Environment control method and system for preserving a spicy chicken product
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUIZHOU GUIFUDUO FOOD CO LTD
- Filing Date
- 2026-02-09
- Publication Date
- 2026-06-02
Smart Images
Figure CN122131856A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent storage technology, and in particular to an environmental control method and system for preserving spicy chicken products. Background Technology
[0002] The flavor and quality stability of spicy chicken are greatly affected by storage environment parameters. Currently, the industry's commonly used environmental control solutions are mostly based on static threshold settings or proportional-integral-derivative feedback regulation. When dealing with non-homogeneous foods like spicy chicken, which have complex matrices (high oil content, high capsaicin content, and low water activity), the existing static regulation mechanisms are insufficient to cope with their dynamic oxidative rancidity and texture deterioration. As a result, aroma loss and excessive oil acid value still occur frequently during long-term storage.
[0003] Existing static control methods lack a deep understanding of changes in food quality and cannot establish a dynamic mapping relationship between environmental fluctuations and quality loss, resulting in control strategies often lagging behind the actual deterioration process. For spicy chicken products, simple temperature and humidity control is insufficient to suppress its complex biochemical reactions. Without predictive and synergistic optimization of environmental parameters, it is difficult to meet the requirements of high-end pre-prepared foods for authentic flavor. Therefore, how to guide environmental decisions through high-precision quality prediction models and overcome control bottlenecks under multiple constraints using efficient global search algorithms has become a pressing challenge in this technical field. Summary of the Invention
[0004] In view of the aforementioned existing problems, the present invention is proposed.
[0005] Therefore, this invention provides an environmental control method for the preservation of spicy chicken products, which solves the problems of rigid environmental control strategies, insufficient perception of quality loss, and low efficiency of complex multi-objective optimization in existing spicy chicken product storage technologies.
[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:
[0007] In a first aspect, the present invention provides an environmental control method for the preservation of spicy chicken products, which includes collecting environmental status data in real time by means of sensors deployed in the storage space of spicy chicken products.
[0008] Construct a machine learning quality prediction model, input environmental state data and storage duration, and output the predicted quality loss value of spicy chicken products;
[0009] Based on the predicted quality loss value, a multi-objective fitness function is constructed, and the improved gray wolf optimization algorithm is used to iteratively update the environmental control decision vector to find the environmental control decision vector with the optimal fitness global solution.
[0010] The optimal environmental control decision vector is transformed into control commands to drive the operation of environmental conditioning equipment until the environmental parameters of the storage space reach the target value.
[0011] As a preferred embodiment of the environmental control method for preserving spicy chicken products according to the present invention, the real-time acquisition of environmental status data includes continuously acquiring the environmental status data through a sensor array deployed in the storage space at a preset sampling period; the sensor array includes a temperature sensor, a humidity sensor, and an oxygen concentration sensor.
[0012] The environmental condition data includes at least ambient temperature, relative humidity, and oxygen concentration; when placed in the storage space, the initial physicochemical properties of the spicy chicken products are obtained, including moisture content, water activity, peroxide value, acid value, and salinity.
[0013] As a preferred embodiment of the environmental control method for preserving spicy chicken products according to the present invention, the construction of the machine learning quality prediction model includes: receiving real-time environmental state data, combining it with the current storage duration parameter in a vectorized manner to construct an input feature vector; and inputting the constructed feature vector into the pre-trained machine learning quality prediction model.
[0014] The model employs a hybrid architecture of LSTM (Long Short-Term Memory) network and fully connected layers. When the feature vector is input, the LSTM layer uses the forget gate and input gate to combine the hidden state of the previous time step to capture the dynamic trend of environmental state data changing over time. The output of the LSTM layer flows into the fully connected layer for high-dimensional space mapping, and the activation function maps the low-dimensional environmental state features to the physicochemical property space of spicy chicken. The weight parameters learned during the training phase are used to simulate the nonlinear coupling relationship between environmental factors and the peroxide value, acid value, and water activity in the physicochemical properties of spicy chicken.
[0015] The model outputs a comprehensive quality score for the spicy chicken, which is then converted into a pre-made quality loss value. An ideal quality benchmark value is set, and the pre-made quality loss value is obtained by weighting the deviation and sensitivity coefficient between the ideal quality benchmark value and the comprehensive quality score, reflecting the rate of quality decay.
[0016] As a preferred embodiment of the environmental control method for preserving spicy chicken products according to the present invention, the step of constructing a multi-objective fitness function based on the predicted quality loss value includes: delineating a three-dimensional solution space for environmental state data according to the physical constraints of the spicy chicken preservation process; and generating a hybrid mapping strategy within the three-dimensional solution space. There are 1 initial gray wolf individuals, each of which is instantiated as a set of environmental control decision vectors; the initial gray wolf individuals are uniformly and randomly distributed in the solution space.
[0017] A multi-objective fitness assessment is performed, in which each initial gray wolf individual is sequentially substituted into the machine learning quality prediction model and the pre-made quality loss calculation; through weighted calculation, the comprehensive fitness score corresponding to the initial gray wolf individual is obtained, and the weighted cost of quality loss and operating cost is obtained.
[0018] Regarding the first In each iteration, a gray wolf representing a specific temperature, relative humidity, and oxygen concentration setting is used to calculate a multi-target fitness value; this multi-target fitness value is obtained by a weighted sum of three parts:
[0019] The first part is the predicted quality loss value calculated from the prefabricated quality loss, which is correlated by a first weighting coefficient; the second part is the ratio of compressor energy consumption to maximum energy consumption calculated based on the current target temperature, which is correlated by a second weighting coefficient; the third part is the ratio of nitrogen consumption required to maintain the target oxygen concentration to the maximum consumption, which is correlated by a third weighting coefficient. The first weighting coefficient is set to be greater than the sum of the latter two to establish the principle of quality priority.
[0020] As a preferred embodiment of the environmental control method for preserving spicy chicken products according to the present invention, the step of iteratively updating the environmental control decision vector using the improved gray wolf optimization algorithm includes: classifying the social hierarchy and ranking the gray wolf population based on a fitness score algorithm; the top three individuals with the best performance and lowest cost are respectively marked as... Wolf, wolves and The wolves represent the optimal solution, the second-best solution, and the third-best solution, respectively; the three wolves constitute the core of decision guidance for the current iteration cycle, and the vector position coordinates serve as the direction of population evolution.
[0021] After receiving the position information of the alpha wolf, the remaining gray wolves adjust their control parameters according to the encirclement mechanism formula and converge toward a better solution space region; the updated position vector is determined by the average of the approximation vectors calculated for the three alpha wolves respectively.
[0022] As a preferred embodiment of the environmental control method for preserving spicy chicken products according to the present invention, the environmental control decision vector for finding the optimal global solution includes introducing a Cauchy mutation mechanism during the iteration process, and introducing a random perturbation based on the quality loss gradient during position update to force some gray wolf individuals to jump out of the current gathering area to explore.
[0023] During the iteration process, a Cauchy mutation mechanism that introduces gradients is used when determining... When the wolf's position changes by less than a set threshold in a preset number of iterations, it is considered to have stalled. At this point, [the wolf's position is then...]. The wolf performs a mutation operation to generate a new position; the new position is calculated as follows: Based on the wolf's current position, a random number perturbation of standard Cauchy distribution is superimposed, and a term proportional to the quality loss gradient is subtracted.
[0024] The quality loss gradient is the partial derivative vector of the quality loss with respect to environmental parameters, obtained through backpropagation using a machine learning quality prediction model. The algorithm is forced to break out of the stagnation zone using the negative gradient direction and evolve abruptly in the direction of the fastest quality improvement.
[0025] Check if the updated position exceeds the physical constraint boundary of the device. If it does, force correction to the constraint boundary value. If the termination condition is not met, return to the fitness evaluation step to continue the next iteration.
[0026] Termination conditions include reaching the maximum number of iterations or the variance of the population fitness value falling below a preset convergence threshold, at which point iteration stops; and the population at the top of the population hierarchy... The position vector corresponding to the wolf is the global optimal environmental control decision vector, which includes the target temperature value, target relative humidity value, and target oxygen concentration threshold that minimize the rate of quality loss of spicy chicken products and optimize system energy consumption cost at the current moment.
[0027] As a preferred embodiment of the environmental control method for preserving spicy chicken products according to the present invention, the operation of the driving environmental adjustment equipment includes: receiving the global optimal environmental control decision vector and resolving it into independent setpoints for different actuators; starting multi-channel PID closed-loop control logic to adjust the operating frequency of the variable frequency compressor and the opening of the expansion valve for the temperature setpoint; coordinating the power of the ultrasonic humidifier and the dehumidification wheel to maintain the moisture balance of the microenvironment for the humidity setpoint; and controlling the on / off ratio of the nitrogen replacement valve through a pulse width modulation signal to suppress the oxygen content below the safety line to inhibit oxidative rancidity for the oxygen concentration threshold.
[0028] Secondly, the present invention provides an environmental control system for the preservation of spicy chicken products, including a sensing module comprising a temperature sensor, a humidity sensor, a gas concentration sensor, and a visual monitoring unit for collecting real-time data.
[0029] The computing module has a built-in microprocessor and memory. The microprocessor runs the program to perform machine learning inference and an improved gray wolf optimization algorithm.
[0030] The execution module, including a variable frequency refrigeration compressor, an ultrasonic humidification and dehumidification integrated unit, and a proportional-integral controlled nitrogen valve, is used to respond to the control commands output by the calculation module;
[0031] The interactive module is used to display current environmental parameters, predicted shelf life, and energy consumption status.
[0032] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the environmental control method for preserving spicy chicken products as described in the first aspect of the present invention.
[0033] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the environmental control method for preserving spicy chicken products as described in the first aspect of the present invention.
[0034] The beneficial effects of this invention are as follows: By constructing a machine learning quality prediction model, complex environmental data (temperature, humidity, etc.) are nonlinearly mapped to storage time. This enables the system to quantify the flavor deterioration and oil oxidation degree of spicy chicken during storage, realizing the shift from controlling environmental parameters to controlling quality indicators, thus improving the scientific nature of preservation decisions. By introducing an improved Gray Wolf Optimization Algorithm (IGWO), the problem of environmental control decision vectors easily getting trapped in local optima under multivariable and strongly coupled conditions is solved. The improved algorithm has stronger global search capabilities and convergence accuracy, and can dynamically find the optimal control instructions that satisfy multi-objective fitness functions, ensuring that the storage space is always in an ideal state that is most conducive to suppressing the quality loss of spicy chicken. Attached Figure Description
[0035] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0036] Figure 1 A flowchart of environmental control methods for preserving spicy chicken products.
[0037] Figure 2 A schematic diagram of the overall algorithm flow for environmental control methods in the preservation of spicy chicken products.
[0038] Figure 3 A schematic diagram of computer equipment for environmental control methods in the preservation of spicy chicken products. Detailed Implementation
[0039] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
[0040] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.
[0041] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.
[0042] Reference Figures 1-2 This is one embodiment of the present invention, which provides an environmental control method for preserving spicy chicken products, comprising the following steps:
[0043] S1: Real-time environmental status data is collected by sensors deployed in the storage space of spicy chicken products.
[0044] Furthermore, the real-time acquisition of environmental status data includes continuously acquiring the environmental status data through a sensor array deployed in the storage space at a preset sampling period; the sensor array includes a temperature sensor, a humidity sensor, and an oxygen concentration sensor.
[0045] The environmental condition data includes at least ambient temperature, relative humidity, and oxygen concentration; when placed in the storage space, the initial physicochemical properties of the spicy chicken products are obtained, including moisture content, water activity, peroxide value, acid value, and salinity.
[0046] S2: Build a machine learning quality prediction model, input environmental state data and storage duration, and output the predicted quality loss value of spicy chicken products.
[0047] Furthermore, the construction of the machine learning quality prediction model includes receiving the collected real-time environmental state data, combining it with the current storage duration parameter in a vectorized manner to construct an input feature vector, and inputting the constructed feature vector into the pre-trained machine learning quality prediction model.
[0048] The model adopts a hybrid architecture of LSTM long short-term memory network and fully connected layers. When the feature vector is input, the LSTM layer uses the forget gate and the input gate to combine the hidden state of the previous time step to capture the dynamic trend of environmental state data changing over time.
[0049] A digital proxy model capable of quantifying the impact of environmental parameters on the quality of spicy chicken in real time and with high accuracy is established. The specific algorithm execution process includes three consecutive sub-stages: data preprocessing, model feature extraction and nonlinear mapping, and quantitative output of quality loss values.
[0050] Receive the real-time environmental status data collected by the sensor array in step S1. This data stream contains the current ambient temperature. relative humidity and oxygen concentration To eliminate differences between different physical units (such as temperature in degrees Celsius and humidity percentage), the system compares the above data stream with the current storage duration. The data is combined and imported into the preprocessing module. The preprocessing module uses the ZScore normalization method to transform the raw data into input feature vectors that conform to a standard normal distribution. :
[0051]
[0052] in, and These represent the mean and standard deviation of each environmental parameter in the training set. This process ensures the smoothness of the input data in the numerical space, laying the foundation for stable solution of the subsequent model.
[0053] The output of the LSTM layer flows into the fully connected layer for high-dimensional space mapping. The low-dimensional environmental state features are mapped to the physicochemical property space of the spicy chicken through the activation function. The weight parameters learned during the training phase are used to simulate the nonlinear coupling relationship between environmental factors and peroxide value, acid value and water activity in the physicochemical properties of the spicy chicken.
[0054] The model outputs a comprehensive quality score for the spicy chicken, which is then converted into a pre-made quality loss value. An ideal quality benchmark value is set, and the pre-made quality loss value is obtained by weighting the deviation and sensitivity coefficient between the ideal quality benchmark value and the comprehensive quality score, reflecting the rate of quality decay.
[0055] The constructed feature vector The data is fed into a pre-trained machine learning quality prediction model. This model employs a hybrid architecture of LSTM (Long Short-Term Memory) network and fully connected layers. The data flow within the model is as follows:
[0056] For temporal feature extraction, the feature vector is first passed through an LSTM layer. The model utilizes its internal forget gate and input gate mechanisms, combined with the hidden state from the previous time step, to capture the dynamic trend of environmental parameters changing over time, such as whether the temperature is in a continuous upward trend. For high-dimensional space mapping, the output of the LSTM layer flows into a fully connected layer (DenseLayer). This layer uses an activation function such as ReLU to map the low-dimensional environmental characteristics to a high-dimensional space of the chicken's physicochemical properties.
[0057] In this process, the model uses the weight parameters learned during the training phase to simulate the complex nonlinear coupling relationship between the environmental factors of peroxide value (POV), acid value (AV), and water activity (Aw).
[0058] The quantitative calculation of quality loss value, after deep model deduction, finally generates a normalized value representing the current overall quality score of the spicy chicken. The range is usually 0 to l, where l represents the fresh state just out of the factory.
[0059] To adapt to the subsequent Grey Wolf optimization algorithm system, this quality score is converted into a predicted quality loss value. Calculation logic. The system sets the ideal quality benchmark value. Loss is defined as the deviation between the baseline value and the predicted comprehensive quality score. :
[0060]
[0061] in, The sensitivity coefficient is represented by the ML value, which is ultimately output by the system as a quantitative evaluation index and passed to the subsequent optimization module. This value not only reflects the current quality status under the local environment but also predicts the rate of quality degradation per unit time under the continuous influence of the current environmental parameters.
[0062] S3: Construct a multi-objective fitness function based on the predicted quality loss value, and use the improved gray wolf optimization algorithm to iteratively update the environmental control decision vector to find the environmental control decision vector with the optimal fitness global solution.
[0063] The construction of a multi-objective fitness function based on the predicted quality loss value includes: defining a three-dimensional solution space for environmental state data according to the physical constraints of the spicy chicken preservation process; and generating a hybrid mapping strategy within the three-dimensional solution space. There are 1 initial gray wolf individuals, each of which is instantiated as a set of environmental control decision vectors; the initial gray wolf individuals are uniformly and randomly distributed in the solution space.
[0064] like Figure 2As shown, a multi-objective fitness evaluation is performed, and each initial gray wolf individual is sequentially substituted into the machine learning quality prediction model and the pre-made quality loss calculation; through weighted calculation, the comprehensive fitness score corresponding to the initial gray wolf individual is obtained, and the weighted cost of quality loss and operating cost is obtained.
[0065] The construction and evaluation of the multidimensional fitness function first involves each individual gray wolf. (representing the first) A set in the next iteration (Set value), calculate its multi-objective fitness value To balance food flavor preservation with equipment operating costs, the following fitness model is adopted:
[0066]
[0067] in, The predicted quality loss value output by the surrogate model in S2 is the core of the weight and is directly related to the peroxide value (POV) of the spicy chicken. Indicates based on the current target temperature Calculate compressor energy consumption to prevent an exponential increase in energy consumption in pursuit of minor quality improvements. This indicates the amount of nitrogen required to maintain the target oxygen concentration. This represents the weighting coefficient, which is set in this embodiment. We must adhere to the principle of prioritizing quality.
[0068] Regarding the first In each iteration, a gray wolf representing a specific temperature, relative humidity, and oxygen concentration setting is used to calculate a multi-target fitness value; this multi-target fitness value is obtained by a weighted sum of three parts:
[0069] The first part is the predicted quality loss value calculated from the prefabricated quality loss, which is correlated by a first weighting coefficient; the second part is the ratio of compressor energy consumption to maximum energy consumption calculated based on the current target temperature, which is correlated by a second weighting coefficient; the third part is the ratio of nitrogen consumption required to maintain the target oxygen concentration to the maximum consumption, which is correlated by a third weighting coefficient. The first weighting coefficient is set to be greater than the sum of the latter two to establish the principle of quality priority.
[0070] The iterative update of the environmental control decision vector using the improved gray wolf optimization algorithm includes: classifying the population into social ranks and ranking the gray wolf population based on a fitness score algorithm; the top three individuals with the best performance and lowest cost are respectively marked as... Wolf, wolves and The wolves represent the optimal solution, the second-best solution, and the third-best solution, respectively. The three wolves form the core of decision guidance for the current iteration cycle, and the vector position coordinates serve as the direction of population evolution.
[0071] Decision-making levels are determined. Based on the calculated... The fitness values are sorted in ascending order (lower fitness values are better). The individual with the lowest fitness value is marked as... Wolf (current best strategy in the environment); the second smallest is marked as... Wolf (suboptimal strategy); mark the third smallest as Wolves (alternative strategy); the remaining individuals are marked as Wolf. The physical significance of this step is that the system constantly locks onto three of the most reliable temperature and humidity combinations as a benchmark.
[0072] Coordinated encirclement and position update, Wolves (poor control strategy) based on The system updates its own parameters at different positions in an attempt to find a better solution. The update formula is as follows:
[0073]
[0074]
[0075] Final location update:
[0076]
[0077] in, Indicates the current iteration number (the th iteration). At any given time (time), the current position of the ordinary wolf (i.e., the environmental control decision vector to be optimized). In this project, it represents a specific set of optional control parameters (such as temperature, humidity, wind speed, etc.). This indicates that after this iteration update, the decision vector will be updated in the next iteration (the...). (times) predicted location. These represent the positions of the wolves ranked first, second, and third in fitness within the current population. Technically, they represent the three "near-optimal" environmental control schemes currently identified by the system. Wolves refer to the ordinary individuals in a population other than the top three; they are responsible for... The search is conducted under the guidance of [the relevant authority / guideline].
[0078] Distance control and disturbance parameters These represent the encirclement distances between the ordinary wolf and the three alpha wolves. This value determines the step size by which the current search point moves closer to the optimal solution. Oscillator: Usually a random number. Its function is to simulate the random movement of prey in nature, increase the randomness of the algorithm, and prevent the control system from getting stuck in local optima (i.e., prevent the system from rigidly hovering around a certain fixed parameter). Denotes the convergence factor, when At that time, the wolf pack attacks its prey (local precision optimization); when At this time, the wolf pack disperses to search for new prey (global search).
[0079] Intermediate candidate solutions and composite decision, They represent the recipients respectively After guidance, each of the ordinary wolves generates a predicted candidate position. The denominator 3 represents the equally weighted arithmetic average of the optimization suggestions provided by the three alpha wolves, reflecting the comprehensive coordination logic in multi-objective decision-making.
[0080] This process simulates the system making trade-offs and fine-tuning among three optimal solutions. For example, The solution has an extremely low temperature but insufficient dehumidification. The initial humidity was perfect, but the temperature was slightly high; the new location after comprehensive consideration... It can often combine the advantages of both.
[0081] After receiving the position information of the alpha wolf, the remaining gray wolves adjust their control parameters according to the encirclement mechanism formula and converge toward a better solution space region; the updated position vector is determined by the average of the approximation vectors calculated for the three alpha wolves respectively.
[0082] The environmental control decision vector for finding the optimal global solution includes introducing a Cauchy mutation mechanism during the iteration process, and introducing a random perturbation based on the quality loss gradient during position updates to force some gray wolf individuals to jump out of the current gathering area to explore.
[0083] To address the issue of traditional gray wolf algorithms easily getting trapped in local optima when handling nonlinear biochemical decay processes like those in spicy chicken (e.g., simply lowering the temperature while ignoring humidity leads to a tougher texture), this step introduces a gradient-guided mutation mechanism. When the algorithm determines... The wolf's position is in a continuous In the next iteration, the change magnitude is less than the threshold. When (i.e., when stagnation occurs), for Wolves perform mutation operations to generate The formula is expressed as:
[0084]
[0085] in, It represents a standard Cauchy distribution random number, and its long-tailed characteristic is used to increase perturbation to simulate sudden environmental changes (such as opening and closing doors).
[0086] The gradient vector representing quality loss with respect to environmental parameters. .
[0087] This was obtained through backpropagation from the neural network model of S2. It indicates the direction of the fastest quality decline. The negative sign in the formula... This means that the algorithm is forced to mutate in the direction of the fastest quality improvement, rather than blindly jumping around randomly.
[0088] Check the updated Does it exceed the physical capabilities of the equipment (e.g.) or If the value exceeds the limit, it will be forcibly pulled back to the boundary. If the termination condition is not met, it will return to continue the next iteration.
[0089] During the iteration process, a Cauchy mutation mechanism that introduces gradients is used when determining... When the wolf's position changes by less than a set threshold in a preset number of iterations, it is considered to have stalled. At this point, [the wolf's position is then...]. The wolf performs a mutation operation to generate a new position; the new position is calculated as follows: Based on the wolf's current position, a random number perturbation of a standard Cauchy distribution is superimposed, and a term proportional to the quality loss gradient is subtracted.
[0090] The quality loss gradient is the partial derivative vector of quality loss with respect to environmental parameters obtained through backpropagation using a machine learning quality prediction model. The algorithm is forced to jump out of the stagnation zone using the negative gradient direction and evolve abruptly towards the gradient direction with the fastest quality improvement.
[0091] Check if the updated position exceeds the physical constraint boundary of the device. If it does, force correction to the constraint boundary value. If the termination condition is not met, return to the fitness evaluation step to continue the next iteration.
[0092] Furthermore, after the aforementioned gradient-guided mutation and multi-round iterative evolution process, the algorithm stops iterating when it meets the preset termination conditions. These termination conditions include reaching the maximum number of iterations or the variance of the population fitness value falling below a preset convergence threshold. At this point, the position vector corresponding to the α wolf at the top of the population hierarchy is locked as the globally optimal environmental control decision vector. This vector precisely contains a set of target temperature values, target relative humidity values, and target oxygen concentration thresholds that, at the current moment, minimize the rate of quality loss in the spicy chicken product and optimize system energy consumption costs.
[0093] S4: Transform the optimal environmental control decision vector into control commands to drive the environmental conditioning equipment to operate until the environmental parameters of the storage space reach the target value.
[0094] Furthermore, the system implements the instruction conversion and closed-loop control for the specific execution of the optimal decision. The control system first receives the global optimal environmental control decision vector output from step S3, and then resolves it into independent setpoints for different actuators.
[0095] Subsequently, the system activates multi-channel PID closed-loop control logic. For the temperature setpoint, it adjusts the operating frequency of the variable frequency compressor and the opening of the expansion valve to quickly eliminate heat load deviation and prevent abnormal crystallization of oil due to temperature fluctuations. For the humidity setpoint, it coordinates the power of the ultrasonic humidifier and dehumidification impeller to maintain the moisture balance of the microenvironment and preserve the crispness of the spicy chicken products. For the oxygen concentration threshold, it precisely controls the on / off ratio of the nitrogen replacement valve through pulse width modulation (PWM) signals to suppress the oxygen content below the safety line and inhibit oxidative rancidity.
[0096] It should be noted that during the execution process, the system continuously monitors sensor feedback and dynamically corrects the output intensity of the actuator until the actual environmental parameters in the storage space stably converge to the target value, thereby completing the closed loop of a single environmental control cycle.
[0097] This embodiment also provides an environmental control system for preserving spicy chicken products, including:
[0098] The sensing module includes a temperature sensor, a humidity sensor, a gas concentration sensor, and a visual monitoring unit, which are used to collect real-time data.
[0099] The computing module has a built-in microprocessor and memory. The microprocessor runs the program to perform machine learning inference and an improved gray wolf optimization algorithm.
[0100] The execution module, including a variable frequency refrigeration compressor, an ultrasonic humidifier and dehumidifier, and a proportional-integral controlled nitrogen valve, is used to respond to control commands output by the calculation module.
[0101] The interactive module is used to display current environmental parameters, predicted shelf life, and energy consumption status.
[0102] This embodiment also provides a computer device, such as... Figure 3 As shown, the environmental control method for preserving spicy chicken products includes: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the environmental control method for preserving spicy chicken products as proposed in the above embodiments.
[0103] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.
[0104] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the environmental control method for preserving spicy chicken products as proposed in the above embodiment. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.
[0105] In summary, this invention achieves the following: Real-time collection of environmental status data is achieved through sensors deployed within the storage space of spicy chicken products; a machine learning quality prediction model is constructed, taking the environmental status data and storage duration as input, and outputting the predicted quality loss value of the spicy chicken products; a multi-objective fitness function is constructed based on the predicted quality loss value, and an improved gray wolf optimization algorithm is used to iteratively update the environmental control decision vector to find the environmental control decision vector with the optimal global fitness solution; the optimal environmental control decision vector is then converted into control commands to drive the environmental regulation equipment until the environmental parameters of the storage space reach the target value.
[0106] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.
Claims
1. A method for environmental control in the preservation of spicy chicken products, characterized in that: include, Environmental status data is collected in real time by sensors deployed in the storage space of spicy chicken products; Construct a machine learning quality prediction model, input environmental state data and storage duration, and output the predicted quality loss value of spicy chicken products; Based on the predicted quality loss value, a multi-objective fitness function is constructed, and the improved gray wolf optimization algorithm is used to iteratively update the environmental control decision vector to find the environmental control decision vector with the optimal fitness global solution. The optimal environmental control decision vector is transformed into control commands to drive the environmental conditioning equipment to operate until the environmental parameters of the storage space reach the target value.
2. The environmental control method for preserving spicy chicken products as described in claim 1, characterized in that: The real-time acquisition of environmental status data includes continuously acquiring the environmental status data through a sensor array deployed in the storage space at a preset sampling period; the sensor array includes a temperature sensor, a humidity sensor, and an oxygen concentration sensor. The environmental condition data includes at least ambient temperature, relative humidity, and oxygen concentration; when placed in the storage space, the initial physicochemical properties of the spicy chicken products are obtained, including moisture content, water activity, peroxide value, acid value, and salinity.
3. The environmental control method for preserving spicy chicken products as described in claim 2, characterized in that: The construction of the machine learning quality prediction model includes receiving real-time environmental state data, combining it with the current storage duration parameter in a vectorized manner to construct an input feature vector, and inputting the constructed feature vector into the pre-trained machine learning quality prediction model. The model employs a hybrid architecture of LSTM (Long Short-Term Memory) network and fully connected layers. When the feature vector is input, the LSTM layer uses the forget gate and input gate to combine the hidden state from the previous time step to capture the dynamic trend of environmental state data changing over time. The output of the LSTM layer flows into the fully connected layer for high-dimensional space mapping, and the low-dimensional environmental state features are mapped to the physicochemical property space of spicy chicken through the activation function. The weight parameters learned during the training phase are used to simulate the nonlinear coupling relationship between environmental factors and the peroxide value, acid value, and water activity in the physicochemical properties of spicy chicken. After deep modeling, the current overall quality score of the spicy chicken is output, and the overall quality score is converted into a pre-made quality loss value. An ideal quality benchmark value is set, and the pre-made quality loss value is obtained by weighting the deviation and sensitivity coefficient of the ideal quality benchmark value and the overall quality score, which reflects the quality decay rate.
4. The environmental control method for preserving spicy chicken products as described in claim 3, characterized in that: The construction of a multi-objective fitness function based on the predicted quality loss value includes: defining a three-dimensional solution space for environmental state data according to the physical constraints of the spicy chicken preservation process; and generating a hybrid mapping strategy within the three-dimensional solution space. There are 1 initial gray wolf individuals, each of which is instantiated as a set of environmental control decision vectors; the initial gray wolf individuals are uniformly and randomly distributed in the solution space. A multi-objective fitness assessment is performed, in which each initial gray wolf individual is sequentially substituted into the machine learning quality prediction model and the pre-made quality loss calculation; through weighted calculation, the comprehensive fitness score corresponding to the initial gray wolf individual is obtained, and the weighted cost of quality loss and operating cost is obtained. Regarding the first In each iteration, a gray wolf representing a specific temperature, relative humidity, and oxygen concentration setting is used to calculate a multi-target fitness value; this multi-target fitness value is obtained by a weighted sum of three parts: The first part is the predicted quality loss value calculated from the prefabricated quality loss, which is correlated by a first weighting coefficient; the second part is the ratio of compressor energy consumption to maximum energy consumption calculated based on the current target temperature, which is correlated by a second weighting coefficient; the third part is the ratio of nitrogen consumption required to maintain the target oxygen concentration to the maximum consumption, which is correlated by a third weighting coefficient. The first weighting coefficient is set to be greater than the sum of the latter two to establish the principle of quality priority.
5. The environmental control method for preserving spicy chicken products as described in claim 4, characterized in that: The iterative update of the environmental control decision vector using the improved gray wolf optimization algorithm includes classifying the social hierarchy and ranking the gray wolf population based on the fitness score algorithm. The three individuals with the best performance and the lowest cost were labeled as follows: Wolf, wolves and Wolves represent the optimal solution, the second-best solution, and the third-best solution, respectively. The three alpha wolves form the core of decision-making guidance in the current iteration cycle, with the vector position coordinates serving as the direction of population evolution; After receiving the position information of the alpha wolf, the remaining gray wolves adjust their control parameters according to the encirclement mechanism formula and converge toward a better solution space region; the updated position vector is determined by the average of the approximation vectors calculated for the three alpha wolves respectively.
6. The environmental control method for preserving spicy chicken products as described in claim 5, characterized in that: The environmental control decision vector for finding the optimal global solution includes introducing a Cauchy mutation mechanism during the iteration process and introducing a random perturbation based on the quality loss gradient during position updates, forcing some gray wolf individuals to jump out of the current gathering area to explore. During the iteration process, a Cauchy mutation mechanism that introduces gradients is used when determining... When the wolf's position changes by less than a set threshold in a preset number of iterations, it is considered to have stalled. At this point, [the wolf's position is then...]. The wolf performs a mutation operation to generate a new position; the new position is calculated as follows: Based on the wolf's current position, a random number perturbation of standard Cauchy distribution is superimposed, and a term proportional to the quality loss gradient is subtracted. The quality loss gradient is the partial derivative vector of quality loss with respect to environmental parameters obtained through backpropagation using a machine learning quality prediction model; the algorithm is forced to jump out of the stagnation zone using the negative gradient direction and undergo a mutation evolution towards the gradient direction with the fastest quality improvement. Check if the updated position exceeds the physical constraint boundary of the device. If it does, force correction to the constraint boundary value. If the termination condition is not met, return to the fitness evaluation step to continue the next iteration. Termination conditions include reaching the maximum number of iterations or the variance of the population fitness value falling below a preset convergence threshold, at which point iteration stops; and the population at the top of the population hierarchy... The position vector corresponding to the wolf is the global optimal environmental control decision vector, which includes the target temperature value, target relative humidity value, and target oxygen concentration threshold that minimize the rate of quality loss of spicy chicken products and optimize system energy consumption cost at the current moment.
7. The environmental control method for preserving spicy chicken products as described in claim 6, characterized in that: The operation of the driving environment regulation equipment includes: receiving the globally optimal environmental control decision vector and resolving it into independent setpoints for different actuators; starting multi-channel PID closed-loop control logic to adjust the operating frequency of the variable frequency compressor and the opening of the expansion valve for the temperature setpoint; coordinating the power control of the ultrasonic humidifier and the dehumidification impeller for the humidity setpoint to maintain the moisture balance of the microenvironment; and controlling the on / off ratio of the nitrogen replacement valve through a pulse width modulation signal to suppress the oxygen content below the safety line to inhibit oxidative rancidity for the oxygen concentration threshold.
8. An environmental control system for preserving spicy chicken products, based on the environmental control method for preserving spicy chicken products according to any one of claims 1 to 7, characterized in that: It includes a sensing module, comprising a temperature sensor, a humidity sensor, a gas concentration sensor, and a visual monitoring unit, for collecting real-time data; The computing module has a built-in microprocessor and memory. The microprocessor runs the program to perform machine learning inference and an improved gray wolf optimization algorithm. The execution module, including a variable frequency refrigeration compressor, an ultrasonic humidification and dehumidification integrated unit, and a proportional-integral controlled nitrogen valve, is used to respond to the control commands output by the calculation module; The interactive module is used to display current environmental parameters, predicted shelf life, and energy consumption status.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the environmental control method for preserving spicy chicken products according to any one of claims 1 to 7.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the environmental control method for preserving spicy chicken products according to any one of claims 1 to 7.