A method for preparing a chicken breast sensory analog based on morphology and texture and a chicken breast sensory analog
Patent Information
- Application Number
- CN202611496196.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-24
- Publication Date
- 2026-10-30
AI Technical Summary
整个配方优化过程以实现材料内部均质、硬度精准可控为核心，仅能体现单一硬度数值，无法复现鸡胸肉特有的外形轮廓、自然下垂形态与粘弹性质地触感，不具备鸡胸肉感官分级的对标参考价值，不能直接移植到鸡胸肉的感官评判当中，无法满足木质化鸡胸肉分级的人员标准化培训与生产现场对标需求，因此亟需研发能够高度还原真实鸡胸肉感官特征的专用模拟物
[0033]本发明提供了一种基于形态和质地的鸡胸肉感官模拟物制备方法及鸡胸肉感官模拟物，通过对鸡胸肉进行形态基准和硬度基准测定，构建制备材料重量与水凝胶成品硬度的回归关系模型和BP-ANN模型，通过回归关系模型确了各原料的作用规律与交互效应，优选最佳的配方组合，通过BP-ANN模型提升了非线性体系的预测精度进行交叉验证，确保最优配方的准确性与可重复性。在制备好鸡胸肉感官模拟物后，将其作为参照物，无需深厚专业背景即可快速掌握分级尺度；未经训练的人员经短期培训即可达到较高的分级准确率，且评定员间一致性优秀，可直接应用于屠宰生产线的人工快速分级，也可用于分级人员的标准化培训。
Smart Images
Figure CN123008433A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the technical field of chicken breast quality grading, and more particularly to a method for preparing a chicken breast sensory analog based on morphology and texture, and the chicken breast sensory analog. Background Technology
[0002] In the traditional poultry processing industry, chicken breast quality is typically graded by manual palpation. Trained graders assess the firmness of the chicken breast by touch. This grading method heavily relies on the grader's personal experience. In practice, there is a lack of standardized physical references among graders from different production companies and regions. Grading criteria often depend on the experience of individual teams and literature descriptions, leading to significant differences in judgment standards among different evaluators. This results in poor consistency in grading results for the same sample, and the industry lacks standardized sensory evaluation references. Currently available standardized hardness reference samples (such as rubber hardness standard blocks) are mainly used for hardness measurement calibration, instrument verification, and traceability of industrial homogeneous materials such as rubber and plastics. These are mostly internally homogeneous rigid materials. Their research and development is based on rubber matrix (natural or synthetic rubber) as the base material, combined with functional additives such as vulcanizing agents, accelerators, reinforcing fillers, and softeners. The target hardness value is the sole optimization objective. Through methods such as single-factor gradient experiments and orthogonal experimental design, the dosage ratio of each component is gradually adjusted. The vulcanization characteristics test and finished product hardness test are used for verification at the same time. After repeated iterations, the formula combination that can stably achieve the target hardness is selected. The entire formula optimization process focuses on achieving internal homogeneity and precise control of hardness in the material. It can only reflect a single hardness value and cannot reproduce the unique shape, natural drooping shape, and viscoelastic texture of chicken breast. It does not have the benchmark reference value for the sensory grading of chicken breast and cannot be directly transferred to the sensory evaluation of chicken breast. It cannot meet the needs of standardized training for personnel and benchmarking on the production site for grading lignified chicken breast. Therefore, there is an urgent need to develop a special simulator that can highly reproduce the sensory characteristics of real chicken breast. Summary of the Invention
[0003] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method for preparing a chicken breast sensory analog based on morphology and texture, as well as the chicken breast sensory analog. By preparing the chicken breast sensory analog as a permanent reference standard, the grading standards of graders can be effectively unified, the grading senses of graders can be calibrated, and the error of manual palpation grading can be reduced.
[0004] To achieve the above objectives, the present invention is implemented using the following technical solution:
[0005] In a first aspect, the present invention provides a method for preparing a sensory analog of chicken breast based on morphology and texture, comprising:
[0006] The morphological and hardness benchmarks of chicken breast with target lignification grade were determined. The morphological benchmarks included thickness benchmarks, external dimensions benchmarks, and tail drooping angle benchmarks.
[0007] Using the weight of each material used in the preparation of the chicken breast sensory simulant as a continuous independent variable and the thickness as a classification factor, an experiment was conducted using a Box-Behnken design. The hardness of the hydrogel product was used as the response value to construct a regression model between the weight of the preparation materials and the hardness of the hydrogel product.
[0008] With the hardness benchmark as the optimization objective, boundary constraints are set for each of the continuous independent variables, and the regression relationship model is optimized and solved to obtain the optimal weight combination for the prepared material.
[0009] A simulant mold is prepared based on the morphological reference, and a sensory simulant of chicken breast with a target lignification grade is prepared according to the optimal weight combination of the simulant mold and the preparation materials using the preparation process.
[0010] Optionally, the determination of the morphological reference includes:
[0011] Take a target number of chicken breasts of the target lignification grade, and measure its thickness, external dimensions, and tail drooping angle when the head is held naturally in the air by pinching the side of the head. Take the average values of the thickness, external dimensions, and tail drooping angle as the thickness reference, external dimensions reference, and tail drooping angle reference, respectively.
[0012] The determination of the hardness standard includes:
[0013] Take the target number of chicken breasts of the target lignification grade, measure their hardness in the Total Texture Analysis (TPA) mode, and take the average value of the hardness as the hardness benchmark.
[0014] Optionally, the preparation materials include high-acyl gellan gum, bovine bone gelatin, calcium chloride, and ultrapure water pre-dissolved with anhydrous calcium chloride, with the weights of the high-acyl gellan gum, bovine bone gelatin, and calcium chloride being continuous independent variables, and a fixed volume of ultrapure water pre-dissolved with anhydrous calcium chloride being set.
[0015] Optionally, the boundary constraint of the high acyl gellan gum is 1 to 2 g, the boundary constraint of the bovine bone gelatin is 9 to 27 g, the boundary constraint of the calcium chloride is 0 to 0.4 g, and the fixed volume of the ultrapure water pre-dissolved with anhydrous calcium chloride is 240 mL.
[0016] Optionally, the preparation of the sensory simulant of chicken breast with the target lignification grade based on the optimal weight combination of the simulant mold and the preparation materials according to the preparation process includes:
[0017] Prepare the materials according to the optimal weight combination of the materials, add a fixed volume of ultrapure water pre-dissolved with anhydrous calcium chloride, seal and place in a 90℃ water bath, and stir magnetically at 600 rpm for 20 min until the powder is completely dissolved to form a mixture.
[0018] The mixture was rapidly injected into the simulant mold, transferred to a 4°C environment and left to stand for 12 hours. After full gelation, the mixture was demolded to obtain a sensory simulant of chicken breast with the target lignification grade.
[0019] Optionally, optimizing the regression model to obtain the optimal weight combination for preparing the material includes:
[0020] The steepest descent method is used to iteratively search within the boundary constraints of each continuous independent variable to obtain the weight combinations in which the output hardness of the regression model deviates from the hardness benchmark by less than the deviation threshold as candidate weight combinations.
[0021] Taking into account the ease of weighing each preparation material and the simplicity of the preparation process, the candidate weight combinations are optimized and fine-tuned to determine the optimal weight combination.
[0022] Optionally, it also includes constructing a BP-ANN model, wherein the input of the BP-ANN model is the normalized data of the weight and thickness of each prepared material, and the output is the hardness of the hydrogel product;
[0023] The training process of the BP-ANN model includes:
[0024] Based on the experimental data from the Box-Behnken design, an expanded dataset was generated.
[0025] The dataset is divided into a training set, a validation set, and a test set according to a preset ratio. By comparing the mean square error of the validation set corresponding to different numbers of hidden layer neurons, the optimal number of hidden layer neurons is determined to be 8.
[0026] The BP-ANN model with the network topology determined is trained using the training set, the validation set, and the test set to obtain a trained BP-ANN model.
[0027] The trained BP-ANN model is used for validation during the construction of the regression relationship model and for validation of the optimal weight combination of the prepared materials.
[0028] Optionally, the expanded dataset generated based on the experimental data from the Box-Behnken design includes:
[0029] Single-factor gradient experiment data: Take one continuous independent variable as the target continuous independent variable, fix the remaining continuous independent variables as the intermediate level, set dense gradient points in the boundary constraints of the target continuous independent variable, form a weight combination, and determine the hardness of its hydrogel product according to the same procedure.
[0030] Preliminary experimental verification data: The weight combination of the boundary constraints of each of the continuous independent variables was taken and the hardness of the hydrogel product was measured according to the same procedure.
[0031] Secondly, the present invention provides a chicken breast sensory analogue, which is prepared by the chicken breast sensory analogue preparation method described above.
[0032] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:
[0033] This invention provides a method for preparing a morphological and textural sensory simulant of chicken breast, as well as the simulant itself. By measuring the morphological and hardness benchmarks of chicken breast, a regression model and a BP-ANN model are constructed to correlate the weight of the prepared materials with the hardness of the hydrogel product. The regression model clarifies the interaction and interaction of each raw material, allowing for the selection of the optimal formulation combination. The BP-ANN model improves the prediction accuracy of the nonlinear system, and cross-validation ensures the accuracy and repeatability of the optimal formulation. After preparing the chicken breast sensory simulant, it can be used as a reference, allowing for rapid mastery of grading standards without requiring extensive professional background. Untrained personnel can achieve high grading accuracy with short-term training, and there is excellent consistency among assessors. It can be directly applied to rapid manual grading in slaughtering production lines, and can also be used for standardized training of grading personnel. Attached Figure Description
[0034] Figure 1 This is a schematic flowchart of a method for preparing a sensory analog of chicken breast based on morphology and texture, provided in an embodiment of the present invention.
[0035] Figure 2 This is a schematic diagram comparing the performance of the BP-ANN model and the Regression Relationship Model (RSM) provided in this embodiment of the invention. Detailed Implementation
[0036] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.
[0037] Example 1
[0038] like Figure 1 As shown, this embodiment of the invention provides a method for preparing a sensory mimic of chicken breast based on morphology and texture, comprising the following steps:
[0039] Step S1: Determine the morphological and hardness benchmarks for the target lignification grade chicken breast. The morphological benchmarks include the thickness benchmark, the external dimensions benchmark, and the tail drooping angle benchmark.
[0040] Specifically, in this embodiment, the lignification grade of chicken breast includes normal chicken breast (NORM), lightly lignified chicken breast (MILD), moderately lignified chicken breast (MOD), and heavily lignified chicken breast (SEV). In this embodiment, sensory simulants of chicken breast are prepared for each type of lignification grade, and the category currently being prepared is marked as the target lignification grade of chicken breast.
[0041] Fresh chicken breast samples were collected from commercial broiler slaughterhouses. Professionals with many years of experience in lignified chicken breast research used manual palpation to classify the samples into the above four grades based on cranial hardness, surface protrusion, and overall morphology. Twenty representative samples from each grade were selected, wrapped in ice packs, and transported to the laboratory at 4°C for morphological and hardness benchmark measurements.
[0042] The determination of morphological standards includes:
[0043] Take the target amount of chicken breast of the target lignification grade, and measure its thickness, external dimensions, and tail droop angle when the head is held naturally in the air. Take the average values of thickness, external dimensions, and tail droop angle as the thickness benchmark, external dimensions benchmark, and tail droop angle benchmark, respectively. The tail droop angle is an important morphological indicator reflecting the viscoelasticity and hardness of the meat. It can be used to guide the overall length-width ratio and thickness gradient design of the mold to ensure the matching of the basic morphological proportions.
[0044] The determination of hardness standards includes:
[0045] Take the target number of chicken breasts of the target lignification grade, measure their hardness in the Total Texture Analysis (TPA) mode, and take the average hardness as the hardness benchmark.
[0046] Experimental measurements showed that the hardness of normal chicken breast (NORM), slightly lignified chicken breast (MILD), moderately lignified chicken breast (MOD), and heavily lignified chicken breast (SEV) were 4159.18 gf, 6671.28 gf, 9432.31 gf, and 12039.55 gf, respectively.
[0047] Step S2: Using the weight of each material used in the preparation of the chicken breast sensory simulant as a continuous independent variable and thickness as a classification factor, an experiment was conducted using a Box-Behnken design. The hardness of the hydrogel product was used as the response value to construct a regression model between the weight of the preparation materials and the hardness of the hydrogel product.
[0048] Specifically, in this embodiment, the preparation materials include high acyl gellan gum, bovine bone gelatin, calcium chloride, and ultrapure water pre-dissolved with anhydrous calcium chloride. The weights of high acyl gellan gum, bovine bone gelatin, and calcium chloride are continuous independent variables, while a fixed volume of ultrapure water pre-dissolved with anhydrous calcium chloride is set.
[0049] Experiments revealed that the boundary constraints for high-acyl gellan gum were 1-2 g, for bovine bone gelatin 9-27 g, and for calcium chloride 0-0.4 g. 1-2 g of high-acyl gellan gum controlled the gel's toughness and drooping behavior; below 1 g, it failed to maintain structural integrity during bending, while above 2 g resulted in excessive stiffness and loss of the natural drape of the tissue. Below 9 g of bovine bone gelatin, the hydrogel's mechanical strength was insufficient to match the lowest grade of meat firmness; above 27 g, the gel became too brittle, losing the characteristic supple texture of chicken breast. 0-0.4 g of calcium chloride was used to control the ionic cross-linking density; exceeding 0.4 g caused excessive cross-linking, leading to brittle gel and loss of the viscoelasticity characteristic of meat. The fixed volume of ultrapure water pre-dissolved with anhydrous calcium chloride was 240 mL.
[0050] Specifically, in this embodiment, a three-factor, three-level Box-Behnken design was used to conduct the experiment, and the coding levels of the independent variables are shown in Table 1.
[0051] Table 1: Encoding Levels of Independent Variables
[0052]
[0053] Using thickness as a classification factor is to analyze the interaction effect between thickness and the weight of each preparation material, so that the regression model can calculate the formulation hardness at different thicknesses and ensure the formulation accuracy of each grade of simulants.
[0054] Experimental Design: A total of 60 experiments were designed, including factor points and center points. The experimental execution order was completely randomized to eliminate systematic errors. Hydrogels were prepared and their hardness measured using a standardized process in each experiment to obtain the core experimental dataset. Factor points refer to experimental points obtained by pairwise combinations of factors at -1 and 1 levels, used to fit the interaction and secondary effects between factors. Center points refer to experimental points where all factors are at a 0 level, with multiple replicates, used to estimate the pure error of the experiment, test the model's lack of fit, and improve the model's reliability.
[0055] This embodiment provides a multivariate regression model to construct a regression relationship between the weight of the preparation materials and the hardness of the hydrogel product. The regression model adopts a quadratic polynomial regression equation. The significance, goodness of fit and lack of fit of the model are verified by analysis of variance (ANOVA). The main effect and interaction effect of the weight of each preparation material on the hardness of the hydrogel product are clarified to ensure that the model can be used for subsequent formulation optimization.
[0056] Step S3: Using near-hardness benchmark as the optimization objective, set boundary constraints for each continuous independent variable, optimize and solve the regression model to obtain the optimal weight combination for the prepared material, specifically including:
[0057] The steepest descent method is used to iteratively search within the boundary constraints of each continuous independent variable to obtain the weight combination with the output hardness of the regression model and the hardness benchmark deviation less than the deviation threshold as the candidate weight combination.
[0058] Taking into account the ease of weighing each material and the simplicity of the preparation process, the candidate weight combinations were optimized and fine-tuned to determine the optimal weight combination.
[0059] This embodiment was verified through experiments, and the final optimal weight combination is shown in Table 2.
[0060] Table 2: Optimal Weight Combinations
[0061]
[0062] Step S4: Prepare a simulation mold based on morphological benchmarks, and prepare a sensory simulation of chicken breast of the target lignification grade according to the optimal weight combination of the simulation mold and the preparation materials.
[0063] The preparation process is as follows:
[0064] Prepare the materials according to the optimal weight combination of the materials, add a fixed volume of ultrapure water pre-dissolved with anhydrous calcium chloride, seal and place in a 90℃ water bath, and stir magnetically at 600 rpm for 20 min until the powder is completely dissolved to form a mixture.
[0065] The mixture was quickly injected into the simulant mold and placed at 4°C for 12 hours. After full gelation, the mold was removed to obtain a sensory simulant of chicken breast with the target lignification level.
[0066] To improve the accuracy of the sensory simulants prepared in this embodiment, this embodiment also includes constructing a BP-ANN model. The input of the BP-ANN model is the normalized data of the weight and thickness of each prepared material, and the output is the hardness of the hydrogel product.
[0067] The training process of the BP-ANN model includes:
[0068] (1) Based on the experimental data from the Box-Behnken design experiment, expand and generate a dataset:
[0069] Single-factor gradient experiment data: Take one continuous independent variable as the target continuous independent variable, fix the remaining continuous independent variables as the intermediate level, set dense gradient points in the boundary constraints of the target continuous independent variable, form a weight combination, and determine the hardness of its hydrogel product according to the same procedure.
[0070] Preliminary experimental verification data: Take the weight combination of the boundary constraints of each continuous independent variable and measure the hardness of the hydrogel product according to the same procedure;
[0071] Specifically, in this embodiment, based on the aforementioned 60 sets of experiments, 160 additional sets of independent experimental data were added. For bovine bone gelatin as the target continuous independent variable, four additional gradients of 12, 15, 21, and 24 g were set; for high-acyl gellan gum as the target continuous independent variable, an additional gradient of 0.5 g was set; calcium chloride was not required as a target continuous independent variable and no additional gradient was set. Hydrogel samples were prepared in four types of simulant molds at each gradient point, and hardness was measured according to a unified procedure. Finally, these were merged to form a complete dataset of 220 sets, used to expand the data coverage density and improve the fitting accuracy and generalization ability of the neural network.
[0072] (2) Divide the dataset into training set, validation set and test set according to the preset ratio (generally 7:1.5:1.5). By comparing the mean square error of the validation set corresponding to different numbers of hidden layer neurons, determine the optimal number of hidden layer neurons as 8.
[0073] (3) The BP-ANN model with the network topology determined is trained by training set, validation set and test set to obtain the trained BP-ANN model.
[0074] The trained BP-ANN model is used for validation during the construction of regression relationship models and for validation of the optimal weight combination of prepared materials.
[0075] like Figure 2 As shown, this implementation method uses 10-fold cross-validation to verify the robustness of the model. The coefficient of determination (R²), root mean square error (RMSE), and absolute mean deviation (AAD%) are used as core indicators to compare the calculation accuracy of the BP-ANN model and the regression relationship model (RSM). The verification results show that the BP-ANN model has better prediction accuracy in nonlinear formulation systems and can cross-validate the reliability of the RSM formulation optimization results.
[0076] Example 2
[0077] This invention provides a chicken breast sensory analogue, prepared using the method described in Embodiment 1 above. The lignification grades of chicken breast include normal chicken breast (NORM), lightly lignified chicken breast (MILD), moderately lignified chicken breast (MOD), and heavily lignified chicken breast (SEV).
[0078] After preparation, the sensory analogue of chicken breast can be validated in multiple dimensions to ensure its accuracy.
[0079] 1. Mechanical property verification: TPA hardness, compressive force, stress-strain, and compression creep tests were conducted to compare the mechanical parameters of the simulated material and real chicken breast at the same grade. The verification results showed that there was no significant difference in hardness and compressive force between the simulated material and the real sample at each grade, and the stress-strain trend and creep behavior were highly consistent, meeting the requirements for texture simulation.
[0080] 2. Morphological verification: The scenario of pinching the side of the skull and letting it hang naturally during artificial grading was simulated. The drooping angle of the simulated tail was measured and compared with the real sample of the same grade. The verification results showed that the relative error of the drooping angle of the four grades was ≤2%, and the morphological drooping characteristics were highly consistent.
[0081] 3. Microstructure verification: Scanning electron microscopy (SEM) and Fourier transform infrared spectroscopy (FTIR) were used to verify the gradient of the double network structure from loose to dense in the quaternary analog, as well as the gradient changes of intermolecular hydrogen bonds and ionic bonds, thus explaining the formation mechanism of the texture gradient at the microscopic level.
[0082] The four grades of chicken breast sensory simulants that have passed the above verification can be used as a unified physical standard for sensory grading of lignified chicken breast and personnel training.
[0083] The final application of the technical solution uses a four-level simulant as a physical reference to achieve standardized grading of lignified chicken breast and output clear grading results.
[0084] 1. Standardized Training for Assessors: Recruit personnel with no experience in assessing lignified chicken breast and conduct a 2-day standardized training: Day 1 includes theoretical instruction, explaining the causes and grading characteristics of lignified chicken breast, while assessors establish a clear understanding of grading standards by touching and observing sensory simulants of four levels of chicken breast; Day 2 involves calibration practice using coded real chicken breast samples, with assessors using the chicken breast sensory simulants as a reference for grading. Immediate feedback and corrections are provided after each round of practice, and personnel with a final grading accuracy rate of ≥85% are qualified for formal grading.
[0085] 2. Formal grading operation procedure: A complete set of sensory simulants of chicken breast in four grades are placed at the grading site as permanent reference standards; the assessor palpates the cranial region of the chicken breast to be tested to sense its firmness, and lifts the cranial end to observe its natural drooping shape. The sample to be tested is compared with the four grades of simulants one by one, and the sample to be tested is determined to be the grade with the closest characteristics; the determination time for a single sample is ≤40s, which is suitable for the rapid grading operation requirements of the slaughtering production line.
[0086] 3. Verification of the grading effect: The evaluation results of 400 real blind test samples (100 samples per grade) by 20 evaluators are shown in Table 3.
[0087] Table 3: Evaluation results of 20 evaluators on 400 real blind test samples
[0088]
[0089] The overall classification accuracy reached 90.75%, and the intragroup correlation coefficient (ICC) among assessors reached 0.89, demonstrating excellent assessment consistency. This proves that the simulant can effectively unify classification standards and reduce human judgment errors.
[0090] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.
Claims
1. A method for preparing a sensory mimic of chicken breast based on morphology and texture, characterized in that, include: The morphological and hardness benchmarks of chicken breast with target lignification grade were determined. The morphological benchmarks included thickness benchmarks, external dimensions benchmarks, and tail drooping angle benchmarks. Using the weight of each material used in the preparation of the chicken breast sensory simulant as a continuous independent variable and the thickness as a classification factor, an experiment was conducted using a Box-Behnken design. The hardness of the hydrogel product was used as the response value to construct a regression model between the weight of the preparation materials and the hardness of the hydrogel product. With the hardness benchmark as the optimization objective, boundary constraints are set for each of the continuous independent variables, and the regression relationship model is optimized and solved to obtain the optimal weight combination for the prepared material. A simulant mold is prepared based on the morphological reference, and a sensory simulant of chicken breast with a target lignification grade is prepared according to the optimal weight combination of the simulant mold and the preparation materials using the preparation process.
2. The method for preparing a sensory mimic of chicken breast based on morphology and texture according to claim 1, characterized in that, The determination of the morphological reference includes: Take a target number of chicken breasts of the target lignification grade, and measure its thickness, external dimensions, and tail drooping angle when the head is held naturally in the air by pinching the side of the head. Take the average values of the thickness, external dimensions, and tail drooping angle as the thickness reference, external dimensions reference, and tail drooping angle reference, respectively. The determination of the hardness standard includes: Take the target number of chicken breasts of the target lignification grade, measure their hardness in the Total Texture Analysis (TPA) mode, and take the average value of the hardness as the hardness benchmark.
3. The method for preparing a sensory mimic of chicken breast based on morphology and texture according to claim 1, characterized in that, The preparation materials include high acyl gellan gum, bovine bone gelatin, calcium chloride, and ultrapure water pre-dissolved with anhydrous calcium chloride. The weights of the high acyl gellan gum, bovine bone gelatin, and calcium chloride are continuous independent variables, while a fixed volume of ultrapure water pre-dissolved with anhydrous calcium chloride is set.
4. The method for preparing a sensory mimic of chicken breast based on morphology and texture according to claim 3, characterized in that, The boundary constraint of the high acyl gellan gum is 1 to 2 g, the boundary constraint of the bovine bone gelatin is 9 to 27 g, the boundary constraint of the calcium chloride is 0 to 0.4 g, and the fixed volume of the ultrapure water pre-dissolved with anhydrous calcium chloride is 240 mL.
5. The method for preparing a sensory mimic of chicken breast based on morphology and texture according to claim 3, characterized in that, The preparation of sensory simulants of chicken breast with the target lignification grade based on the optimal weight combination of the simulant mold and the preparation materials according to the preparation process includes: Prepare the materials according to the optimal weight combination of the materials, add a fixed volume of ultrapure water pre-dissolved with anhydrous calcium chloride, seal and place in a 90℃ water bath, and stir magnetically at 600 rpm for 20 min until the powder is completely dissolved to form a mixture. The mixture was rapidly injected into the simulant mold, transferred to a 4°C environment and left to stand for 12 hours. After full gelation, the mixture was demolded to obtain a sensory simulant of chicken breast with the target lignification grade.
6. The method for preparing a sensory mimic of chicken breast based on morphology and texture according to claim 1, characterized in that, The optimization and solution of the regression model to obtain the optimal weight combination for the prepared materials includes: The steepest descent method is used to iteratively search within the boundary constraints of each continuous independent variable to obtain the weight combinations in which the output hardness of the regression model deviates from the hardness benchmark by less than the deviation threshold as candidate weight combinations. Taking into account the ease of weighing each preparation material and the simplicity of the preparation process, the candidate weight combinations are optimized and fine-tuned to determine the optimal weight combination.
7. The method for preparing a sensory mimic of chicken breast based on morphology and texture according to claim 1, characterized in that, It also includes constructing a BP-ANN model, the input of which is the normalized data of the weight and thickness of each prepared material, and the output is the hardness of the hydrogel product; The training process of the BP-ANN model includes: Based on the experimental data from the Box-Behnken design, an expanded dataset was generated. The dataset is divided into a training set, a validation set, and a test set according to a preset ratio. By comparing the mean square error of the validation set corresponding to different numbers of hidden layer neurons, the optimal number of hidden layer neurons is determined to be 8. The BP-ANN model with the network topology determined is trained using the training set, the validation set, and the test set to obtain a trained BP-ANN model. The trained BP-ANN model is used for validation during the construction of the regression relationship model and for validation of the optimal weight combination of the prepared materials.
8. The method for preparing a morphology and texture-based sensory analog of chicken breast according to claim 7, characterized in that, The dataset generated by expanding upon the experimental data from the Box-Behnken design includes: Single-factor gradient experiment data: Take one continuous independent variable as the target continuous independent variable, fix the remaining continuous independent variables as the intermediate level, set dense gradient points in the boundary constraints of the target continuous independent variable, form a weight combination, and determine the hardness of its hydrogel product according to the same procedure. Preliminary experimental verification data: The weight combination of the boundary constraints of each of the continuous independent variables was taken and the hardness of the hydrogel product was measured according to the same procedure.
9. A sensory analog of chicken breast, characterized in that, It is prepared using the method for preparing chicken breast sensory analogues as described in any one of claims 1-8.