of the mutton carcasses to be pickled. Specifically, a laser rangefinder or measuring tape can be used to vertically measure the distance between the center points of adjacent carcasses to obtain the spacing data. The weight data includes volume, etc. Then, set up several mutton carcass experimental groups, each mutton carcass experimental group including several mutton carcasses, to obtain the mutton carcass experimental group set. Several quality index types for mutton after aging were defined, resulting in a set of quality index types. These included mutton carcass pH value, mutton carcass temperature drop rate, mutton carcass meat color, mutton carcass juice loss rate, and the proportion of muttony odor substances (4-methyloctanoic acid). Specifically, the mutton carcass pH value, mutton carcass temperature drop rate, mutton carcass meat color, tenderness, and mutton carcass juice loss rate were measured using pH test paper, a thermometer, a spectrophotometer, and an electronic scale, respectively. The proportion of muttony odor substances was obtained by scanning the carcass surface using neospectral imaging (NIR) in the 400-1700 nm band, combined with a chemometric model to invert the spatial distribution ratio of the odor substances. S12. Collect the average weight data of multiple sheep carcasses in each sheep carcass experimental group and the initial arrangement spacing data between them to obtain the sheep carcass experimental weight dataset and the sheep carcass experimental initial arrangement spacing dataset; then, according to the existing mutton aging standard process, divide the mutton aging process into stages to obtain the mutton aging stage set; the mutton aging stage set includes the preparation stage (0-2 hours after slaughter) and the core aging stage (2-48 hours after slaughter); specifically, the preparation stage (0-2 hours) adopts a rapid cooling mode to immediately inhibit microbial activity and shorten the reproduction window of pathogens such as Salmonella by more than 70%; the standardized humidity control (90%RH) in the core stage (2-48 hours) prevents the surface dry film from breaking and blocks the external contamination penetration path; Precise measurement of initial carcass spacing (laser ranging / vertical calibration with measuring tape) eliminates the uneven cold air distribution caused by traditional manual estimation, ensuring a uniform temperature and humidity environment for each piece of mutton during the aging process. The experimental group's categorized design allows for targeted verification of optimal aging parameters for different weights of sheep carcasses (e.g., lambs and adult sheep), avoiding meat quality differences caused by a "one-size-fits-all" standard. The phased aging process (preparation stage + core stage) dynamically adjusts the cooling intensity, employing a rapid cooling mode within the critical two hours after slaughter, followed by a maintenance mode, saving approximately 25% of energy compared to traditional continuous strong cooling. Spacing data is linked to the aging stage to control the cold storage fan speed, avoiding energy waste caused by excessive airflow to tightly packed carcass areas. The experimental group's dataset provides training samples for subsequent intelligent aging processes. S2. Set the variables to be adjusted for the preparation stage and the core aging stage of mutton aging, and train the weight and initial arrangement spacing data in S1, and divide the test data. S2 includes the following steps: S21. In conjunction with the aforementioned lamb aging stage set, define the equipment types used in the preparation stage and the core aging stage to obtain a first equipment type set and a second equipment type set; the first equipment type set includes a high-pressure blower (wind speed ≥ 8m / s), an evaporator, and a blood washing sprayer (used to rinse away residual blood stains and reduce subsequent aging odors); the second equipment type set includes a titanium-plated double-wall temperature-controlled chamber and an ultrasonic humidifier (used to maintain 85%-90% humidity and prevent the carcass surface from dehydrating and hardening). S22. Based on the first equipment type set and the second equipment type set, set the types of variables to be adjusted for the preparation stage and the core acid removal stage, and obtain the first variable type set and the second variable type set; the first variable type set includes fan speed (high pressure fan), evaporator power and spray water pressure (blood cleaning spray machine); the second variable type set includes main circulation velocity, auxiliary airflow velocity (titanium-plated double-wall temperature control chamber) and humidifier power (ultrasonic humidifier). S23. Divide the sheep carcass experimental group set into training and testing groups to obtain sheep carcass experimental training groups and sheep carcass experimental testing groups; in conjunction with sheep carcass experimental weight dataset and sheep carcass experimental initial arrangement spacing dataset, obtain sheep carcass weight and initial arrangement spacing data corresponding to sheep carcass experimental training groups and sheep carcass experimental testing groups to obtain sheep carcass experimental training weight dataset, sheep carcass experimental testing weight dataset, sheep carcass experimental training initial arrangement spacing dataset, and sheep carcass experimental testing initial arrangement spacing dataset; wherein, the division ratio in the process of dividing the training group and testing group can be adaptively set according to the actual situation, such as 7:3 or 8:2. The preparation stage equipment set (high-pressure fan + evaporator + blood washing sprayer) achieves a rapid 4°C drop in the carcass's core temperature and removal of surface blood stains through the triple action of rapid cooling by high-speed airflow, precise heat absorption by the evaporator, and high-pressure rinsing, reducing the risk of microbial growth and odor residue from the source. The core deacidification stage equipment set (titanium-plated double-walled temperature-controlled chamber + ultrasonic humidifier) ​​provides dynamic temperature and humidity balance control to maximize lactate-degrading enzyme activity while preventing muscle fiber dehydration and contraction leading to hardening. By setting the first set of variables (fan speed / evaporator power / ... The spray pressure allows for subsequent parameter linkage adjustment, thereby shortening the pre-cooling time and reducing energy consumption; the second variable set (main and auxiliary airflow velocity / humidification power) adopts a gradient airflow organization design to reduce the wind speed difference between the carcass gaps and improve the uniformity of acid removal; in addition, the main circulation of the titanium-plated chamber promotes the fragmentation of myofibrils, which increases the amount of IMP umami substances generated and reduces the amount of muttony substances (4-methyloctanoic acid) residue; ultrasonic humidification forms a nanoscale water film, reducing the weight loss due to dryness during acid removal and delaying myoglobin oxidation; the evaporator frequency conversion technology can reduce the power consumption in the pre-cooling stage; S3. Based on the training data divided in S2, construct and train a well-trained mapping model for the end indicators of the preparation stage and the core acid removal stage. S3 includes the following steps: S31. Construct a mapping model between the quality index data of sheep carcasses before the start of the preparation stage and the core acid removal stage, the applied variable data during the process, the sheep carcass experimental training weight data, and the initial arrangement spacing data of sheep carcasses experimental training, and the various index data after the corresponding stage. Train the mapping model based on the sheep carcass experimental training group, the first variable type set, the quality index type set after acid removal, the sheep carcass experimental training weight dataset, and the initial arrangement spacing dataset of sheep carcasses experimental training to obtain the trained preparation end index mapping model set and the trained core acid removal end index mapping model set. S31 includes the following steps: S311. Before the preparation phase begins, quality index data of multiple sheep carcasses in each group of the sheep carcass experimental training group are collected to obtain the initial quality index dataset of the sheep carcass experiment. According to the first variable type set and the quality index type set after acid removal, variables for the preparation phase are applied to multiple sheep carcasses in each group of the sheep carcass experimental training group, and the corresponding applied variable data are collected to obtain the first sheep carcass experimental training variable dataset. After the preparation phase ends, the first sheep carcass experimental training group after the end of the preparation phase is obtained, and the corresponding quality index data after acid removal is collected to obtain the first sheep carcass experimental quality index dataset. A mapping model is constructed between the initial quality index data of sheep carcasses, the experimental training variable data during the preparation phase, the sheep carcass experimental training weight data, and the initial arrangement spacing data of sheep carcasses experimental training, and the various index data after the preparation phase, respectively, to obtain the preparation-end index mapping model set. The initial quality index dataset, sheep carcass experimental training weight dataset, sheep carcass experimental training initial arrangement spacing dataset, first sheep carcass experimental training variable dataset, and first sheep carcass experimental quality index dataset are used to train each mapping model in the preparation-end index mapping model set. After the training operation is completed, the trained preparation-end index mapping model set is obtained. The structure of the prepared end indicator mapping model is shown in Table 1 below: Table 1

[0022] Where: Input: Input; Conv1D: Convolutional layer; 64: Number of convolutional kernels (number of output channels), determining the number of feature maps extracted; kernel_size=3: Kernel width is 3 time steps, controlling the size of the local receptive field; strides=1: Sliding stride is 1, maintaining temporal resolution; BatchNorm: Batch normalization, standardizing each batch of data to accelerate training convergence and stabilize gradients; ReLU: Activation function, correcting linear units by setting negative values ​​to zero and retaining positive values, introducing nonlinearity. Features; MaxPooling1D: Pooling layer; pool_size=2: Pooling window width is 2, compressing feature dimension by 50%; LSTM: Long Short-Term Memory network; 128: Number of hidden units, controlling memory capacity; return_sequences=False: Only output the result of the last time step; Dense: Fully connected layer; 64: Number of neurons, realizing feature space transformation; activation='relu': Activation function is the same as above; Output layer; S312. Based on the second variable type set and the quality index type set after acid removal, apply the core acid removal stage variables to multiple sheep carcasses in each group of the sheep carcass experiment training group after the first stage, collect the corresponding applied variable data, and obtain the second sheep carcass experiment training variable dataset; after the core acid removal stage, collect the corresponding quality index data after acid removal, and obtain the second sheep carcass experiment quality index dataset. A mapping model was constructed between the first sheep carcass experimental quality index data, the experimental training variable data during the core aging stage, the sheep carcass experimental training weight data, and the initial arrangement spacing data of the sheep carcass experimental training, and the various index data after the core aging stage, respectively, to obtain the core aging end index mapping model set. The first sheep carcass experimental quality index dataset, the sheep carcass experimental training weight dataset, the sheep carcass experimental training initial arrangement spacing dataset, the second sheep carcass experimental training variable dataset, and the second sheep carcass experimental quality index dataset were used to train each mapping model in the core aging end index mapping model set. After the training operation was completed, the trained core aging end index mapping model set was obtained. The structure of the core acid removal completion index mapping model is shown in Table 2 below: Table 2 ; Wherein, GRU: Gated Recurrent Unit; 256: Number of hidden units, controlling the network's memory capacity; return_sequences=True: Outputs the results of all time steps (preserving the complete temporal sequence); MultiHeadAttention: Multi-Head Attention Mechanism; num_heads=4: Number of parallel attention heads, enhancing feature extraction capabilities; key_dim=64: Key vector dimension of each attention head; LayerNormalization: Layer normalization, standardizing all features of a single sample to stabilize the training process; 128: Number of neurons; activation='swish': Activation function (Sigmoid weighted linear unit), smooth and non-monotonic; Dropout: Random deactivation; 0.3: 30% of neurons randomly deactivate to prevent overfitting; Sigmoid: Activation function; When processing structured data (quality / weight indicators) using XGBoost, its built-in feature importance assessment can automatically identify key influencing factors, such as the correlation between fat thickness and muscle pH. The 1D-CNN+LSTM hybrid network can jointly model spatiotemporal features (interval / process parameter time series) to capture local fluctuations and long-term dependencies, such as the lag effect of temperature fluctuations on acid removal rate. The GRU+multi-head attention mechanism can dynamically adjust the parameter sensitivity of different process stages (such as pre-cooling and isothermal) by memorizing historical states through 256 hidden units and combining them with the 4-head attention weight allocation. The smooth gradient properties of the Swish activation function effectively solve the neuron death problem of traditional ReLU when process parameters change abruptly; the GAT model quantifies the correlation strength (2-hop neighborhood) of the acid removal distance between sheep carcasses through an 8-head attention mechanism, simulating the interaction effect between airflow organization and carcasses in actual cold storage; node features fuse weight data and process parameters, and edge weights reflect the influence of spatial position, realizing a digital twin of the three-dimensional acid removal environment; the batch normalization layer ensures that data of different dimensions (such as temperature 0-4℃ and humidity 70%-90%) maintains consistent distribution at input, accelerating model convergence; the 30% random inactivation rate of the Dropout layer enhances generalization ability and prevents overfitting to small sample experimental data; compared with traditional statistical models, spatiotemporal feature joint modeling reduces prediction error by 15%-24%; key process parameters are identified through attention weight visualization (such as num_heads=4), reducing the number of trial and error experiments by up to 40%; the topology analysis function of graph neural networks can optimize the cold storage carcass arrangement scheme and improve space utilization by more than 20%; By constructing a phased mapping model between the preparation stage and the core aging stage, a complete causal chain analysis is achieved from the initial carcass state (weight / spacing) to process intervention variables and then to the final quality indicators, overcoming the limitations of traditional single-stage quality assessment. During model training, key variable combinations (such as the synergistic effect of specific initial spacing and wind speed variables) are automatically identified, revealing process optimization paths that are difficult to capture through human experience. Based on the trained mapping model, the tenderness / water retention indicators at the aging endpoint can be directly predicted by inputting preparation stage data (such as the rate of pH decrease), allowing for adjustments to environmental parameters (temperature / humidity) 24-48 hours in advance, reducing the aging failure rate by more than 60%. Through virtual mapping models… The model replaces some physical experiments, reducing the required sample size of a single batch of sheep carcasses by 40%, while avoiding raw material waste caused by trial and error; the predicted value of the core acid removal end index output by the model can be linked to the refrigeration system to achieve on-demand cooling (such as automatically raising the storage temperature by 2°C after predicting that the water retention meets the standard), saving energy by 15%-20%; it matches differentiated acid removal schemes according to the initial weight dataset (such as automatically extending the acid removal time by 1.5 hours for heavy carcasses), solving the quality fluctuation problem caused by the "one-size-fits-all" approach of traditional production lines; the process fault tolerance is enhanced: when an abnormal initial pH is detected in a batch, the model automatically generates a combination of compensatory variables (such as increasing humidity and extending the settling time) to control the defect rate to within 5%; S4. Input the test data divided in S2 into the corresponding end index mapping model in S3 for mapping and adjust the variables to be adjusted in the preparation stage and the core acid removal stage during the test. S4 includes the following steps: S41. In conjunction with the sheep carcass experimental test group and the set of quality index types after acid removal, set the standard intervals for various index data of the sheep carcass experimental test group during the preparation stage and after the core acid removal stage, and obtain the first test quality index standard interval set and the second test quality index standard interval set; before the start of the preparation stage, collect the quality index data of multiple sheep carcasses in each group of the sheep carcass experimental test group to obtain the initial dataset of sheep carcass test quality. Then, based on the first variable type set and the second variable type set, obtain the acid removal application variable data corresponding to the preparation stage and the core acid removal stage of the sheep carcass experimental test group, and obtain the first test initial variable dataset and the second test initial variable dataset. S42. Input the initial index data, the sheep carcass experimental test weight dataset, the first test initial variable dataset, and the sheep carcass experimental test initial arrangement spacing dataset from the initial index dataset of the sheep carcass test into the corresponding mapping model in the set of the prepared end index mapping model for mapping, and obtain the first test quality index dataset. If there are quality index data in the first test quality index dataset that are not located within the corresponding quality index standard interval of the first test quality index standard interval set, the initial arrangement spacing dataset and the initial variable dataset of the sheep carcass experiment test are adjusted until there are no quality index data in the first test quality index dataset that are not located within the corresponding quality index standard interval of the first test quality index standard interval set, thus obtaining the sheep carcass experiment test preparation arrangement spacing dataset, the first test final variable dataset, and the first final test quality index dataset; otherwise, no adjustment is required, and the first test quality index dataset is used as the first final test quality index dataset. S43. Input the data of each indicator in the first final test quality indicator dataset, the second test initial variable dataset, the sheep carcass test preparation arrangement spacing dataset, and the sheep carcass test weight dataset into the corresponding mapping model in the trained core acid removal end indicator mapping model set for mapping, and obtain the second test quality indicator dataset. If any quality indicator data in the second test quality indicator dataset is not located within the corresponding quality indicator standard interval of the second test quality indicator standard interval set, the sheep carcass experimental test preparation arrangement spacing dataset and the second test initial variable dataset are adjusted until no quality indicator data in the second test quality indicator dataset is not located within the corresponding quality indicator standard interval of the second test quality indicator standard interval set, thus obtaining the sheep carcass experimental test core arrangement spacing dataset, the second test final variable dataset, and the second final test quality indicator dataset; otherwise, no adjustment is required, and the second test quality indicator dataset is used as the second final test quality indicator dataset. By setting two-tiered quality indicator standard ranges for the preparation and core acid removal stages, a quantifiable process target system is formed. The model's output quality data is compared with the standard ranges in real time, automatically triggering a parameter adjustment mechanism to ensure that the final parameters guarantee the sheep carcass reaches the optimal acid removal state. Static features such as initial quality indicators, weight data, and spacing are integrated with dynamic time-series data of process variables (temperature, humidity, wind speed). The XGBoost regression model's efficient processing capability for structured data (max_depth=8 to prevent overfitting) and the spatiotemporal feature extraction advantages of 1D-CNN+LSTM (64 convolutional kernels to capture local fluctuations) are utilized to achieve cross-dimensional parameter collaborative optimization. When quality indicators deviate from the standard... During interval processing; the batch normalization layer (BatchNorm) eliminates the difference in data dimensions between different sensors, and the dropout layer (0.3 inactivation rate) improves the model's generalization ability, enabling the system to maintain stable output under raw material differences (such as different breeds of sheep carcasses) and environmental fluctuations (changes in cold storage load); through dynamic optimization of arrangement spacing (three-level adjustment of initial spacing → preparation spacing → core spacing), the cold storage space utilization rate is improved by 15%-20%, while reducing the problem of uneven local acid discharge caused by excessive stacking; attention weight visualization (such as num_heads=4) reveals the nonlinear relationship between variables (such as the synergistic effect of humidity and wind speed), forming a reusable process knowledge base and shortening the process commissioning cycle of new plants by 40%; S5. Apply the adjusted variables from S4 to the actual acid removal operation during the test process; adjust the end index mapping model in S3 based on the operation results to obtain the final end index mapping model. S5 includes the following steps: S51. The sheep carcass experimental test preparation spacing dataset and the first test final variable dataset are used in the actual preparation stage of the sheep carcass experimental test group; after the preparation stage is completed, the corresponding data of various types of quality indicators after acid removal are collected according to the quality indicator type set after acid removal to obtain the first actual quality indicator dataset. Set a first error threshold; calculate the Euclidean distance between the first actual quality index dataset and the first final test quality index dataset to obtain the first experimental error data; if the first experimental error data is greater than or equal to the first error threshold, return to S31 to continue training the trained indicator mapping model set to be terminated, and repeat S41, S42, S43 and S51 until the first experimental error data is less than the first error threshold, to obtain the final indicator mapping model set to be terminated; otherwise, use the trained indicator mapping model set to be terminated as the final indicator mapping model set to be terminated. S52. The core arrangement spacing dataset and the second final variable dataset of the sheep carcass experiment test are used in the actual core acid removal stage of the sheep carcass experiment test group; after the core acid removal stage is completed, the corresponding quality index data of various types after acid removal are collected to obtain the second actual quality index dataset. Set a second error threshold; calculate the Euclidean distance between the second actual quality index dataset and the second final test quality index dataset to obtain the second experimental error data; if the second experimental error data is greater than or equal to the second error threshold, return to S31 to continue training the trained core acid removal end index mapping model set, and repeat S41, S42, S43, S51 and S52 until the second experimental error data is less than the second error threshold, thus obtaining the final core acid removal end index mapping model set; otherwise, use the trained core acid removal end index mapping model set as the final core acid removal end index mapping model set. By setting independent error thresholds for the preparation and core acid removal stages, and using Euclidean distance to quantify the deviation between predicted and measured values, the model is automatically retrained when the error exceeds the threshold, ensuring that the model's prediction accuracy is improved by more than 23%. By using the spacing data (spatial dimension) and process variables (time dimension) as joint inputs, and modeling the spatiotemporal correlation through a GRU neural network, the parameter adjustments in the core acid removal stage can automatically compensate for the accumulated errors in the preparation stage. Through a dual verification mechanism (preparation stage + core stage), the final product compliance rate is increased from 82% in the traditional process to 98%, and the coefficient of variation of quality indicators between different batches is controlled within 5%. S6. Input the weight, initial spacing data and corresponding aging variable data of the mutton to be aged in S1 into the final end index mapping model in S5 for mapping and adjust the variables and spacing in the aging process of the mutton to be aged. S6 includes the following steps: S61. Based on the first variable type set and the second variable type set, obtain the acid removal applied variable data corresponding to the preparation stage and the core acid removal stage in the current acid removal process, and obtain the first current initial variable dataset and the second current initial variable dataset; then, based on the acid removal quality index type set, collect the quality index data of the sheep carcass corresponding to the current multiple pieces of mutton to be acid removed, and obtain the current initial quality index dataset. S62. Input the data of each indicator in the current initial quality indicator dataset, the first current initial variable dataset, the initial carcass spacing set to be acidified, and the carcass weight dataset to be acidified into the model corresponding to the final preparation end indicator mapping model set for mapping, and obtain the current preparation quality indicator dataset. If any quality indicator data in the current preparation quality indicator dataset is not located within the quality indicator standard interval corresponding to the first test quality indicator standard interval set, the initial sheep carcass spacing set to be leached and the first current initial variable dataset are adjusted until no quality indicator data in the current preparation quality indicator dataset is not located within the quality indicator standard interval corresponding to the first test quality indicator standard interval set, thus obtaining the sheep carcass spacing set to be leached, the first current final variable dataset, and the current final preparation quality indicator dataset; otherwise, no adjustment is needed, and the current preparation quality indicator dataset is used as the current final preparation quality indicator dataset. S63. Input the data of each indicator in the current final preparation quality indicator dataset, the second current initial variable dataset, the carcass spacing set of sheep to be quenched and the carcass weight dataset of sheep to be quenched into the model corresponding to the final core quenching end indicator mapping model set for mapping, and obtain the current core quality indicator dataset. If any quality indicator data in the current core quality indicator dataset is not located within the corresponding quality indicator standard interval of the second test quality indicator standard interval set, the carcass spacing set of sheep to be quenched and the second current initial variable dataset are adjusted until no quality indicator data in the current core quality indicator dataset is not located within the corresponding quality indicator standard interval of the second test quality indicator standard interval set, thus obtaining the carcass spacing set of sheep to be quenched, the second current final variable dataset, and the current final core quality indicator dataset; otherwise, no adjustment is needed, and the current core quality indicator dataset is used as the current final core quality indicator dataset. For example, taking a certain sheep carcass aging process as an example, as follows: Preparation phase standards: pH dynamic curve: initial 6.3±0.2 → 24-hour target 5.9±0.1 (monitored online using a Mettler Toledo pH electrode); temperature gradient: 38℃→4℃ (pre-cooling rate 2℃ / h, surface temperature difference monitored by infrared thermal imager ≤0.8℃); Core standards for the acid removal stage: Tenderness index: Shear force 3.2±0.3kg (measured by TA.XT Plus texture analyzer); Water retention (sheep carcass juice loss rate): Centrifugation loss rate ≤28% (centrifugation at 3000 rpm for 10 minutes); Color stability (meat color of lamb carcass): L value ≥ 50, a value ≤ 14 (detected by HunterLab colorimeter); Data collection: Initial data of sheep carcasses, Group A (5 groups) { List of sheep carcass numbers "carcass_id": ["A01","A02","A03","A04","A05"], Carcass weight (unit: kilograms) "weight_kg": [18.6,17.9,19.2,18.1,17.5], Carcass spacing data (unit: cm): For example, 5.2 shows the spacing data between carcasses A01 and A02; "carcass_spacing"": [5.2,4.8,6.1,5.0,4.5], initial pH value "initial_pH": [6.31,6.28,6.35,6.29,6.33] } Process parameters are shown in Table 3 below: Table 3 ; Preparation phase exception handling: Scenario: A03 carcass predicted pH=6.15 (exceeding the upper limit of 6.0) Adjustments: Spacing increased from 60cm to 75cm (based on GAT model spatial optimization); wind speed increased to 0.7m / s for 2 hours; Six hours later, the measured pH was 5.92, indicating that the core acid removal phase had begun. Core acid removal stage optimization: Input feature dimensions: weight data, environmental time series: 5-minute temperature and humidity sequence; Control effect: When the predicted tenderness is 3.6kg, the system automatically extends the aging time by 8 hours and increases the humidity to 73%. The final measured shear force was 3.1 kg; The final verification results are shown in Table 4 below: Table 4 ; A progressive quality assurance mechanism is formed through graded control in the preparation and core acid removal stages. Each stage has an independent quality standard range and variable adjustment mechanism to ensure that parameters at each stage accurately match process requirements. A closed-loop logic of "detection-mapping-adjustment" is adopted to correct parameter combinations that deviate from the standard in real time. A dynamic correlation model is integrated between key variables such as temperature, spacing, and weight and quality indicators. Through the classification management of variable type sets (first / second variable sets), parameter decoupling control of different process stages is achieved. The coupling adjustment of spacing parameters (initial / preparation / core three-state evolution) with variable data optimizes cold air circulation efficiency. The model set automatically matches the optimal parameter combination, replacing traditional experience-based judgment. The quality indicator standard range provides a quantitative evaluation benchmark, avoiding subjective bias. The iterative adjustment algorithm ensures that the final output parameters necessarily meet the preset quality standards. It fully records the variable dataset and quality indicator dataset at each stage, forming a digital process archive. It provides data support for establishing differentiated aging programs for different varieties / cuts of mutton. It avoids quality loss caused by over-aging or under-aging, increasing the rate of high-quality products by more than 30%. The dynamic adjustment mechanism reduces energy waste, lowering cooling energy consumption by 15%-20%. The standardized process shortens the operator training cycle and reduces the rate of human error. S7. Repeat S6 multiple times and select the variable data of the acid removal process with the lowest energy consumption as the current final variable dataset. S7 includes the following steps: S71. Based on the fact that there are no quality indicator data in the current core quality indicator dataset that are not located within the quality indicator standard interval corresponding to the second test quality indicator standard interval set, and that there are no quality indicator data in the current preparation quality indicator dataset that are not located within the quality indicator standard interval corresponding to the first test quality indicator standard interval set, repeat S62 and S63 to obtain multiple sets of adjusted data on the acid removal application variables and the corresponding sheep carcass spacing data corresponding to the preparation stage and core acid removal stage in the current acid removal process. These are denoted as the current candidate variable dataset and the current candidate sheep carcass spacing dataset. S72. Evaluate the energy consumption corresponding to the acid removal application variable data of each group in the current acid removal process in the current candidate variable dataset for the preparation stage and the core acid removal stage, and obtain the current candidate energy consumption dataset; select the current candidate variable data corresponding to the minimum energy consumption data in the current candidate energy consumption dataset as the current final variable dataset; S73. Combine the current final variable dataset and the corresponding current candidate carcass spacing dataset with the carcass spacing data to perform deacidification treatment on the current multiple pieces of mutton to be deacidified; Based on dual verification of quality indicators, a closed-loop system of "generation-screening-verification" for parameter combinations is established to ensure that each set of variable data passes the rigorous standard test in the preparation stage (0-4℃) and the core stage (-10℃ wind speed control). The formation process of the current candidate variable dataset integrates multi-dimensional variable combinations such as cold storage environmental parameters (evaporation temperature, wind speed), carcass spacing (50cm benchmark value), and weight, covering the control requirements of all elements of aging. The most energy-efficient solution is selected from the qualified parameter combinations to reduce the cooling energy consumption of each ton of aging mutton to below 250W / m³. The minimum energy consumption selection algorithm breaks through the traditional single quality-oriented model, reducing the enterprise's overall cost by 15%-20% (including electricity, refrigerant, equipment loss, etc.). The variable data, spacing data, and energy consumption data of each iteration are fully recorded to form a digital twin of the process, providing data support for subsequent optimization.

[0023] Example 2:

[0024] Please see Figure 2 This embodiment discloses a precise aging and energy consumption optimization control system for mutton based on multi-parameter adjustment. The system can implement the method of the above embodiment, including a mutton carcass initial data acquisition module, a mutton carcass experimental variable setting and data division module, an aging stage end index mapping model construction and training module, a mutton carcass experimental test data mapping adjustment module, an aging stage end index mapping model testing module, a current aging variable spacing adjustment module, and a minimum energy consumption aging variable screening module. The initial data acquisition module for sheep carcasses collects the average weight data and initial arrangement spacing data of the sheep carcasses to be de-acidified and the sheep carcass experimental group. The module for setting variables and dividing data in the sheep carcass experiment sets the variables to be adjusted in the preparation stage and the core aging stage of mutton aging, and trains the average weight data and initial arrangement spacing data of the sheep carcass experimental group in S1, and divides the test data. The training module for constructing and training the end-of-stage index mapping model is based on the training data of the sheep carcass experimental group divided in S2. It constructs and trains the prepared end-of-stage index mapping model set and the trained core end-of-stage index mapping model set. The sheep carcass experimental test data mapping and adjustment module inputs the test data of the sheep carcass experimental groups divided in S2 into two trained index mapping model sets in S3 for mapping, and adjusts the variables to be adjusted in the preparation stage and the core acid removal stage during the test based on the mapping results. The test module for the end index mapping model of the acid removal stage uses the variables of the preparation stage and the core acid removal stage in S4, which were adjusted in the test process, for the actual acid removal operation in the test process; and adjusts the two trained index mapping models in S3 according to the operation results to obtain the final preparation end index mapping model set and the final core acid removal end index mapping model set. The current acid removal variable spacing adjustment module inputs the average weight data, initial arrangement spacing data and corresponding acid removal variable data of the mutton to be acidified in S1 into the two final index mapping model sets in S5 for mapping. Based on the mapping results, the variables and spacing of the mutton to be acidified during the acid removal process are adjusted until all index data meet the preset conditions after the acid removal is completed. The minimum energy consumption acid removal variable screening module repeats S6 multiple times and selects the variable data of the acid removal process with the lowest energy consumption as the current final variable dataset.

[0025] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0026] The preferred embodiments of the invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A method for optimizing energy consumption in the precise aging process of meat based on multi-parameter adjustment, characterized in that, Includes the following steps: S1. Collect the weight and initial spacing data of the meat to be de-acidified and the corresponding carcasses in the experimental group; S2. Set the variables to be adjusted for the preparation stage and the core aging stage of meat aging, and train the weight and initial arrangement spacing data in S1, and divide the test data. S3. Based on the training data divided in S2, construct and train a well-trained mapping model for the end indicators of the preparation stage and the core acid removal stage. S4. Input the test data divided in S2 into the corresponding end index mapping model in S3 for mapping and adjust the variables to be adjusted in the preparation stage and the core acid removal stage during the test. S5. Apply the adjusted variables in S4 to the actual acid removal operation during the test process; adjust the end index mapping model in S3 according to the operation results to obtain the final end index mapping model. S6. Input the weight, initial spacing data and corresponding aging variable data of the meat to be aged in S1 into the final end index mapping model in S5 for mapping and adjust the variables and spacing of the meat to be aged during the aging process. S7. Repeat S6 multiple times, and select the variable data of the acid removal process with the lowest energy consumption as the current final variable dataset.

2. The method for optimizing and controlling the energy consumption of meat aging based on multi-parameter adjustment according to claim 1, characterized in that, S1 includes the following steps: S11. Obtain multiple pieces of meat to be pickled and measure the initial spacing between the carcasses corresponding to the multiple pieces of meat to be pickled and the average weight data of the carcasses, to obtain the initial carcass spacing set and the carcass weight data set; then set up several carcass experimental groups, each carcass experimental group including several carcasses, to obtain the carcass experimental group set; Several types of quality indicators for meat after aging are defined to obtain a set of quality indicator types after aging; the set of quality indicator types after aging includes carcass pH value, carcass temperature drop rate, carcass meat color, carcass juice loss rate, and carcass odor substance ratio; S12. Collect the average weight data of multiple carcasses in each carcass experimental group in the carcass experimental group set and the initial arrangement spacing data between them to obtain the carcass experimental weight dataset and the carcass experimental initial arrangement spacing dataset; then, according to the existing meat aging standard process, divide the meat aging process into stages to obtain the meat aging stage set; the meat aging stage set includes the preparation stage and the core aging stage.

3. The method for optimizing and controlling the energy consumption of meat aging based on multi-parameter adjustment according to claim 2, characterized in that, S2 includes the following steps: S21. In conjunction with the meat aging stage set, set the equipment types used in the preparation stage and the core aging stage to obtain a first equipment type set and a second equipment type set; S22. Based on the first equipment type set and the second equipment type set, set the types of variables to be adjusted for the preparation stage and the core acid removal stage, and obtain the first variable type set and the second variable type set; S23. Divide the carcass experimental set into training and testing groups to obtain carcass experimental training groups and carcass experimental testing groups; in conjunction with the carcass experimental weight dataset and the carcass experimental initial arrangement spacing dataset, obtain the carcass weight and initial arrangement spacing data corresponding to the carcass experimental training group and the carcass experimental testing group to obtain the carcass experimental training weight dataset, the carcass experimental testing weight dataset, the carcass experimental training initial arrangement spacing dataset, and the carcass experimental testing initial arrangement spacing dataset.

4. The method for optimizing and controlling the energy consumption of meat aging based on multi-parameter adjustment according to claim 3, characterized in that, S3 includes the following steps: S31. Construct a mapping model between the carcass quality index data before the start of the preparation stage and the core acid removal stage, the applied variable data during the process, the carcass experimental training weight data, and the initial arrangement spacing data of the carcass experimental training, and the various index data after the corresponding stage. Train the mapping model based on the carcass experimental training group, the first variable type set, the quality index type set after acid removal, the carcass experimental training weight dataset, and the initial arrangement spacing dataset of the carcass experimental training to obtain the trained preparation end index mapping model set and the trained core acid removal end index mapping model set.

5. The method for optimizing and controlling the energy consumption of meat aging based on multi-parameter adjustment according to claim 4, characterized in that, S4 includes the following steps: S41. Before the start of the preparation stage, collect the quality index data of multiple carcasses in each carcass experimental test group to obtain the initial quality index dataset of carcass test; obtain the acid removal application variable data corresponding to the preparation stage and the core acid removal stage of the carcass experimental test group to obtain the first test initial variable dataset and the second test initial variable dataset. S42. Input the initial index data of each initial index in the initial index dataset of carcass test quality, the weight dataset of carcass experiment test, the initial variable dataset of the first test, and the initial spacing dataset of carcass experiment test into the corresponding mapping model in the trained set of ready-to-end index mapping models for mapping; adjust the initial spacing dataset of carcass experiment test and the initial variable dataset of the first test according to the mapping results to obtain the final variable dataset of the first test. S43. Input the data of each indicator after the preparation stage in S42, the data of carcass arrangement spacing, the dataset of initial variables for the second test, the dataset of carcass experimental test preparation arrangement spacing, and the dataset of carcass experimental test weight into the corresponding mapping model in the set of trained core acid removal end indicator mapping models for mapping, and obtain the dataset of quality indicators for the second test. Based on the second test quality index dataset, the carcass experimental test preparation arrangement spacing dataset and the second test initial variable dataset are adjusted to obtain the second test final variable dataset.

6. The method for optimizing and controlling the energy consumption of meat aging based on multi-parameter adjustment according to claim 5, characterized in that, S5 includes the following steps: S51. Use the carcass spacing data and the first test final variable dataset after the preparation stage in S42 in the actual preparation stage of the carcass experimental test group; after the preparation stage is completed, the first actual quality index dataset is obtained. Based on the first actual quality index dataset, determine whether to return to S31 to continue training the trained ready-to-end index mapping model set, and repeat S41, S42, S43 and S51 to obtain the final ready-to-end index mapping model set. S52. Use the carcass spacing data and the second test final variable dataset after the core acid removal stage in S43 in the actual core acid removal stage of the carcass experimental test group; after the core acid removal stage, obtain the second actual quality index dataset. Based on the second actual quality index dataset, determine whether to return to S31 to continue training the trained core acid removal end index mapping model set, and repeat S41, S42, S43, S51 and S52 to obtain the final core acid removal end index mapping model set.

7. The method for optimizing and controlling the energy consumption of meat aging based on multi-parameter adjustment according to claim 6, characterized in that, S6 includes the following steps: S61. Based on the set of quality indicators after aging, collect the quality indicator data of the carcasses corresponding to the current multiple pieces of meat to be aged, and obtain the current initial quality indicator dataset. S62. Input the data of each indicator in the current initial quality indicator dataset, the data of acid removal applied during the preparation stage, the initial carcass spacing set to be acid removed, and the carcass weight dataset to be acid removed into the model corresponding to the final preparation and completion indicator mapping model set for mapping. Adjust the initial carcass spacing set to be acid removed and the data of acid removal applied during the preparation stage according to the mapping results. S63. Input the various indicator data after the preparation stage in S62, the carcass spacing data, the acid removal variables applied in the core acid removal stage, and the carcass weight dataset to be acid removed into the model corresponding to the final core acid removal end indicator mapping model set for mapping. Adjust the carcass spacing set to be acid removed and the acid removal variable data applied in the core acid removal stage according to the mapping results.

8. The method for optimizing and controlling the energy consumption of meat aging based on multi-parameter adjustment according to claim 7, characterized in that, S7 includes the following steps: S71. Repeat S62 and S63 multiple times to obtain multiple sets of data on the acid removal applied variables and the corresponding carcass spacing data corresponding to the preparation stage and the core acid removal stage in the current acid removal process. These are denoted as the current candidate variable dataset and the current candidate carcass spacing dataset. S72. Evaluate the energy consumption of the acid removal application variable data corresponding to the preparation stage and the core acid removal stage in each group of the current acid removal process in the current candidate variable dataset; select the current candidate variable data corresponding to the minimum energy consumption data as the current final variable dataset. S73. Combine the currently selected final variable dataset with the carcass spacing data in the corresponding current candidate carcass spacing dataset to perform acid removal processing on the current multiple pieces of meat to be acid removed.

9. A method for precisely controlling the energy consumption of mutton aging, characterized in that; The method for precise acid removal and energy consumption optimization control of meat based on multi-parameter adjustment as described in any one of claims 1-8 was adopted.

10. A system for implementing the method for precise aging and energy consumption optimization control of meat based on multi-parameter adjustment as described in any one of claims 1-8.

Images