Preparation process and synergistic regulation method of quick-frozen egg tart

Abstract

The application belongs to the technical field of food processing, and discloses a preparation process and a synergistic regulation method of quick-frozen egg tarts, which comprises the following steps: a. selecting egg tart liquid raw materials, and preparing a targeted modified egg tart liquid through step-by-step dissolution, two-stage homogenization and segmented pasteurization; b. sending pre-prepared flaky puff pastry into an adjustable precision mold and a vacuum suction device to complete shape pressing and molding; c. completing quantitative and accurate filling of tart liquid through a high-precision filling pump and a flow regulating valve; d. completing gradient baking and curing of the egg tarts; e. completing gradient humidity control and anti-condensation cooling, controlling the temperature and relative humidity of the cooling environment in a process, and completing directional cooling of the egg tarts; f. completing modified atmosphere packaging of the cooled egg tarts in a ten-thousand-level clean environment; g. controlling the temperature and wind speed of the quick-freezing environment in a process, completing temperature-controlled freezing of the egg tarts, and directionally regulating the generation size and distribution of ice crystals in the freezing process; and after batch verification and metal detection, the egg tarts are stored in a warehouse, and finally, high freeze-thaw stability quick-frozen egg tart products are obtained.

Description

Technical Field

[0001] This manual relates to the field of food processing technology, and in particular to a preparation process and synergistic control method for quick-frozen egg tarts. Background Technology

[0002] With the large-scale development of the frozen food industry and the upgrading of residents' consumption, pre-baked frozen foods have seen their market size continue to expand due to their core advantages of high convenience, good flavor reproduction, and long shelf life. Egg tarts, as a classic Western-style baked product, have a wide audience and strong market demand. Finished frozen egg tarts can be produced in a centralized, end-to-end factory, requiring only thawing and rebaking at the point of sale, eliminating the need for secondary filling. This significantly reduces the operational threshold and labor costs for retail stores, making them a core product in baking, new retail, and catering channels, with a market growth rate significantly higher than that of semi-finished egg tart combo products.

[0003] The industrial production of frozen egg tarts places stringent demands on the synergy of the entire process, product freeze-thaw stability, batch-to-batch quality consistency, and food safety control capabilities. Currently, existing technologies for the research and production of frozen egg tarts still have significant limitations, primarily in the following aspects:

[0004] 1. Existing technologies mostly focus on fragmented optimization of single processes, failing to form a collaborative process system across the entire process, and thus unable to solve core quality pain points. Some existing technologies only optimize the egg tart filling formula by adding additives such as colloids and modified starches to improve the static stability of the filling itself, without considering the compatibility of the rheological properties of the egg tart filling with the physical properties of the egg tart crust, filling, baking, and quick-freezing processes. This fails to solve problems such as ice crystal growth during frozen storage, water leakage and collapse after thawing and re-baking, and deterioration in taste. Some existing technologies only optimize the puffing and freezing processes of the egg tart crust to improve the stability of the crust when stored alone, without forming a synergistic match with the characteristics of the egg tart filling. This easily leads to defects such as filling penetration, puffing layering damage, and layering of the finished product's taste, resulting in extremely poor quality consistency between the same batch.

[0005] 2. Existing processes lack gradient-based coordinated control across the entire baking, cooling, and quick-freezing process, resulting in poor freeze-thaw stability and short shelf life. Current baking processes often use fixed top and bottom heating temperatures, failing to address the varying needs of crust setting, filling maturation, and surface color setting, easily leading to problems like burnt crust, undercooked filling, and uneven coloring. The cooling process lacks gradient humidity control, causing condensation to form on the surface of hot tarts upon cooling, leading to microbial growth after packaging and significantly shortening shelf life. The quick-freezing process often uses a single-temperature rapid freezing mode, unable to directionally control the size and distribution of ice crystals, easily forming large ice crystals that damage the gel structure of the filling and the layered structure of the crust, resulting in severe deterioration in taste after frozen storage. Currently available quick-frozen tarts typically have a shelf life of less than 6 months at -18℃, failing to achieve long-term stable storage.

[0006] In summary, existing technologies cannot achieve coordinated operation of the entire process of quick-frozen egg tart preparation, and cannot solve the core technical problems of poor freeze-thaw stability of finished products, low quality consistency between batches, low production efficiency and pass rate, and short shelf life. They are also difficult to adapt to the large-scale, flexible industrial mass production needs of modern food factories. Summary of the Invention

[0007] This invention addresses the problems existing in the prior art by proposing a preparation process and synergistic control method for quick-frozen egg tarts. Through the synergistic effect of the process and method, the qualified rate of quick-frozen egg tart products can be improved; single-shift production capacity can be increased and unit product energy consumption can be reduced; the product can be frozen at -18℃ for up to 12 months and refrigerated for up to 45 days. The food safety compliance, sensory quality, and freeze-thaw stability are significantly better than the prior art, and it can fully meet the needs of large-scale, flexible industrial mass production in modern food factories.

[0008] To achieve the above objectives, this application provides the following technical solution:

[0009] Firstly, a process for preparing quick-frozen egg tarts includes the following steps: a. Targeted modification preparation of egg tart filling, selecting egg tart filling raw materials, and preparing targeted modified egg tart filling through stepwise mixing, two-stage homogenization, and segmented pasteurization, while simultaneously detecting the rheological properties parameters of the egg tart filling, and transmitting the detection results to subsequent processes in real time to provide a basis for parameter adaptation in subsequent processes;

[0010] b. Precision shaping of egg tart crusts: The pre-made puff pastry dough is fed into an adjustable precision mold and a vacuum adsorption device to complete the shaping and pressing process. At the same time, the key physical properties of the egg tart crusts, such as thickness, porosity, and edge integrity, are detected, and the test results are transmitted to the subsequent filling process in real time.

[0011] c. Filling and tart crust matching: Based on the rheological properties of the filling and the physical properties of the crust, the filling control parameters are matched and adjusted. The filling is quantitatively and accurately filled by a high-precision filling pump and flow regulating valve. The combined weight, filling weight, and crust weight of a single egg tart are detected simultaneously, and the detection results are transmitted to the subsequent baking process.

[0012] d. Segmented temperature-controlled baking: Based on the characteristic parameters of the egg tart filling, the parameters of the egg tart crust, and the weight parameters passed in the previous step, a multi-segment baking process curve with independent temperature control of the top and bottom heating elements is generated to complete the gradient baking and maturation of the egg tart. The product quality characteristic parameters are detected simultaneously during the baking process, and the baking curve is corrected in real time based on the detection results.

[0013] e. Gradient humidity control and anti-condensation cooling: Based on the core state parameters of the center temperature and moisture content of the egg tart after baking, a gradient cooling curve is generated to control the temperature and relative humidity of the cooling environment in each process to complete the directional cooling of the egg tart. At the same time, the appearance and shape parameters of the product are detected during the cooling process.

[0014] f. Aseptic modified atmosphere packaging: After cooling, the egg tarts are sealed in a Class 10,000 clean environment, and a full range of verifications of packaging compliance and product appearance are performed simultaneously, eliminating substandard products.

[0015] g. Stepped ice crystal control quick-freezing and finished product warehousing: Based on the initial temperature and moisture content parameters of the packaged egg tarts, a stepped temperature control quick-freezing curve is generated. The temperature and wind speed of the quick-freezing environment are controlled in stages to complete the temperature-controlled freezing of the egg tarts. The size and distribution of ice crystals generated during the freezing process are directionally controlled. After batch verification and metal detection, the tarts are put into storage, and finally, a high freeze-thaw stability quick-frozen egg tart product is obtained.

[0016] Optionally, in step a, the egg tart filling ingredients, by weight, are: 30-35 parts pure milk, 10-15 parts light cream, 15-20 parts egg white liquid, 5-10 parts egg yolk liquid, 10-15 parts white sugar, 5-10 parts whole milk powder, 2-8 parts corn starch, and 2-8 parts citrus fiber. The amount of citrus fiber added is positively correlated with the porosity of the egg tart crust in step b. When the porosity of the egg tart crust is 18%-22%, the amount of citrus fiber added is 2-4 parts; when the porosity of the egg tart crust is 22%-25%, the amount of citrus fiber added is 4-8 parts.

[0017] Optionally, step a, the targeted modification preparation of the egg tart filling, specifically includes:

[0018] a1. Base material premixing: Heat pure milk to 40-60℃, add white sugar, whole milk powder, corn starch and citrus fiber in sequence, stir at medium speed until completely dissolved to obtain base material mixture;

[0019] a2. Egg liquid preparation: Cool the base mixture to 0-10℃, add light cream, egg white liquid, and egg yolk liquid, and stir to obtain the egg tart filling stock solution;

[0020] a3. Two-stage homogenization: The egg tart filling is subjected to two-stage homogenization. The first-stage homogenization pressure is 10-15MPa, the second-stage homogenization pressure is 3-5MPa, and the homogenization temperature is controlled at 4-10℃.

[0021] a4. Segmented pasteurization: The homogenized egg tart mixture is pasteurized in segments. The first pasteurization segment is at a temperature of 60-65℃ for 10-20 minutes; the second pasteurization segment is at a temperature of 70-75℃ for 3-5 minutes.

[0022] a5. Cool and set aside: After sterilization, the egg tart filling is rapidly cooled to 0-4℃ and refrigerated for later use. Simultaneously, the viscosity, solids content, and water content rheological properties of the egg tart filling are tested.

[0023] Optionally, the multi-stage baking process curve in step d includes the sequential execution of the puff pastry setting temperature zone, the internal maturation temperature zone, and the color setting temperature zone, with the total baking time controlled at 14-15 minutes.

[0024] The puff pastry shaping temperature zone is: bottom heat temperature of the oven 270-280℃, top heat temperature of the oven 175-186℃, and baking time of 4-5 minutes.

[0025] The internal ripening temperature zone is: bottom heat temperature of the oven is 260-270℃, top heat temperature of the oven is 166-175℃, baking time is 7-8 minutes, and the center temperature of the egg tart rises to above 90℃.

[0026] The color-setting temperature zone is as follows: bottom heat temperature of the oven is 255-265℃, top heat temperature of the oven is 170-180℃, and the baking time is 2-3 minutes.

[0027] After the entire baking process is completed, the center temperature of the egg tart should be controlled within the range of 95-120℃; the acceptable standard for black spots and foreign objects is: the number of black spots with a diameter ≤0.5mm and not connected is ≤2, and there are no black spots with a diameter >0.5mm.

[0028] Secondly, the present invention provides a method for controlling the precision of food processing, based on the preparation process of quick-frozen egg tarts in the first aspect, comprising the following steps:

[0029] S1. Real-time collection of material property parameters of egg tart filling, physical property parameters of egg tart crust, process operation parameters, and environmental status parameters, synchronous acquisition of test data, and construction of a standardized production dataset in all dimensions;

[0030] S2. Using the characteristic parameters of the egg tart filling and the physical properties of the egg tart crust collected in S1 as input, the target process parameters and execution curve are output through a pre-trained parameter matching model, so as to realize the coordinated linkage and pre-adaptation of parameters in each step of the preparation process.

[0031] S3. Real-time acquisition of image data and physicochemical index data of products in production at each process, and completion of product defect identification, classification and quality compliance judgment through deep learning defect detection model, outputting non-conforming product rejection instructions and quality deviation data;

[0032] S4. Taking the quality deviation data output by S3, the real-time deviation between the process operating parameters and the target parameters as input, the process parameter adjustment amount is output through the adaptive control algorithm and sent to the execution equipment of the corresponding process in real time to correct parameter fluctuations and realize closed-loop quality control of the preparation process.

[0033] S5. Based on a standardized production dataset with full dimensions, the system optimizes multiple core performance indicators of the entire production process by using a multi-objective optimization algorithm to find the global optimal combination of process parameters, which is then distributed to the production line for execution. Simultaneously, the system iteratively optimizes the parameter matching model, defect detection model, and adaptive control algorithm based on newly added production data.

[0034] Optionally, in step S2, the parameter matching model is constructed using a Long Short-Term Memory (LSTM) network, as shown in the following formula:

[0035] Forget Gate Calculation: (1)

[0036] Input gate calculation: (2)

[0037] Cell status update: (3)

[0038] Output gate calculation: (4)

[0039] in, The vector of egg tart filling properties and egg tart crust physical property parameters input at time t; This is the output of the hidden layer at time t-1; , , , These are the weight matrices for each gating control; , , , These are the bias vectors for each gate; It is the sigmoid activation function; For Hadamah accumulation; Let be the cell state at time t; This is the vector of target process parameters for the entire process output by the model at time t.

[0040] Optionally, in S2, the parameter matching model is also coupled with the food rheological constitutive equation, and the rheological properties of the egg tart filling are characterized using the Herschel-Bulkley model, as shown in the following formula:

[0041] (5)

[0042] in, Shear stress; K is the yield stress; K is the consistency coefficient; is the shear rate; n is the flow behavior exponent;

[0043] The migration of moisture inside the egg tart during baking is characterized by Fick's second law diffusion equation, as shown in the following formula:

[0044] (6)

[0045] Where C is the internal moisture content of the egg tart; t is time; and D(T) is the temperature-dependent moisture diffusion coefficient. For the Laplace operator.

[0046] Optionally, in step S3, the deep learning defect detection model is constructed by fusing the SAM segmentation model and the YOLO object detection model. The formula for calculating the total loss function of the model is:

[0047] (7)

[0048] in, This is the intersection-union ratio (IU) between the predicted bounding box and the ground truth bounding box. is the squared Euclidean distance between the center point of the predicted bounding box and the center point of the true bounding box; c is the diagonal length of the smallest bounding rectangle containing both boxes; These are the weighting coefficients; This is a parameter for aspect ratio consistency.

[0049] Optionally, in S5, the multi-objective optimization algorithm adopts the third-generation non-dominated sorting genetic algorithm NSGA-III, and the formula for calculating the objective function vector is as follows:

[0050] (8)

[0051] Where x is the decision variable vector, i.e. the controllable process parameters throughout the entire process; The objective function is to maximize production efficiency; The objective function is to maximize the finished product qualification rate. The objective function is to minimize the energy consumption per unit product. The objective function is to maximize the sensory score of the finished product; The objective function is to maximize the freeze-thaw stability of the product.

[0052] The constraints are:

[0053] (9)

[0054] in, These are inequality constraints, corresponding to the upper and lower threshold values ​​of each process parameter; These are equality constraints, corresponding to the coupling relationships between process parameters.

[0055] Optionally, step S5 also includes rapid parameter tuning for multi-specification product changeover scenarios, using a Gaussian process Bayesian optimization algorithm. The formula for calculating the posterior distribution of the Gaussian process regression is:

[0056] (10)

[0057] Mean calculation: (11)

[0058] Variance calculation: (12)

[0059] Where X is the observed process parameter matrix; y is the observed quality index vector; Let K be the vector of process parameters to be predicted; K is the kernel function matrix. This is the kernel function vector between the point to be predicted and the observation point; I is the noise variance; I is the identity matrix;

[0060] The desired improvement to the acquisition function calculation formula is:

[0061] (13)

[0062] (14)

[0063] in, This represents the best observed quality index value. To explore and utilize the balance coefficient; The cumulative distribution function of the standard normal distribution; It is the probability density function of the standard normal distribution.

[0064] The beneficial effects of this invention are as follows:

[0065] 1. This invention establishes a standardized, end-to-end quick-frozen egg tart preparation process system, fundamentally different from existing technologies. It breaks through the limitations of existing technologies that only focus on egg tart filling or front-end semi-finished product processes. At the same time, through "pre-control of material properties - parameter linkage and adaptation between processes - closed-loop feedback optimization of finished product quality", it realizes parameter linkage and synergy between processes, completely solving the problems of isolated processes and large batch-to-batch quality differences in existing technologies.

[0066] 2. This invention overcomes the limitations of fixed parameter execution in existing technologies. By introducing industrial process control, operations research optimization, food rheology, computer vision, and Bayesian statistics, it constructs a control system deeply integrated with the preparation process: LSTM temporal networks enable pre-prediction and adaptation of parameters across all processes, solving the industry problem of significant production delays; food rheological constitutive equations coupled with neural networks achieve precise matching of egg tart liquid characteristics with process parameters; the NSGA-III algorithm enables multi-objective global optimization; and the Bayesian optimization algorithm enables rapid tuning of production changeover parameters, shortening the production changeover debugging cycle.

[0067] 3. The synergistic effect of the process and method of the present invention can improve the qualified rate of quick-frozen egg tart products; increase single-shift production capacity and reduce unit product energy consumption; the product can be frozen at -18℃ for up to 12 months and refrigerated for up to 45 days. The food safety compliance, sensory quality and freeze-thaw stability are significantly better than the existing technology, and can be fully adapted to the large-scale and flexible industrial mass production needs of modern food factories. Attached Figure Description

[0068] This specification will be further described by way of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. The same numbers in the drawings denote the same structures or steps.

[0069] Figure 1 This is a schematic diagram of the preparation process of the quick-frozen egg tart according to Embodiment 1 of this application.

[0070] Figure 2 This is a schematic diagram of the food processing precision control method of Embodiment 2 of this application. Detailed Implementation

[0071] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely one preferred embodiment of this application and are only used to explain this application. They do not limit the scope of protection of this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0072] Example 1:

[0073] like Figure 1 As shown, a process for preparing a quick-frozen egg tart includes the following steps:

[0074] a. Targeted modification preparation of egg tart filling: Select egg tart filling raw materials and prepare targeted modified egg tart filling through stepwise mixing, two-stage homogenization and segmented pasteurization. Simultaneously detect the rheological properties of the egg tart filling and transmit the test results to the subsequent process in real time to provide a basis for parameter adaptation of the subsequent process.

[0075] b. Precision shaping of egg tart crusts: The pre-made puff pastry dough is fed into an adjustable precision mold and a vacuum adsorption device to complete the shaping and pressing process. At the same time, the key physical properties of the egg tart crusts, such as thickness, porosity, and edge integrity, are detected, and the test results are transmitted to the subsequent filling process in real time.

[0076] c. Filling and tart crust matching: Based on the rheological properties of the filling and the physical properties of the crust, the filling control parameters are matched and adjusted. The filling is quantitatively and accurately filled by a high-precision filling pump and flow regulating valve. The combined weight, filling weight, and crust weight of a single egg tart are detected simultaneously, and the detection results are transmitted to the subsequent baking process.

[0077] d. Segmented temperature-controlled baking: Based on the characteristic parameters of the egg tart filling, the parameters of the egg tart crust, and the weight parameters passed in the previous step, a multi-segment baking process curve with independent temperature control of the top and bottom heating elements is generated to complete the gradient baking and maturation of the egg tart. The product quality characteristic parameters are detected simultaneously during the baking process, and the baking curve is corrected in real time based on the detection results.

[0078] e. Gradient humidity control and anti-condensation cooling: Based on the core state parameters of the center temperature and moisture content of the egg tart after baking, a gradient cooling curve is generated to control the temperature and relative humidity of the cooling environment in each process to complete the directional cooling of the egg tart. At the same time, the appearance and shape parameters of the product are detected during the cooling process.

[0079] f. Aseptic modified atmosphere packaging: After cooling, the egg tarts are sealed in a Class 10,000 clean environment, and a full range of verifications of packaging compliance and product appearance are performed simultaneously, eliminating substandard products.

[0080] g. Stepped ice crystal control quick-freezing and finished product warehousing: Based on the initial temperature and moisture content parameters of the packaged egg tarts, a stepped temperature control quick-freezing curve is generated. The temperature and wind speed of the quick-freezing environment are controlled in stages to complete the temperature-controlled freezing of the egg tarts. The size and distribution of ice crystals generated during the freezing process are directionally controlled. After batch verification and metal detection, the tarts are put into storage, and finally, a high freeze-thaw stability quick-frozen egg tart product is obtained.

[0081] Through the above steps, this invention establishes a standardized, end-to-end quick-frozen egg tart preparation process system, fundamentally different from existing technologies. It overcomes the limitations of existing technologies that focus only on the egg tart filling or early-stage semi-finished product processes. Furthermore, through "pre-control of material properties - inter-process parameter linkage and adaptation - closed-loop feedback optimization of finished product quality," it achieves parameter linkage and synergy between processes, completely solving the problems of isolated processes and large batch-to-batch quality variations in existing technologies. The weight deviation of mass-produced finished products is ≤±2%, and the sensory quality consistency is ≥98%. The targeted modification of the egg tart filling and the end-to-end process adaptation technology fundamentally improve the freeze-thaw stability of the finished product. This invention achieves precise control of the rheological properties of egg tart filling through two-stage homogenization and segmented pasteurization. At the same time, it establishes a positive correlation matching mechanism between the amount of citrus fiber added and the porosity of the egg tart crust, realizing a deep adaptation of the egg tart filling properties with the physical properties of the crust, filling, baking, and quick-freezing processes. Through a step-by-step ice crystal control quick-freezing process, it forms a synergistic effect with the citrus fiber in the egg tart filling, doubly inhibiting the formation of large-sized ice crystals, so that the average diameter of ice crystals inside the quick-frozen egg tarts is ≤40μm. After being frozen and stored at -18℃ for 12 months, the taste retention rate after thawing and rebaking is ≥96%.

[0082] In step a, the egg tart filling ingredients, by weight, are: 30-35 parts pure milk, 10-15 parts light cream, 15-20 parts egg white liquid, 5-10 parts egg yolk liquid, 10-15 parts white sugar, 5-10 parts whole milk powder, 2-8 parts corn starch, and 2-8 parts citrus fiber. The amount of citrus fiber added is positively correlated with the porosity of the egg tart crust in step b. When the porosity of the egg tart crust is 18%-22%, the amount of citrus fiber added is 2-4 parts; when the porosity of the egg tart crust is 22%-25%, the amount of citrus fiber added is 4-8 parts.

[0083] The targeted modification preparation of the egg tart filling in step a specifically includes:

[0084] a1. Base material premixing: Heat pure milk to 40-60℃, add white sugar, whole milk powder, corn starch and citrus fiber in sequence, stir at medium speed until completely dissolved to obtain base material mixture;

[0085] a2. Egg liquid preparation: Cool the base mixture to 0-10℃, add light cream, egg white liquid, and egg yolk liquid, and stir to obtain the egg tart filling stock solution;

[0086] a3. Two-stage homogenization: The egg tart filling is subjected to two-stage homogenization. The first-stage homogenization pressure is 10-15MPa, the second-stage homogenization pressure is 3-5MPa, and the homogenization temperature is controlled at 4-10℃.

[0087] a4. Segmented pasteurization: The homogenized egg tart mixture is pasteurized in segments. The first pasteurization segment is at a temperature of 60-65℃ for 10-20 minutes; the second pasteurization segment is at a temperature of 70-75℃ for 3-5 minutes.

[0088] a5. Cool and set aside: After sterilization, the egg tart filling is rapidly cooled to 0-4℃ and refrigerated for later use. Simultaneously, the viscosity, solids content, and water content rheological properties of the egg tart filling are tested.

[0089] In step b, the vacuum adsorption pressure is controlled within the range of 0.6-0.8 MPa, the mold pressing pressure is controlled within the range of 2-3 MPa, and the pressing time is controlled within the range of 1-1.5 s; the thickness of the tart crust after molding is controlled within the range of 1.3-2.0 mm, the porosity is controlled within the range of 18%-25%, and the edge integrity deviation is ≤ ±0.3 mm.

[0090] In step c, the filling positioning accuracy is controlled within ±0.5mm, the filling flow rate is controlled within the range of 10-12ml / s, and the single filling time is ≤0.8s; the weight of a single egg tart combination is controlled within the range of 29g-32g, of which the weight of the tart filling is controlled within the range of 15g-17g, and the weight of the tart crust is controlled within the range of 13g-15g.

[0091] The multi-stage baking process curve in step d includes the sequential execution of the puff pastry setting temperature zone, the internal maturation temperature zone, and the color setting temperature zone, with the total baking time controlled at 14-15 minutes. The baking process includes the following temperature zones: Puff pastry setting zone: bottom heat 270-280℃, top heat 175-186℃, baking time 4-5 minutes; internal maturation zone: bottom heat 260-270℃, top heat 166-175℃, baking time 7-8 minutes, with the center temperature of the egg tart reaching above 90℃; color setting zone: bottom heat 255-265℃, top heat 170-180℃, baking time 2-3 minutes; after the entire baking process, the center temperature of the egg tart should be controlled within the range of 95-120℃; the acceptable standard for black spots / foreign objects is: ≤2 black spots with a diameter ≤0.5mm and no connected black spots with a diameter >0.5mm.

[0092] In step e, the gradient cooling and humidity control curve is divided into a rapid cooling section and a constant temperature and humidity control section. In the rapid cooling section, the ambient temperature is 10-15℃, the relative humidity is 45%-55%, and the cooling time is 6-8 minutes, reducing the center temperature of the egg tart from above 95℃ to 40℃. In the constant temperature and humidity control section, the ambient temperature is 0-10℃, the relative humidity is 35%-45%, and the cooling time is 4-6 minutes, reducing the center temperature of the egg tart to the bagging threshold of 0-25℃. After cooling, the indentation of the egg tart filling is controlled within the range of 0-0.8cm, and the difference in length between the long and short sides of the egg tart is <0.3cm.

[0093] In step f, the inner packaging opening diameter is controlled within the range of 5.7-6.2cm, the weight of a single bag is controlled within the range of 25-29g, and the single bag specification is 20 pieces / tray / bag; the modified atmosphere preservation gas is a mixture of 70% N2 + 30% CO2 by volume, and the residual oxygen content inside the packaging is ≤1%.

[0094] In step g, the stepped temperature-controlled quick-freezing curve is divided into a rapid ice crystal formation zone and a constant-temperature freezing zone. Specifically, in the rapid ice crystal formation zone, the quick-freezing environment temperature is -35℃ to -30℃, the wind speed is 3-4 m / s, and the quick-freezing time is 12-15 min, rapidly reducing the egg tart core temperature from 25℃ to -1℃. In the constant-temperature freezing zone, the quick-freezing environment temperature is -25℃ to -18℃, the wind speed is 1-2 m / s, and the quick-freezing time is 20-25 min, reducing the egg tart core temperature to the threshold range of -30℃ to -12℃. After quick-freezing, the average diameter of the ice crystals inside the egg tart is ≤40 μm.

[0095] Example 2:

[0096] like Figure 2 As shown, a synergistic regulation method for quick-frozen egg tarts, based on the quick-frozen egg tart preparation process described in Example 1, includes the following steps:

[0097] S1. Real-time collection of material property parameters of egg tart filling, physical property parameters of egg tart crust, process operation parameters, and environmental status parameters, synchronous acquisition of test data, and construction of a standardized production dataset in all dimensions;

[0098] S2. Using the characteristic parameters of the egg tart filling and the physical properties of the egg tart crust collected in S1 as input, the target process parameters and execution curve are output through a pre-trained parameter matching model, so as to realize the coordinated linkage and pre-adaptation of parameters in each step of the preparation process.

[0099] S3. Real-time acquisition of image data and physicochemical index data of products in production at each process, and completion of product defect identification, classification and quality compliance judgment through deep learning defect detection model, outputting non-conforming product rejection instructions and quality deviation data;

[0100] S4. Taking the quality deviation data output by S3, the real-time deviation between the process operating parameters and the target parameters as input, the process parameter adjustment amount is output through the adaptive control algorithm and sent to the execution equipment of the corresponding process in real time to correct parameter fluctuations and realize closed-loop quality control of the preparation process.

[0101] S5. Based on a standardized production dataset with full dimensions, the system optimizes multiple core performance indicators of the entire production process by using a multi-objective optimization algorithm to find the global optimal combination of process parameters, which is then distributed to the production line for execution. Simultaneously, the system iteratively optimizes the parameter matching model, defect detection model, and adaptive control algorithm based on newly added production data.

[0102] Through the above steps, this invention overcomes the limitations of fixed parameter execution in existing technologies. By introducing industrial process control, operations research optimization, food rheology, computer vision, and Bayesian statistics, it constructs a control system deeply integrated with the preparation process: LSTM temporal networks enable pre-prediction and adaptation of parameters across all processes, solving the industry problem of significant production delays; food rheological constitutive equations coupled with neural networks achieve precise matching of egg tart liquid characteristics with process parameters; the SAM-YOLO fusion model enables high-precision detection of minute defects down to 0.1mm, with an accuracy rate ≥98.5%; the NSGA-III algorithm achieves multi-objective global optimization; and the Bayesian optimization algorithm enables rapid tuning of changeover parameters, shortening the changeover debugging cycle. The synergistic effect of the process and method of this invention can improve the pass rate of quick-frozen egg tart products; increase single-shift production capacity and reduce unit product energy consumption; the product can be frozen at -18℃ for up to 12 months and refrigerated for up to 45 days. The food safety compliance, sensory quality and freeze-thaw stability are significantly better than existing technologies, and can be fully adapted to the large-scale, flexible industrial mass production needs of modern food factories.

[0103] In step S2, the parameter matching model is constructed using a Long Short-Term Memory (LSTM) network, as shown in the following formula:

[0104] Forget Gate Calculation: (1)

[0105] Input gate calculation: (2)

[0106] Cell status update: (3)

[0107] Output gate calculation: (4)

[0108] in, The vector of egg tart filling properties and egg tart crust physical property parameters input at time t; This is the output of the hidden layer at time t-1; , , , These are the weight matrices for each gating control; , , , These are the bias vectors for each gate; It is the sigmoid activation function; For Hadamah accumulation; Let be the cell state at time t; This is the vector of target process parameters for the entire process output by the model at time t.

[0109] Specifically, in step S2, the parameter matching model is also coupled with the food rheological constitutive equation. The rheological properties of the egg tart filling are characterized using the Herschel-Bulkley model, and the specific formula is shown below:

[0110] (5)

[0111] in, Shear stress; K is the yield stress; K is the consistency coefficient; is the shear rate; n is the flow behavior exponent;

[0112] The migration of moisture inside the egg tart during baking is characterized by Fick's second law diffusion equation, as shown in the following formula:

[0113] (6)

[0114] Where C is the internal moisture content of the egg tart; t is time; and D(T) is the temperature-dependent moisture diffusion coefficient. For the Laplace operator.

[0115] In step S3, the deep learning defect detection model is constructed by fusing the SAM segmentation model and the YOLO object detection model. The fusion method is as follows: first, the YOLOv8 backbone network is used to detect defects in the input image, outputting defect candidate boxes. Then, the candidate box regions are input into the SAM model for accurate segmentation, extracting the size and morphological features of the defects, and completing defect classification and compliance judgment. This can achieve high-precision detection of minute defects down to 0.1mm. The formula for calculating the total loss function of the model is:

[0116] (7)

[0117] in, This is the intersection-union ratio (IU) between the predicted bounding box and the ground truth bounding box. is the squared Euclidean distance between the center point of the predicted bounding box and the center point of the true bounding box; c is the diagonal length of the smallest bounding rectangle containing both boxes; These are the weighting coefficients; This is a parameter for aspect ratio consistency.

[0118] In step S4, the adaptive control algorithm employs a fuzzy adaptive PID algorithm. It takes the deviation *e* and the rate of change of the deviation between the actual value and the preset standard value of the process parameter as input, and the adjustment amounts of the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the PID controller as output. The PID parameters are tuned using fuzzy rules to achieve adaptive closed-loop control of the process parameters.

[0119] In step S5, the multi-objective optimization algorithm uses the third-generation non-dominated sorting genetic algorithm NSGA-III, and the formula for calculating the objective function vector is as follows:

[0120] (8)

[0121] Where x is the decision variable vector, i.e. the controllable process parameters throughout the entire process; The objective function is to maximize production efficiency; The objective function is to maximize the finished product qualification rate. The objective function is to minimize the energy consumption per unit product. The objective function is to maximize the sensory score of the finished product; The objective function is to maximize the freeze-thaw stability of the product.

[0122] The constraints are:

[0123] (9)

[0124] in, These are inequality constraints, corresponding to the upper and lower threshold values ​​of each process parameter; These are equality constraints, corresponding to the coupling relationships between process parameters.

[0125] Step S5 also includes rapid parameter tuning for multi-specification product changeover scenarios, using the Gaussian process Bayesian optimization algorithm. The formula for calculating the posterior distribution of the Gaussian process regression is as follows:

[0126] (10)

[0127] Mean calculation: (11)

[0128] Variance calculation: (12)

[0129] Where X is the observed process parameter matrix; y is the observed quality index vector; Let K be the vector of process parameters to be predicted; K is the kernel function matrix. This is the kernel function vector between the point to be predicted and the observation point; I is the noise variance; I is the identity matrix;

[0130] The desired improvement to the acquisition function calculation formula is:

[0131] (13)

[0132] (14)

[0133] in, This represents the best observed quality index value. To explore and utilize the balance coefficient; The cumulative distribution function of the standard normal distribution; It is the probability density function of the standard normal distribution.

[0134] The above-described specific embodiments are preferred embodiments of the preparation process and synergistic control method for quick-frozen egg tarts of this application, and are not intended to limit the specific scope of this application. The scope of this application includes but is not limited to the specific embodiments described herein. All equivalent changes made in accordance with the shape and structure of this application are within the protection scope of this application.

Claims

1. A process for preparing quick-frozen egg tarts, characterized in that, Includes the following steps: a. Targeted modification preparation of egg tart filling: Select egg tart filling raw materials and prepare targeted modified egg tart filling through stepwise mixing, two-stage homogenization and segmented pasteurization. Simultaneously detect the rheological properties of the egg tart filling and transmit the test results to the subsequent process in real time to provide a basis for parameter adaptation of the subsequent process. b. Precision shaping of egg tart crusts: The pre-made puff pastry dough is fed into an adjustable precision mold and a vacuum adsorption device to complete the shaping and pressing process. At the same time, the key physical properties of the egg tart crusts, such as thickness, porosity, and edge integrity, are detected, and the test results are transmitted to the subsequent filling process in real time. c. Filling and tart crust matching: Based on the rheological properties of the filling and the physical properties of the crust, the filling control parameters are matched and adjusted. The filling is quantitatively and accurately filled by a high-precision filling pump and flow regulating valve. The combined weight, filling weight, and crust weight of a single egg tart are detected simultaneously, and the detection results are transmitted to the subsequent baking process. d. Segmented temperature-controlled baking: Based on the characteristic parameters of the egg tart filling, the parameters of the egg tart crust, and the weight parameters passed in the previous step, a multi-segment baking process curve with independent temperature control of the top and bottom heating elements is generated to complete the gradient baking and maturation of the egg tart. The product quality characteristic parameters are detected simultaneously during the baking process, and the baking curve is corrected in real time based on the detection results. e. Gradient humidity control and anti-condensation cooling: Based on the core state parameters of the center temperature and moisture content of the egg tart after baking, a gradient cooling curve is generated to control the temperature and relative humidity of the cooling environment in each process to complete the directional cooling of the egg tart. At the same time, the appearance and shape parameters of the product are detected during the cooling process. f. Aseptic modified atmosphere packaging: After the egg tarts have cooled, they are sealed in a Class 10,000 clean environment. At the same time, the packaging compliance and product appearance are fully verified, and unqualified products are rejected. g. Stepped ice crystal control quick-freezing and finished product warehousing: Based on the initial temperature and moisture content parameters of the packaged egg tarts, a stepped temperature control quick-freezing curve is generated. The temperature and wind speed of the quick-freezing environment are controlled in stages to complete the temperature-controlled freezing of the egg tarts. The size and distribution of ice crystals generated during the freezing process are directionally controlled. After batch verification and metal detection, the tarts are put into storage, and finally, a high freeze-thaw stability quick-frozen egg tart product is obtained.

2. The preparation process of quick-frozen egg tarts according to claim 1, characterized in that, In step a, the egg tart filling ingredients, by weight, are: 30-35 parts pure milk, 10-15 parts light cream, 15-20 parts egg white liquid, 5-10 parts egg yolk liquid, 10-15 parts white sugar, 5-10 parts whole milk powder, 2-8 parts corn starch, and 2-8 parts citrus fiber. The amount of citrus fiber added is positively correlated with the porosity of the egg tart crust in step b. When the porosity of the egg tart crust is 18%-22%, the amount of citrus fiber added is 2-4 parts; when the porosity of the egg tart crust is 22%-25%, the amount of citrus fiber added is 4-8 parts.

3. The preparation process of quick-frozen egg tarts according to claim 1, characterized in that, The targeted modification preparation of the egg tart filling in step a specifically includes: a1. Base material premixing: Heat pure milk to 40-60℃, add white sugar, whole milk powder, corn starch and citrus fiber in sequence, stir at medium speed until completely dissolved to obtain base material mixture; a2. Egg liquid preparation: Cool the base mixture to 0-10℃, add light cream, egg white liquid, and egg yolk liquid, and stir to obtain the egg tart filling stock solution; a3. Two-stage homogenization: The egg tart filling is subjected to two-stage homogenization. The first-stage homogenization pressure is 10-15MPa, the second-stage homogenization pressure is 3-5MPa, and the homogenization temperature is controlled at 4-10℃. a4. Segmented pasteurization: The homogenized egg tart mixture is pasteurized in segments. The first pasteurization segment is at a temperature of 60-65℃ for 10-20 minutes; the second pasteurization segment is at a temperature of 70-75℃ for 3-5 minutes. a5. Cool and set aside: After sterilization, the egg tart filling is rapidly cooled to 0-4℃ and refrigerated for later use. Simultaneously, the viscosity, solids content, and water content rheological properties of the egg tart filling are tested.

4. The preparation process of quick-frozen egg tarts according to claim 1, characterized in that, The multi-stage baking process curve in step d includes sequentially executed puffing and shaping temperature zone, internal maturation temperature zone, and color shaping temperature zone, with the total baking time controlled at 14-15 minutes. The puff pastry shaping temperature zone is: bottom heat temperature of the oven 270-280℃, top heat temperature of the oven 175-186℃, and baking time of 4-5 minutes. The internal ripening temperature zone is: bottom heat temperature of the oven is 260-270℃, top heat temperature of the oven is 166-175℃, baking time is 7-8 minutes, and the center temperature of the egg tart rises to above 90℃. The color-setting temperature zone is as follows: bottom heat temperature of the oven is 255-265℃, top heat temperature of the oven is 170-180℃, and the baking time is 2-3 minutes. After the entire baking process is completed, the center temperature of the egg tart should be controlled within the range of 95-120℃; the acceptable standard for black spots and foreign objects is: the number of black spots with a diameter ≤0.5mm and not connected is ≤2, and there are no black spots with a diameter >0.5mm.

5. A synergistic regulation method for quick-frozen egg tarts, applicable to the preparation process described in any one of claims 1-4, characterized in that, Includes the following steps: S1. Real-time collection of material property parameters of egg tart filling, physical property parameters of egg tart crust, process operation parameters, and environmental status parameters, synchronous acquisition of test data, and construction of a standardized production dataset in all dimensions; S2. Using the characteristic parameters of the egg tart filling and the physical properties of the egg tart crust collected in S1 as input, the target process parameters and execution curve are output through a pre-trained parameter matching model, so as to realize the coordinated linkage and pre-adaptation of parameters in each step of the preparation process. S3. Real-time acquisition of image data and physicochemical index data of products in production at each process, and completion of product defect identification, classification and quality compliance judgment through deep learning defect detection model, outputting non-conforming product rejection instructions and quality deviation data; S4. Taking the quality deviation data output by S3, the real-time deviation between the process operating parameters and the target parameters as input, the process parameter adjustment amount is output through the adaptive control algorithm and sent to the execution equipment of the corresponding process in real time to correct parameter fluctuations and realize closed-loop quality control of the preparation process. S5. Based on a standardized production dataset with full dimensions, the system optimizes multiple core performance indicators of the entire production process by using a multi-objective optimization algorithm to find the global optimal combination of process parameters, which is then distributed to the production line for execution. Simultaneously, the system iteratively optimizes the parameter matching model, defect detection model, and adaptive control algorithm based on newly added production data.

6. The synergistic regulation method for quick-frozen egg tarts according to claim 5, characterized in that, In step S2, the parameter matching model is constructed using a Long Short-Term Memory (LSTM) network, and the specific formula is shown below: Forget Gate Calculation: (1) Input gate calculation: (2) Cell status update: (3) Output gate calculation: (4) in, The vector of egg tart filling properties and egg tart crust physical property parameters input at time t; This is the output of the hidden layer at time t-1; , , , These are the weight matrices for each gating control; , , , These are the bias vectors for each gate; It is the sigmoid activation function; For Hadamah accumulation; Let be the cell state at time t; This is the vector of target process parameters for the entire process output by the model at time t.

7. The synergistic regulation method for quick-frozen egg tarts according to claim 5, characterized in that, In S2, the parameter matching model is also coupled with the food rheological constitutive equation. The rheological properties of the egg tart filling are characterized by the Herschel-Bulkley model, and the specific formula is shown below: （5） in, Shear stress; K is the yield stress; K is the consistency coefficient; is the shear rate; n is the flow behavior exponent; The migration of moisture inside the egg tart during baking is characterized by Fick's second law diffusion equation, as shown in the following formula: （6） Where C is the internal moisture content of the egg tart; t is time; and D(T) is the temperature-dependent moisture diffusion coefficient. For the Laplace operator.

8. The synergistic regulation method for quick-frozen egg tarts according to claim 5, characterized in that, In step S3, the deep learning defect detection model is constructed by fusing the SAM segmentation model and the YOLO object detection model. The formula for calculating the total loss function of the model is as follows: （7） in, This is the intersection-union ratio (IU) between the predicted bounding box and the ground truth bounding box. is the squared Euclidean distance between the center point of the predicted bounding box and the center point of the true bounding box; c is the diagonal length of the smallest bounding rectangle containing both boxes; These are the weighting coefficients; This is a parameter for aspect ratio consistency.

9. The synergistic regulation method for quick-frozen egg tarts according to claim 5, characterized in that, In S5, the multi-objective optimization algorithm adopts the third-generation non-dominated sorting genetic algorithm NSGA-III, and the formula for calculating the objective function vector is as follows: （8） Where x is the decision variable vector, i.e. the controllable process parameters throughout the entire process; The objective function is to maximize production efficiency; The objective function is to maximize the finished product qualification rate. The objective function is to minimize the energy consumption per unit product. The objective function is to maximize the sensory score of the finished product; The objective function is to maximize the freeze-thaw stability of the product. The constraints are: （9） in, These are inequality constraints, corresponding to the upper and lower threshold values ​​of each process parameter; These are equality constraints, corresponding to the coupling relationships between process parameters.

10. The synergistic regulation method for quick-frozen egg tarts according to claim 5, characterized in that, Step S5 also includes rapid parameter tuning for multi-specification product changeover scenarios, using a Gaussian process Bayesian optimization algorithm. The posterior distribution calculation formula for Gaussian process regression is as follows: （10） Mean calculation: (11) Variance calculation: (12) Where X is the observed process parameter matrix; y is the observed quality index vector; Let K be the vector of process parameters to be predicted; K is the kernel function matrix. This is the kernel function vector between the point to be predicted and the observation point; I is the noise variance; I is the identity matrix; The desired improvement to the acquisition function calculation formula is: （13） （14） in, This represents the best observed quality index value. To explore and utilize the balance coefficient; The cumulative distribution function of the standard normal distribution; It is the probability density function of the standard normal distribution.

Images