Energy-saving and heat-preserving self-adaptive dynamic control method and system for roasting equipment

Abstract

The application provides a roasting equipment energy-saving heat preservation self-adaptive dynamic control method and system, relates to the technical field of data processing, and the method comprises the following steps: step 1, real-time collection of food material surface temperature distribution, thickness gradient and environmental temperature and humidity data, extraction of three-dimensional heat diffusion shape characteristics, and generation of an initial shape characteristic vector; step 2, based on the real-time deviation value of the set target temperature and the measured food material core temperature, the shape characteristic vector is combined to calculate the pseudo partial derivative estimate value, and the periodic intermittent power parameters, including the duty cycle and the period, are generated according to the pseudo partial derivative estimate value. Through multi-source data acquisition analysis, dynamic calculation and optimization of parameters, intelligent matching of the knowledge base, precise temperature control and efficient energy saving are realized, and the service life of the equipment is prolonged while the cooking quality of the food material is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a frying and baking equipment energy-saving and heat preservation adaptive dynamic control method and system. BACKGROUND

[0002] In the prior art, the dynamic perception of the equipment to the thermal diffusion characteristics of the food material and the strategy adjustment capability need to be improved. The heating quality and energy utilization problems caused by temperature control deviation in actual application need to be optimized.

[0003] Traditional control methods mostly use fixed power or simple PID adjustment, and the correlation mechanism between the three-dimensional thermal diffusion characteristics of the food material (such as thickness gradient and surface curvature) and the heating parameters needs to be improved. When processing different forms of food materials, local uneven heating may occur, which needs to be improved through parameter adjustment, affecting the processing efficiency. The heat preservation technology needs to be improved in terms of combining the dynamic adjustment strategy of the food material form and the environmental temperature and humidity. The compensation capability for heat loss when the environment changes is insufficient, which may cause temperature fluctuations during the heat preservation stage. The global uniform heating mode of the traditional equipment is not perfect in terms of differentiated regulation mechanism for the thermal demand of different regions of the food material. Complex geometric form food materials are prone to local overcooking or undercooking, which needs manual intervention and adjustment, affecting the processing consistency.

[0004] Some studies optimize heating control through sensor data or logical algorithms, but there are still limitations in establishing the direct mapping relationship between the thermal diffusion characteristics of the food material and the heating parameters. When processing food materials of different thicknesses or curvatures, the optimization effect of temperature control accuracy and energy utilization rate needs to be improved. SUMMARY

[0005] The technical problem to be solved by the present application is to provide a frying and baking equipment energy-saving and heat preservation adaptive dynamic control method and system to realize accurate temperature control and energy-saving and heat preservation during the frying and baking process.

[0006] To solve the above technical problems, the technical solutions of the present application are as follows:

[0007] In a first aspect, a frying and baking equipment energy-saving and heat preservation adaptive dynamic control method is provided, which comprises the following steps:

[0008] Step 1: Real-time acquisition of food material surface temperature distribution, thickness gradient and environmental temperature and humidity data, extraction of three-dimensional thermal diffusion form characteristics, and generation of an initial form characteristic vector;

[0009] Step 2: Based on the real-time deviation value of the set target temperature and the measured core temperature of the food material, the form characteristic vector is combined to calculate the pseudo partial derivative estimate value, and the periodic intermittent power parameters including the duty cycle and the period are generated according to the pseudo partial derivative estimate value;

[0010] Step 3, input the morphological feature vector into the dynamic compensation function, output the morphological correction coefficient, and dynamically correct the pseudo partial derivative estimate value based on the periodic intermittent power parameter, to generate an adaptive partial derivative;

[0011] Step 4, input the adaptive partial derivative into the dynamic power distributor, combine the intermittent power constraint condition and the temperature error tolerance, optimize the step weight and the correction strength through the directional parameter search algorithm, and generate a spatial power distribution parameter set;

[0012] Step 5, store the spatial power distribution parameter set in association with the corresponding morphological feature vector and the periodic power parameter in the heat diffusion knowledge base, and automatically match the spatial power distribution parameter set when a new food material is detected.

[0013] Further, real-time acquisition of food material surface temperature distribution, thickness gradient and environmental temperature and humidity data, extraction of three-dimensional heat diffusion morphological features, generation of initial morphological feature vector, including:

[0014] Real-time acquisition of food material surface temperature distribution data by infrared thermal imaging array, acquisition of food material thickness gradient data by laser thickness sensor, and acquisition of environmental temperature and humidity data by temperature and humidity sensor;

[0015] Based on the surface temperature distribution, thickness gradient and environmental temperature and humidity data, the three-dimensional heat diffusion morphological features are extracted;

[0016] Feature fusion of three-dimensional heat diffusion morphological features to generate initial morphological feature vector.

[0017] Further, the three-dimensional heat diffusion morphological features include:

[0018] Geometric shape factor, i.e. heat diffusion rate parameter of food material surface curvature distribution;

[0019] Thickness distribution factor, i.e. heat conduction efficiency parameter of thickness difference of different regions of food material;

[0020] Surface curvature factor, i.e. convective heat transfer efficiency parameter of concave and convex surface of food material.

[0021] Further, based on the real-time deviation value of the set target temperature and the measured food material core temperature, combined with the morphological feature vector, the pseudo partial derivative estimate value is calculated, and the periodic intermittent power parameter including duty cycle and period is generated according to the pseudo partial derivative estimate value, including:

[0022] Obtain the initial morphological feature vector, and calculate the deviation value of the set target temperature and the measured food material core temperature in real time;

[0023] The bias value and the initial morphology feature vector are input into a pseudo partial derivative estimator, the thickness distribution factor in the initial morphology feature vector is processed, a thickness weight coefficient is generated, and the thickness weight coefficient is fused with historical bias data, a pseudo partial derivative estimation value is output by rolling optimization through a dynamic weighted least squares algorithm, and the pseudo partial derivative estimation value represents the influence intensity of unit power change on the core temperature.

[0024] The pseudo partial derivative estimation value is input into an intermittent power generator, a basic duty ratio is calculated, wherein the basic duty ratio is inversely proportional to the pseudo partial derivative estimation value; a geometric shape factor in the initial morphology feature vector is obtained, the periodicity value is curvature compensated according to the geometric shape factor, and a periodic intermittent power parameter is output, including the duty ratio optimized by the pseudo partial derivative and the periodicity value compensated by the geometric shape factor.

[0025] Further, the morphology feature vector is input into a dynamic compensation function, a morphology correction coefficient is output, and the pseudo partial derivative estimation value is dynamically corrected based on the periodic intermittent power parameter to generate an adaptive partial derivative, including:

[0026] The morphology feature vector is input into the dynamic compensation function, the geometric shape factor and the surface curvature factor are subjected to Gaussian kernel nonlinear mapping to generate a curvature-diffusion coupling coefficient;

[0027] Based on the interaction between the coupling coefficient and the thickness distribution factor, the morphology correction coefficient is output to quantify the compensation demand intensity of the food material surface on heat diffusion;

[0028] The periodic intermittent power parameter is obtained, the power density characteristics in the duty ratio and the switching frequency characteristics in the periodicity value are analyzed; the morphology correction coefficient and the power switching frequency characteristics are fused to construct a dynamic weight matrix, and the morphology correction coefficient is taken as a main correction factor to correct the pseudo partial derivative estimation value with variable gain, and an adaptive partial derivative is output to represent the actual influence intensity of unit power on the core temperature after compensation.

[0029] Further, the adaptive partial derivative is input into a dynamic power distributor, the step weight and the correction intensity are optimized through a directional parameter search algorithm in combination with the intermittent power constraint condition and the temperature error tolerance to generate a spatial power distribution parameter set, including:

[0030] The adaptive partial derivative is input into the dynamic power distributor, and the adaptive partial derivative is taken as a gradient descent direction reference to initialize a parameter search path;

[0031] In combination with the intermittent power constraint condition, including the maximum power threshold limit and the minimum intermittent duration constraint, and in combination with the temperature error tolerance, including the core temperature allowable fluctuation range and the surface temperature uniformity threshold, the directional parameter search algorithm is used for axial detection along the adaptive partial derivative direction, wherein the step weight is mapped as a power adjustment step size, and the correction intensity is mapped as a thermal response inertia compensation coefficient.

[0032] In the process of axial detection, a Pareto final solution set is generated within the constraint boundary, including the partition radiation power intensity value, the partition heating duration sequence and the cross-zone power switching timing scheme.

[0033] Further, the spatial power distribution parameter set is associated with the corresponding morphological feature vector and the periodic power parameter and stored in the heat diffusion knowledge base, and when a new food material is detected, the spatial power distribution parameter set is automatically matched, including:

[0034] The spatial power distribution parameter set is associated with the morphological feature vector to establish a primary association mapping, and the spatial power distribution parameter set is associated with the periodic power parameter to establish a secondary adjustment mapping.

[0035] The real-time geometric shape factor in the morphological feature vector is taken as a heat field distribution benchmark key, the real-time duty cycle dynamic range in the periodic power parameter is taken as a power regulation benchmark key, and the spatial power distribution parameter set is associated and stored, so as to construct a heat diffusion knowledge base storage structure.

[0036] When a new food material is detected, the real-time morphological feature vector and the periodic power parameter of the new food material are extracted, the heat diffusion characteristic fitness with the heat field distribution benchmark key is calculated based on the real-time geometric shape factor, and the dynamic compatibility with the power regulation benchmark key is verified based on the real-time duty cycle dynamic range; the spatial power distribution parameter set that passes the double verification is retrieved, and the partition radiation power intensity value, the partition heating duration sequence and the cross-zone power switching timing scheme are loaded.

[0037] In a second aspect, the energy-saving and heat-preserving self-adaptive dynamic control system of the roasting equipment comprises:

[0038] The feature extraction module is configured to collect surface temperature distribution, thickness gradient and environment temperature and humidity data of the food material in real time, extract three-dimensional heat diffusion morphological features, and generate an initial morphological feature vector.

[0039] The power parameter calculation module is configured to calculate a pseudo partial derivative estimate value based on a real-time deviation value of a set target temperature and a measured food material core temperature, in combination with the morphological feature vector, and generate a periodic intermittent power parameter.

[0040] The dynamic correction module is configured to input the morphological feature vector into a dynamic compensation function to output a morphological correction coefficient, and dynamically correct the pseudo partial derivative estimate value based on the periodic intermittent power parameter to generate an adaptive partial derivative.

[0041] The parameter optimization module is configured to input the adaptive partial derivative into a dynamic power distributor, in combination with an intermittent power constraint condition and a temperature error tolerance, and optimize a step weight and a correction strength through a directional parameter search algorithm to generate a spatial power distribution parameter set.

[0042] A matching module is configured to store the spatial power distribution parameter set in the heat diffusion knowledge base in association with the corresponding morphological feature vector and the periodic power parameter, and automatically match the spatial power distribution parameter set when a new food material is detected.

[0043] In a third aspect, a computing device includes:

[0044] one or more processors;

[0045] a memory device storing one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method.

[0046] In a fourth aspect, a computer-readable storage medium stores a program, which, when executed by a processor, implements the method.

[0047] The above-mentioned scheme of the present application at least includes the following beneficial effects:

[0048] By collecting data such as surface temperature and thickness gradient of food materials in real time, combining dynamic compensation and correction mechanism, accurately capturing the change of food material heat conduction characteristics, the core temperature deviation can be controlled within a very small range, avoiding local overcooking or undercooking phenomenon, ensuring uniform heating of food materials, and improving cooking quality and taste. Using the pseudo partial derivative estimate value to generate periodic intermittent power parameters, and combining the directional search algorithm to optimize power distribution, avoiding energy waste caused by continuous high-power heating, reducing operating costs compared with traditional frying and roasting equipment. No matter how the shape and thickness of the food material change, or the environment humidity fluctuates, this method can dynamically adjust the power distribution strategy according to real-time data, quickly adapt to thin fish to thick meat, from humid and rainy weather to dry environment, and ensure stable cooking effect.

[0049] In the power distribution process, strictly follow the intermittent power constraint condition, such as limiting the maximum power threshold and setting the minimum intermittent duration, to avoid long-time high-load operation or frequent start-stop of the equipment, reduce equipment wear and tear, reduce failure rate, prolong the overall service life of the equipment, and reduce maintenance cost. The heat diffusion knowledge base stores the correspondence between different food material morphologies and power parameters, and as the data accumulates, it can automatically match the optimal power distribution scheme for similar food materials. After a new food material is put into the equipment, the parameters can be quickly retrieved, reducing the preheating and debugging time, improving the cooking efficiency, and gradually realizing the intelligent upgrading of the equipment. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 is a flowchart of the energy-saving and heat-preserving self-adaptive dynamic control method of the frying and roasting equipment provided by the embodiment of the present application.

[0051] Figure 2The energy-saving heat preservation self-adaptive dynamic control system schematic diagram of the baking equipment provided by the embodiment of the present application. DETAILED DESCRIPTION

[0052] Exemplary embodiments of the present disclosure will be described in greater detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it is understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that the present disclosure can be more thoroughly understood, and the scope of the present disclosure can be accurately conveyed to those skilled in the art.

[0053] As Figure 1 shown, the embodiment of the present application proposes an energy-saving heat preservation self-adaptive dynamic control method for baking equipment, which comprises the following steps:

[0054] Step 1, real-time acquisition of food material surface temperature distribution, thickness gradient and environmental temperature and humidity data, extraction of three-dimensional heat diffusion shape feature, generation of initial shape feature vector;

[0055] Step 2, based on the real-time deviation value of the set target temperature and the measured food material core temperature, combining the shape feature vector, calculating the pseudo partial derivative estimate value, and generating the periodic intermittent power parameters including duty cycle and period according to the pseudo partial derivative estimate value;

[0056] Step 3, input the shape feature vector into the dynamic compensation function, output the shape correction coefficient, and dynamically correct the pseudo partial derivative estimate value based on the periodic intermittent power parameters to generate the adaptive partial derivative;

[0057] Step 4, input the adaptive partial derivative into the dynamic power distributor, combine the intermittent power constraint condition and the temperature error tolerance, optimize the step weight and the correction strength through the directional parameter search algorithm to generate the spatial power distribution parameter set;

[0058] Step 5, store the spatial power distribution parameter set, the corresponding shape feature vector and the periodic power parameters in the heat diffusion knowledge base, and automatically match the spatial power distribution parameter set when a new food material is detected.

[0059] In the embodiment of the present application, by collecting temperature distribution, thickness gradient and environmental parameters in real time, a three-dimensional heat diffusion feature vector is constructed, the overall modeling of food material morphology and heat conduction characteristics is realized, the problem of one-sided heat field perception in traditional control is solved, and data support is provided for dynamic adjustment. Based on the temperature deviation and the morphology feature, the pseudo partial derivative is calculated, and the periodic intermittent power parameters (duty ratio, period) are generated. The heating power can be dynamically adjusted according to the real-time state of the food material, the energy waste under the constant power mode is reduced, and the energy saving efficiency is improved. The morphology correction coefficient is generated by using the dynamic compensation function, and the partial derivative is corrected by combining the intermittent power parameters, so as to effectively compensate the influence of the morphology differences such as the thickness and curvature of the food material on the heat conduction, reduce the core temperature control deviation, and improve the heating uniformity.

[0060] The space power distribution is optimized by the directional parameter search algorithm, the partition differentiated heating is realized according to the heat demand of different regions of the food material, the local overcooked or undercooked caused by traditional uniform heating is avoided, and the consistency and qualification rate of food material processing are improved. The power parameters and the morphology features are associated and stored in the knowledge base to support automatic parameter matching of new food materials, reduce the cost of manual debugging, enhance the adaptability of the equipment to different food materials, and promote the upgrading of intelligent control.

[0061] In a preferred embodiment of the present application, the above step 1, real-time collection of food material surface temperature distribution, thickness gradient and environmental temperature and humidity data, extraction of three-dimensional heat diffusion morphology features, and generation of initial morphology feature vector, can include:

[0062] Step 100, collecting food material surface temperature distribution data in real time by an infrared thermal imaging array, obtaining food material thickness gradient data by a laser thickness sensor, and collecting environmental temperature and humidity data by a temperature and humidity sensor;

[0063] Step 101, based on the surface temperature distribution, thickness gradient and environmental temperature and humidity data, extracting three-dimensional heat diffusion morphology features, specifically including:

[0064] The geometric shape factor is the heat diffusion rate parameter of the curvature distribution of the food material surface;

[0065] The thickness distribution factor is the heat conduction efficiency parameter of the thickness difference of different regions of the food material;

[0066] The surface curvature factor is the convective heat transfer efficiency parameter of the concave-convex surface of the food material;

[0067] Step 102, feature fusion of the three-dimensional heat diffusion morphology features to generate an initial morphology feature vector.

[0068] In the embodiment of the present application, 16x16 infrared thermal imaging sensor arrays are uniformly arranged on the top of the heating cavity of the baking equipment, the adjacent sensor spacing is 5mm, covering a food material area with a diameter of 20cm, ensuring at least 4 temperature measurement points per square centimeter of surface. The sensor uses a non-cryogenic microbolometer, the response waveband is 8-14pm, the temperature measurement range is -20℃-300℃, the built-in thermoelectric cooler maintains the constant temperature of the detector, avoiding the influence of environmental temperature on the measurement accuracy. The sampling frequency is set to 10Hz, two-point correction is performed before each collection: through the electrically controlled shutter switching, 0℃ and 100℃ blackbody radiation sources are collected respectively, and the offset and gain coefficient table of each pixel (such as the offset of a certain pixel is -0.3℃, and the gain coefficient is 1.02) is established to compensate the temperature deviation in real time; the original temperature matrix is subjected to median filtering (3x3 window) to eliminate salt and pepper noise, and then the temperature field is smoothed through Gaussian filtering (σ=1.0) to reduce the measurement fluctuation caused by surface texture. Every week, the imaging area is calibrated by 9 high-precision platinum resistance thermometers (accuracy ±0.1℃) (corners and center), and a polynomial correction model (quadratic polynomial) of temperature-pixel value is established. A triangulation method laser thickness measurement module (semiconductor laser wavelength 650nm, CMOS image sensor resolution 1280x720) is installed on the side guide rail of the equipment and can move along the X-Y axis in two dimensions, with a scanning range of 30cmx30cm and a vertical measurement accuracy of ±0.05mm.

[0069] Static food material: adopt a zigzag scanning path, the scanning spacing is 0.5mm, and it takes 15 seconds to complete a 20cmx20cm area scanning;

[0070] Dynamic food material (such as food material on a conveyor belt): collect a thickness point every 0.5mm movement triggered by an encoder, and realize dynamic thickness profile reconstruction by cooperating with the motion speed (maximum 10mm / s).

[0071] Abnormal point elimination is performed on the original thickness data: if the thickness difference between a certain point and the adjacent 8 points exceeds 0.5mm, it is determined to be a splash or oil droplet interference, and repaired by bilinear interpolation method; the sparse scanning points are encrypted by cubic spline interpolation to generate a thickness matrix with a resolution of 1mmx1mm, and the edge area (within 5mm from the profile) is encrypted to 0.5mmx0.5mm, ensuring that the concave and convex details are not lost.

[0072] Temperature sensor PT100 platinum resistance (class A precision, ±0.1℃), humidity sensor using capacitive polymer film sensor (precision ±1.5% RH), integrated in stainless steel probe, placed in the heating cavity 10 cm from the food surface, avoiding direct area of heating tube, probe surface coated with oil-repellent coating to reduce pollution. Temperature sampling frequency 10 Hz, humidity 5 Hz, converted to digital signal by 24-bit ADC; According to the sensor's own temperature monitored by the built-in NTC thermistor, the resistance value of PT100 is linearly corrected (for example, PT100 resistance is 109.7Ω at 25℃, and the actual ambient temperature is compensated); Using pre-calibrated humidity-temperature correction table (for example, humidity 40% RH at 25℃, corrected to 38.5% RH at 30℃), the actual environmental humidity is calculated by linear interpolation.

[0073] Step 101, Canny edge detection is performed on the temperature distribution matrix. First, the image is smoothed with a Gaussian filter (σ = 1.5), then the edges are extracted by non-maximum suppression and double threshold segmentation, and finally the broken edges are connected by morphological closing operation (3x3 rectangular kernel) to obtain the complete contour of the food. For each point on the contour, a five-point quadratic polynomial is used to fit the local curve, and the curvature value of the point is calculated. The specific steps are as follows:

[0074] Take the current point and its two adjacent points before and after it to construct a quadratic polynomial y = ax 2 + bx + c, solve the coefficients by least squares method, curvature Where x is the horizontal coordinate (e.g. along the length of the food); y is the vertical coordinate (e.g. along the thickness or height of the food); a is the core parameter of the local surface concave-convex degree, which determines the opening direction and curvature size of the curve (a > 0 for opening upward, a < 0 for opening downward); b is the position of the symmetry axis of the curve; c is the intercept of the curve on the y-axis. For non-contour points (inside the food), the distance to the nearest contour is calculated by distance transformation as a virtual curvature reference value. A curvature-thermal diffusivity lookup table is established, and the thermal diffusivity at different curvatures is measured (e.g. curvature 0.1 mm -1 corresponds to diffusivity 1.2 W / (m·K), curvature -0.1 mm -1 corresponds to 0.8 W / (m·K)), and the calculated curvature value is linearly interpolated to generate a thermal diffusivity parameter matrix, where each element in the matrix corresponds to the diffusivity of a specific location on the surface of the food.

[0075] Thickness distribution factor calculation:

[0076] Based on the laser thickness measurement data, K-means clustering algorithm is used to divide the food into several homogeneous regions. The clustering parameters are set as follows: distance threshold 0.5 mm, minimum region area 5 mm 2For irregular food materials (such as bone-in meat), first identify the key areas (bone, lean meat, fat) through contour convex hull analysis, and then cluster them respectively. Calculate the ratio of the average thickness of each area to the overall average thickness as the thickness difference coefficient (for example, the area average thickness is 10 mm, the overall average is 8 mm, and the coefficient is 1.25). For areas with sudden thickness changes (such as the edge suddenly dropping from 10 mm to 2 mm), introduce a gradient factor, the greater the gradient, the higher the coefficient weight (when the gradient > 1 mm / mm, the weight is multiplied by 1.5). According to the Fourier heat conduction law, establish a thickness-heat conduction efficiency model: the greater the thickness, the lower the efficiency, where α is the material thermal conductivity coefficient (beef α = 0.5 mm / W, vegetable α = 0.3 mm / W), Δh is the thickness difference coefficient -1, and finally the heat conduction efficiency value of each area is obtained. By looking up the table, call the corresponding α value for different food types, and finally generate the thickness distribution factor (heat conduction efficiency parameter matrix) according to the spatial position.

[0077] Surface curvature factor calculation:

[0078] Combined with temperature distribution and geometric shape factor, identify convex, concave and flat areas on the surface of the food material. Convex is defined as curvature > 0.05 mm -1 , concave curvature < -0.05 mm -1 , and the rest is flat. Using a simplified natural convection model, for convex areas, where ΔT is the temperature difference between the surface and the environment (unit ℃); for concave areas , the flat area h = 5. In actual calculation, the shape correction factor (such as the hemispherical convex correction factor 1.2, the V-shaped concave correction factor 0.8) is used to modify the model, and the calculated heat transfer coefficient is normalized to the [0, 1] interval. Take the flat area as the reference (value 1), the convex area value > 1, and the concave area value < 1 to generate the surface curvature factor matrix, and the matrix elements reflect the multiple of the convection heat transfer efficiency of each area relative to the plane.

[0079] Step 102, find the maximum value Vmax and the minimum value Vmin of the thermal diffusion rate parameter matrix, use the minimum-maximum standardization to map the data to [0, 1], for abnormal values (such as noise caused by sudden points), use the 3σ principle to identify and replace with the mean value of adjacent points; similarly, standardize the heat conduction efficiency parameter matrix, but considering that the thickness difference may be negative (such as the area thickness is less than the overall average), use Z-score standardization, and then map the data to [0, 1] through translation transformation. Since the heat transfer coefficient has been normalized, directly perform smoothing processing, and use a 5x5 Gaussian kernel convolution to eliminate local fluctuations, with a kernel function standard deviation σ = 1.0.

[0080] By orthogonal experimental design, the geometric shape factor weight w1=0.4, the thickness distribution factor w2=0.35, the surface curvature factor w3=0.25, the experimental variables include food material types (beef, chicken breast, vegetables), thickness range (2-10mm), surface curvature (plane, convex, concave), and the weight is optimized with heating uniformity as the evaluation index. When the standard deviation of the thickness of the food material is greater than 1.5mm, w2 is automatically increased to 0.45 and w1 is decreased to 0.35; when the surface curvature standard deviation is greater than 0.1mm -1 , w1 is increased to 0.45 and w3 is decreased to 0.2. The dynamic correction threshold is obtained by historical data statistics (such as the thickness standard deviation > 1.5mm with a probability of 15% to trigger correction). The three standardized matrices are subjected to principal component analysis (PCA) respectively, the first three principal components (cumulative variance contribution rate > 90%) are retained for the geometric shape factor, the first two principal components are retained for the thickness distribution factor, and the first two principal components are retained for the surface curvature factor, and the vector dimension is reduced. The weighted sum of the reduced vectors of each factor is calculated according to the weight, for example, when the vector dimension is n, the i-th element of the initial morphological feature vector F(i)=w1×G(i)+w2×T(i)+w3×S(i), wherein G, T and S are the reduced vectors of geometry, thickness and curvature respectively. A spatial coordinate (X, Y) and a time stamp are added to each initial morphological feature vector to form a four-dimensional feature vector (X, Y, T, F), wherein F is the initial morphological feature vector.

[0081] Through infrared thermal imaging, laser thickness measurement and environmental sensing, the surface temperature, thickness gradient and environmental parameters of the food material are collected in real time, the problem of missing thermal field information caused by traditional single-point temperature measurement is solved, and comprehensive data support is provided for dynamic control. Factors such as geometric shape, thickness distribution and surface curvature are extracted to quantify the influence of food material morphology on heat conduction, and a feature model matching the actual heat diffusion process is constructed to improve the pertinence of the control strategy. Through the fusion of multi-dimensional features by standardization and weight allocation, a comprehensive feature vector is formed, which can adapt to the thermal characteristic description of food materials with different morphologies (such as irregular shape and uneven thickness).

[0082] In a preferred embodiment of the present application, step 2 is based on the real-time deviation value of the set target temperature and the measured core temperature of the food material, combined with the morphological feature vector, to calculate the pseudo-derivative estimate value, and according to the pseudo-derivative estimate value, to generate the periodic intermittent power parameter, including duty cycle and period, which can include:

[0083] Step 200, obtaining an initial morphological feature vector, real-time calculating the deviation value of the set target temperature and the measured core temperature of the food material;

[0084] Step 201, input the deviation value and the initial morphological feature vector into the pseudo partial derivative estimator, process the thickness distribution factor in the initial morphological feature vector, generate the thickness weight coefficient, and fuse the thickness weight coefficient with the historical deviation data, and output the pseudo partial derivative estimation value by rolling optimization through the dynamic weighted least squares algorithm, the pseudo partial derivative estimation value represents the influence intensity of unit power change on the core temperature;

[0085] Step 202, input the pseudo partial derivative estimation value into the intermittent power generator, calculate the basic duty ratio, wherein the basic duty ratio is inversely proportional to the pseudo partial derivative estimation value; obtain the geometric shape factor in the initial morphological feature vector, and perform curvature compensation on the periodic value according to the geometric shape factor, and output the periodic intermittent power parameter, including the duty ratio optimized by the pseudo partial derivative and the periodic value compensated by the geometric shape factor.

[0086] In the embodiment of the present application, the three-dimensional thermal diffusion morphological feature fused vector is read from the output result of step 1, which contains geometric shape factors, thickness distribution factors, surface curvature factors and other components (for example, the dimension is 1x n, n is determined by the scanning accuracy of the food material, such as every 1cm 2 corresponding to one feature point). The latest 5 groups of feature vectors are cached and smoothed by using the moving average method:

[0087] The current vector and the previous 4 groups of vectors are assigned weights according to the time distance (such as the latest group weight 0.4, the previous group 0.3, and so on), which eliminates the feature jump caused by the slight movement of the food material or the noise of the sensor. The thickness distribution factor is mapped to [0, 1] (0 represents the thinnest area and 1 represents the thickest area), the geometric shape factor is mapped to [-1, 1] through curvature calculation (negative value for concave surface and positive value for convex surface), and the surface curvature factor is mapped to [0, 2] according to the convective heat transfer efficiency. A K-type thermocouple with a diameter of 0.5mm is buried in the geometric center of the food material (the depth is 1 / 2 of the thickness), and the sampling frequency is set to 10Hz. Before each sampling, the ambient temperature is measured by the PT100 sensor, and the cold end compensation of the thermocouple is performed (for example, when the ambient temperature is 25℃, the thermocouple output voltage needs to be superimposed with the thermal electromotive force value corresponding to 25℃). The target temperature is set by the user (such as the medium degree of beef steak at 60℃), and the real-time deviation value is the difference between the target temperature and the measured temperature. In order to filter high-frequency noise, a first-order low-pass filter is used: current deviation value = 0.7x current measured deviation + 0.3x previous time deviation, and the filter coefficient is dynamically adjusted according to the thermal inertia of the food material (coefficient 0.8 for thick food material and 0.6 for thin food material). A FIFO queue with a length of 50 is established to store the filtered deviation value, and each data point is attached with a time stamp (accurate to 0.1 seconds), when new data is input, the oldest data is removed from the queue tail, ensuring that the queue always contains the deviation data of the last 5 seconds (sampling interval 0.1 seconds).

[0088] Step 201, extract thickness distribution factor component from feature vector, original thickness data (such as 2-10mm) is mapped to [0, 1] by linear mapping. Calculate the standard deviation of the thickness distribution factor, if the standard deviation <0.5 (corresponding to the actual thickness fluctuation <1mm), the weight coefficient is wth=0.5+0.5×thickness distribution factor mean, where the mean ∈ [0, 1], so wth ∈ [0.5, 1]; If there is a region thickness <1 / 2 of the average thickness (such as average thickness 8mm, some region <4mm), the weight coefficient is wth=0.8×thickness distribution factor maximum+0.2×mean, where the maximum ∈ [0.5, 1], the mean ∈ [0.3, 0.7], so wth ∈ [0.62, 0.94]. Traverse each point in the thickness distribution factor, calculate the difference coefficient (regional thickness / average thickness); If more than 30% of the points difference coefficient >1.5, trigger weight enhancement: wth=wth×1.2, where wth=wth×1.2 is essentially a linear amplification operation on the current weight coefficient, the left "wth" is the updated weight, the right "wth" is the weight before updating, 1.2 is a dynamic enhancement factor, used to quantify the weight increment when the thickness mutates, finally truncated to [0.5, 1.2] (to avoid weight over-limit).

[0089] For each deviation value efilter(i) in the history queue, calculate the time difference Δt = current time - ti, where ti is the sampling time of the i-th deviation value in the history queue; The weight function is exponential decay: Where the time constant τ = 10 seconds. The specific performance is:

[0090] When Δt <2 seconds, w(i)>0.8 (such as 1 second weight ≈0.90);

[0091] When Δt = 5 seconds, w(i) ≈0.61;

[0092] When Δt >10 seconds, w(i)<0.37 (such as 15 seconds weight ≈0.22).

[0093] The thickness weight coefficient wth is multiplied by the time decay weight w(i) of each deviation value, and then multiplied by the deviation value itself, and the fusion value =∑(wth×w(i)×efilter(i)) is obtained by accumulation, which reflects the coupling strength of thickness difference and temperature deviation, the greater wth and the greater the absolute value of recent deviation, the greater the absolute value of the fusion value. Assuming that the core temperature change AT and the power change AP satisfy the linear relationship: AT=φ×AP+ε, where φ is the pseudo partial derivative, which is the influence strength of unit power (W) on core temperature (℃), and ε is a random error. The latest (AP, AT) data pairs are stored in a sliding window with a length of 20, and the oldest data is removed each time new data is input. The data in the window needs to satisfy |AP|>50W (filtering small power fluctuations), and |AT|>0.5℃ (ensuring significant temperature change). Each group of data in the window is assigned a weight w=wth×w(i) (w(i) is the time decay weight of the group of data), and the following is calculated respectively:

[0094] Numerator:∑(AP

[0095] Denominator:∑(AP 2 ×w), is the weighted sum of power changes;

[0096] Pseudo partial derivative The result is kept to three decimal places, with units of ℃ / W;

[0097] If φ<0.01℃ / W or φ>0.1℃ / W, take the boundary value and trigger a warning (prompting possible sensor failure or abnormal food material).

[0098] Step 202, the basic duty ratio D is inversely proportional to the pseudo partial derivative φ, and the calculation formula is Where k1, k2 are calibration coefficients (such as k1=0.05, k2=0.2), ensuring that D∈[20%, 80%]; the greater φ (the faster the unit power rises), the smaller the required duty ratio, avoiding overheating; the smaller φ (the slower the unit power rises), the greater the duty ratio, speeding up the heating speed; if D×rated power>90%×maximum power, then reduce the duty ratio in proportion: D=D×(0.9×maximum power)÷(D×rated power), to ensure that the power output does not exceed the upper limit of the device. Extract the geometric shape factor from the feature vector to calculate the average curvature Cavg of the food material surface:

[0099] Convex region: Cavg>0.05mm -1 (such as a hemispherical surface);

[0100] Concave region: Cavg<-0.05mm -1 (such as a spoon-shaped groove);

[0101] Flat area: |Cavg|≤0.05mm -1 .

[0102] Period compensation calculation:

[0103] Base period T0=10s (applicable to flat food materials);

[0104] When the convex surface is dominant, the compensation coefficient A=1-0.5×(Cavg-0.05)÷0.15, so that A∈[0.7, 1], and the final period T=T0×A (accelerate power switching, avoid the convex surface heat too fast);

[0105] When the concave surface is dominant, the compensation coefficient B=1+0.3×(-Cavg-0.05) / 0.15, so that B∈[1, 1.2], and the final period T=T0×B (extend the period, balance the concave heat aggregation);

[0106] Flat area: T=T0; if T<5s, set to 5s (avoid relay frequent action); if T>20s, set to 20s (ensure temperature stability).

[0107] By fusing the morphological features and temperature deviation in real time, the heat characteristic changes (such as thickness attenuation and water evaporation) in the food heating process can be tracked, the response delay of temperature control is shortened, and the heating real-time performance is improved. Based on the thickness distribution factor, a dynamic weight is generated to realize differentiated power regulation for different thickness areas of the food, effectively solving the problem of overcooking in thin areas and undercooking in thick areas caused by traditional uniform heating, and improving the heating uniformity. Intermittent power energy-saving optimization: the inverse ratio adjustment mechanism of duty cycle and pseudo partial derivative can automatically adjust the average power output according to the heat conduction efficiency of the food, avoiding energy waste in the constant power mode; the period adjustment of curvature compensation further optimizes the heat convection efficiency, and comprehensively improves the energy utilization rate. The rolling optimization algorithm continuously updates the pseudo partial derivative estimate value, which can adapt to the sudden change of food morphology (such as turning over) or environmental change (such as opening the door to dissipate heat), quickly re-converge to a stable control state, and reduce temperature fluctuations. The boundary constraint and feasibility verification mechanism in parameter calculation ensure that different power level devices can operate safely, avoid power over-limit or frequent hardware action, and prolong the service life of the device.

[0108] In a preferred embodiment of the present application, the above step 3, inputting the morphological feature vector into the dynamic compensation function, outputting the morphological correction coefficient, and based on the periodic intermittent power parameter, dynamically correcting the pseudo partial derivative estimate value, generating an adaptive partial derivative, can include:

[0109] Step 300, inputting the morphological feature vector into the dynamic compensation function, performing Gaussian kernel nonlinear mapping on the geometric shape factor and surface curvature factor to generate a curvature-diffusion coupling coefficient;

[0110] Step 301, based on the interaction between the coupling coefficient and the thickness distribution factor, output the morphology correction coefficient to quantify the compensation demand strength of the food material surface to heat diffusion;

[0111] Step 302, obtain the periodic intermittent power parameter, analyze the power density characteristics in the duty cycle and the switching frequency characteristics in the period value; fuse the morphology correction coefficient and the power switching frequency characteristics to construct a dynamic weight matrix, and take the morphology correction coefficient as the main correction factor to correct the pseudo partial derivative estimate value with variable gain, and output the adaptive partial derivative representing the actual influence strength of unit power on the core temperature after compensation.

[0112] In the embodiment of the application, the geometric shape factor and the surface curvature factor are accurately separated from the initial morphology feature vector. The geometric shape factor is stored in a matrix form, and each element corresponds to the curvature value of the food material surface 1cm 2 area. The curvature of the convex area is positive (such as the curvature of the semispherical convex part is 0.1mm -1 ), the concave area is negative (such as the curvature of the concave groove is -0.08mm -1 ), and the plane area is close to 0. The surface curvature factor is also stored in a matrix form, and the larger the value is, the higher the convective heat transfer efficiency is (such as the smooth plane takes the value of 0.6, and the rough surface takes the value of 0.9). The minimum-maximum standardization method is used to map the geometric shape factor value to the interval [-1, 1], and the surface curvature factor is mapped to the interval [0, 1]. For example, the original value 0.6 is still 0.6 after standardization (because the original range is 0-1). During the standardization process, the original maximum value is automatically recorded.

[0113] The standardized geometric shape factor and the surface curvature factor are respectively subjected to nonlinear transformation by using the Gaussian kernel function. The core logic of the Gaussian kernel function is to calculate the "similarity" between each factor value and the kernel center (set as 0), and output the weight value between 0 and 1.

[0114] Taking the geometric shape factor as an example: for a region with a standardized curvature value of 0.75, the Gaussian kernel function calculates the distance (i.e. 0.75-0=0.75) between the value and the kernel center 0, and determines the weight of the point in combination with the bandwidth parameter (default 0.3). The closer the distance (the closer the value to 0), the closer the weight to 1; the farther the distance, the closer the weight to 0. The surface curvature factor is the same, for example, the value 0.6 is calculated by the Gaussian kernel, and the corresponding weight is obtained according to the "similarity" between the value and the kernel center.

[0115] The preset weight (geometric factor 60%, surface curvature factor 40%) is multiplied element by element and summed. For example, the geometric factor weight of a certain point is 0.8, and the surface curvature factor weight of the corresponding point is 0.7, then the fused curvature-diffusion coupling coefficient is 0.8*0.6+0.7*0.4=0.76, and finally a coupling coefficient matrix of the same dimension as the original factor is generated, each element reflecting the comprehensive influence degree of the surface morphology of the corresponding region on heat diffusion.

[0116] Step 301: Interact the thickness distribution factor extracted by step 101 with the curvature-diffusion coupling coefficient generated in step 300. The thickness distribution factor is in matrix form, with each element corresponding to the normalized thickness value (0 represents the thinnest place, and 1 represents the thickest place) of the food material 1cm 2 above the surface. Multiply the thickness distribution factor matrix and the coupling coefficient matrix at the corresponding position elements to obtain an intermediate variable matrix. For example, the thickness distribution factor value of a certain region is 0.8 (thicker), and the coupling coefficient value is 0.76, then the intermediate variable is 0.8*0.76=0.608, indicating that the region has higher heat diffusion requirements due to the superposition of thickness and surface morphology. This process comprehensively considers the synergistic effect of internal heat conduction (thickness) and surface diffusion (curvature) of the food material.

[0117] The intermediate variable matrix is normalized by using the minimum-maximum standardization to map the numerical value to the [0, 1] interval to generate a morphology correction coefficient matrix. To eliminate local fluctuations caused by sensor noise or measurement errors, the morphology correction coefficient matrix is smoothed: a 5*5 sliding average filter is used (for larger size food materials, a 3*3 window is used for small size food materials). Taking a certain point as an example, the average value of the surrounding 25 points (5*5 range) is calculated to replace the original value, ensuring that the coefficient changes are continuous and consistent with physical laws. The final output morphology correction coefficient matrix accurately quantifies the heat diffusion compensation requirements of each region on the surface of the food material.

[0118] Step 302: Extract the duty cycle and period value from the periodic intermittent power parameter output in step 202. The duty cycle represents the proportion of heating power turned on in a unit of time (e.g. a duty cycle of 60% means 6 seconds of heating and 4 seconds of shutdown every 10 seconds), reflecting the average power density; the period value determines the power switching frequency (e.g. a period of 8 seconds means completing a heating-shutdown cycle every 8 seconds).

[0119] Feature decomposition is performed on the duty cycle and period value:

[0120] Duty cycle division: divided into high (greater than 70%), medium (30%-70%), and low (less than 30%) three intervals. For example, a duty cycle of 75% is classified as "high", indicating high power density requirements;

[0121] Period value division: divided into fast (less than 5 seconds), medium (5-10 seconds), and slow (more than 10 seconds). For example, a period of 4 seconds is classified as "fast", indicating high-frequency thermal regulation demand;

[0122] Each parameter value is quickly matched with the corresponding characteristic label through table lookup to form a discrete power regulation state description.

[0123] A two-dimensional dynamic weight matrix is constructed, with the row dimension corresponding to the morphological correction coefficient (divided into 11 levels with 0.1 intervals from 0 to 1, such as 0.0, 0.1,..., 1.0), and the column dimension corresponding to the three levels of power switching frequency (fast, medium, and slow). When the morphological correction coefficient is 1.0 (strong compensation demand) and the power switching frequency is slow, the weight is set to 0.8; when the morphological correction coefficient is 0.0 (no compensation demand) and the power switching frequency is fast, the weight is set to 0.2. In the process of establishing the weight matrix, multiple experiments are conducted for different food types (such as beef and chicken breast) and heating stages (warming and holding), and the weight distribution is optimized through cross-validation to ensure that the matrix covers all scene correction needs.

[0124] The morphological correction coefficient is used as the main correction factor, and the pseudo partial derivative estimate value of step 201 is corrected with the dynamic weight matrix. The specific steps are as follows:

[0125] According to the current morphological correction coefficient (such as 0.6) and the power switching frequency feature (such as "medium"), the weight value of the corresponding element (assuming 0.5) is found from the weight matrix;

[0126] The pseudo partial derivative estimate value is multiplied by the weight value to obtain the adaptive partial derivative. For example, the pseudo partial derivative estimate value is 0.06℃ / W, then the adaptive partial derivative is 0.06×0.5=0.03℃ / W, which reflects the actual temperature response strength after considering the food morphology and power regulation.

[0127] To prevent over-correction, the adaptive partial derivative is set with boundary constraints: the lower limit is 80% of the pseudo partial derivative estimate value (i.e. 0.06×0.8=0.048℃ / W), and the upper limit is 120% (i.e. 0.06×1.2=0.072℃ / W). If the calculation result exceeds the range (such as 0.03℃ / W below the lower limit), the boundary value (0.048℃ / W) is directly taken to ensure the stability and reliability of the control parameters and avoid abnormal operation of the equipment.

[0128] By Gaussian kernel mapping and multi-factor fusion, the dynamic compensation function can accurately capture the synergistic effect of food surface curvature and thickness distribution on heat diffusion, effectively solving the uneven heating problem caused by shape differences compared to traditional methods. The linkage optimization of shape correction coefficient and power parameter can adjust the correction strength of pseudo partial derivative and temperature control response speed in real time according to the shape change of food (such as thickness change caused by water evaporation during heating) and power adjustment demand (such as rapid heating or holding stage), and reduce temperature overshoot or lag phenomenon. The generation mechanism of adaptive partial derivative avoids power waste and temperature out of control, such as reducing excessive heating in concave areas and preventing rapid heat dissipation in convex areas, improving food processing quality. The dynamic weight matrix and boundary constraint mechanism ensure that the system can operate stably under different food types (such as slices and thick blocks), heating stages and environmental conditions, improve anti-interference ability and prolong equipment service life.

[0129] In a preferred embodiment of the present application, step 4 inputs the adaptive partial derivative into the dynamic power distributor, combines the intermittent power constraint condition and the temperature error tolerance, optimizes the step weight and correction strength through the directional parameter search algorithm, and generates a spatial power distribution parameter set, which can include:

[0130] Step 400 inputs the adaptive partial derivative into the dynamic power distributor to initialize the parameter search path with the adaptive partial derivative as the gradient descent direction reference;

[0131] Step 401 combines the intermittent power constraint condition, including maximum power threshold limit and minimum intermittent duration constraint, and combines the temperature error tolerance, including core temperature allowable fluctuation range and surface temperature uniformity threshold, and performs axial detection along the adaptive partial derivative direction through the directional parameter search algorithm, wherein the step weight is mapped to the power adjustment step size and the correction strength is mapped to the thermal response inertia compensation coefficient;

[0132] Step 402 generates a Pareto final solution set within the constraint boundary during axial detection, including partition radiation power intensity value, partition heating duration sequence and cross-zone power switching timing scheme.

[0133] In the embodiment of the present application, the adaptive partial derivative obtained from step 302 quantifies the actual impact strength of unit power change on core temperature (for example, 0.03℃ / W means that for every 1W increase in power, the core temperature will rise by 0.03℃). The adaptive partial derivative is used as a gradient descent direction reference, i.e. the target direction of power adjustment: if the current core temperature is lower than the target temperature, the gradient direction is to increase power; if it is higher than the target temperature, it is to reduce power. To ensure the accuracy of the adjustment direction, the adaptive partial derivative is verified twice: whether the current temperature deviation trend (such as continuous 3 times sampling temperature continuously rising or falling) is consistent with the sign of the partial derivative; if not consistent (for example, temperature rises but the partial derivative indicates to increase power), the standby adjustment logic is enabled, and the actual trend of temperature deviation is used as the reference to avoid false adjustment.

[0134] Starting from the current device operating parameters (such as current power, duty cycle, period), an initial search path is constructed in a multi-dimensional parameter space, which covers all adjustable variables, including partition radiation power intensity, heating duration, power switching timing, etc. The step weight is set to a small value by default (corresponding to a small amplitude power adjustment, such as 5W step), and the correction strength is set to a moderate level (thermal response inertia compensation coefficient is 0.5), to ensure the stability of the first adjustment and avoid temperature out of control due to parameter mutation.

[0135] Step 401, the device rated maximum power limit (for example, 2000W), the sum of the radiation power intensity of any partition cannot exceed this threshold; to protect the hardware (such as relay), the single heating or off duration cannot be lower than the safety value (such as 1 second), to prevent device damage caused by high frequency switching. The target temperature tolerance set by the user (such as ±2℃), ensures that the core temperature does not exceed this range during the adjustment process; the surface temperature standard deviation limit calculated by the infrared thermal imaging data (such as ≤3℃), avoids local overheating or overcooling.

[0136] Taking the direction determined by the adaptive partial derivative as the axis, step-by-step detection is performed in the parameter space. Each time, one parameter dimension is adjusted (such as first adjusting the radiation power of partition 1, then adjusting the heating duration), and other parameters remain unchanged, to separate the influence of each parameter on temperature. The step weight is converted into power adjustment step size. For example, a step weight of 0.2 corresponds to a 5W step, so the next time the power is increased or decreased by 5W; convert it into a thermal response inertia compensation coefficient, which is used to adjust the prediction model of temperature change, the higher the coefficient, the stronger the compensation for temperature inertia (for example, after heating, the temperature does not rise immediately, the delay effect needs to be considered). After each parameter adjustment, it is immediately verified whether it violates the intermittent power constraint and temperature error tolerance, if the new parameter causes the power to exceed the upper limit, the intermittent time is too short, or the core temperature exceeds the fluctuation range, the current detection direction is terminated, the previous valid parameter state is rolled back, and the step weight is reduced to re-detect.

[0137] Step 402, under the premise of meeting all constraint conditions, continue to iteratively adjust the parameters along the search direction. Record the current parameter combination and its corresponding temperature response result (such as the core temperature change amount, surface temperature uniformity index) each time. When approaching the constraint boundary, accept suboptimal solutions with a certain probability (for example, accept the parameter combination that makes the surface temperature uniformity slightly drop but can reach the core temperature faster), avoid the algorithm falling into local optimum, and expand the search range. Keep all parameter combinations within the constraint boundary and without other parameter combinations that can simultaneously optimize the core temperature and surface temperature uniformity, form a Pareto frontier, for example, parameter combination A makes the core temperature meet the standard but the surface uniformity is poor, and parameter combination B is the opposite, both are kept. Determine the optimal power distribution of each heating partition (such as 800W for partition 1 and 600W for partition 2); plan the heating-off time of each partition (such as 3 seconds heating and 2 seconds off for partition 1, and the cycle is executed); coordinate the power switching sequence of multiple partitions to avoid current impact caused by simultaneous start (such as 0.5 seconds after partition 1 is off, partition 2 is started); simulate the generated parameter set, predict the temperature change trend through historical data and heat conduction model, if the prediction result does not meet the tolerance requirement, select a suboptimal solution from the solution set for fine tuning, until a feasible spatial power distribution parameter set is generated.

[0138] Combined with the intermittent power and temperature error double constraints, it ensures that the power distribution meets the safe operation of the equipment (such as avoiding overload and high-frequency switching), and improves the heating quality of food. The adaptive partial derivative oriented directional search algorithm can quickly locate the final power parameters, shorten the temperature regulation time, and is especially suitable for the rapid heating stage of thick food. The Pareto solution set generation mechanism considers both the core temperature meeting the standard and the surface uniformity optimization, avoids the trade-off problem caused by single index optimization (such as the core temperature meeting the standard but the surface is burnt), and realizes the double improvement of energy saving and quality. The constraints such as minimum intermittent duration protect the hardware life and prevent the premature damage of relays and other components; the simulated annealing and rollback strategy ensures the stable operation of the system under complex working conditions, prolongs the equipment maintenance period.

[0139] In a preferred embodiment of the present application, the above step 5, the spatial power distribution parameter set is associated with the corresponding morphological feature vector and the periodic power parameter and stored in the heat diffusion knowledge base, and when a new food is detected, the spatial power distribution parameter set is automatically matched, which can include:

[0140] Step 500, establish a primary association mapping between the spatial power distribution parameter set and the morphological feature vector; and establish a secondary adjustment mapping between the spatial power distribution parameter set and the periodic power parameter;

[0141] Step 501, with the real-time geometric shape factor in the morphological feature vector as the heat field distribution benchmark key, with the real-time duty cycle dynamic range in the periodic power parameter as the power regulation benchmark key, and with the spatial power distribution parameter set associatedly stored, a thermal diffusion knowledge base storage structure is constructed;

[0142] Step 502, when a new food material is detected, the real-time morphological feature vector and the periodic power parameter of the new food material are extracted, the thermal diffusion characteristic fitness with the heat field distribution benchmark key is calculated based on the real-time geometric shape factor, and the dynamic compatibility with the power regulation benchmark key is verified based on the real-time duty cycle dynamic range; the spatial power distribution parameter set that passes the double verification is retrieved, and the partition radiation power intensity value, the partition heating duration sequence, and the cross-zone power switching timing scheme are loaded.

[0143] In the embodiment of the present application, the spatial power distribution parameter set (including partition radiation power intensity, heating duration sequence, and power switching timing) generated in step 402 is directly associated with the morphological feature vector (which records information such as the geometric shape, thickness distribution, and surface curvature of the food material), and each spatial power distribution parameter set corresponds to a unique morphological feature vector identifier (such as a feature code generated by a hash algorithm), ensuring that the parameter set corresponds to the food material morphology one-to-one.

[0144] During the association process, the morphological feature vector is compressed and stored, redundant data (such as repeated edge region features) is removed, and key feature points (such as curvature extreme points and thickness mutation points) are retained, reducing the vector dimension from 100 to about 20, to save storage space, and at the same time, an index relationship between the morphological feature vector and the original complete data is established. The spatial power distribution parameter set and the periodic power parameter (duty cycle, period value) are established for secondary mapping, which is used to fine-tune the power distribution strategy:

[0145] The duty cycle in the periodic power parameter determines the average power density, and the period value affects the power switching frequency, both of which jointly act on the specific execution of the spatial power distribution parameter set.

[0146] The secondary mapping adopts a hierarchical association method:

[0147] First layer: classified by duty cycle interval (such as low: less than 30%, medium: 30%-70%, high: greater than 70%);

[0148] Second layer: within each duty cycle interval, further subdivided by period value (such as fast: less than 5 seconds, medium: 5-10 seconds, slow: greater than 10 seconds).

[0149] For example, when the spatial power allocation parameter set corresponds to a duty cycle of 60% and a cycle value of 8 seconds, the system stores it under the "medium duty cycle-medium speed cycle" subdirectory and records the adjustment rules for the partition power intensity and heating duration (e.g. for every 10% increase in duty cycle, the partition 1 power is increased by 50W).

[0150] Step 501, extract the real-time geometric shape factor from the morphological feature vector as the core identifier representing the thermal field distribution characteristics of the food material. The geometric shape factor includes the curvature distribution information of convex, concave, and flat surfaces. The curvature range is divided into multiple intervals (e.g. convex: 0.05-0.1 mm -1 , concave: -0.1-0.05 mm-1), and each interval corresponds to a unique code. For example, the surface of a certain food material is mainly convex, with a curvature concentrated in 0.06-0.08 mm -1 , then its thermal field distribution reference key is "convex-0.06-0.08 mm -1 ". Extract the real-time duty cycle dynamic range from the periodic power parameters as the core identifier of the power adjustment strategy. Divide the duty cycle range into several segments (e.g. 20%-40%, 40%-60%) and combine historical data to label the power adjustment characteristics corresponding to each segment (e.g. for a constant temperature stage for thick food materials when the duty cycle is 40%-60%).

[0151] The heat diffusion knowledge base uses a tree-like hierarchical structure for storage:

[0152] Root node: classified by food material type (e.g. beef, chicken, fish);

[0153] First-level child node: thermal field distribution reference key (geometric shape factor classification);

[0154] Second-level child node: power adjustment reference key (duty cycle dynamic range);

[0155] Leaf node: stores the corresponding spatial power allocation parameter set.

[0156] For example, a piece of beef with a convex geometric shape (curvature 0.07 mm -1 ) and a duty cycle dynamic range of 40%-60%, its parameter set storage path is: knowledge base -> beef -> convex-0.05-0.1 mm -1 -> 40%-60% duty cycle -> specific parameter set file.

[0157] Construct a hash index table for the thermal field distribution reference key and the power adjustment reference key, respectively, to record the parameter set storage location corresponding to each key value. When retrieving, the target node can be directly located.

[0158] Step 502: When new food is detected, its morphological features are acquired in real time using sensors (such as laser scanners or infrared thermal imagers) to generate a real-time morphological feature vector. Periodic power parameters are set according to the current heating requirements (e.g., initial duty cycle 50%, period 10 seconds), and the real-time geometric shape factor of the new food is extracted and matched with the thermal field distribution reference bond. Based on the curvature distribution of the geometric shape factor, the Euclidean distance between the new food and the reference bond is calculated, with a distance ≤ 0.01 mm. -1 The timing is determined to be compatible. The real-time duty cycle dynamic range of the new ingredient is extracted and compared with the power adjustment baseline. The verification logic includes:

[0159] Does the duty cycle range completely or partially cover the reference key interval?

[0160] Does the power regulation characteristic corresponding to the current duty cycle match the needs of the food (e.g., thick food requires a high duty cycle to maintain heating)?

[0161] Only when the thermal diffusion characteristics and dynamic compatibility both pass verification is the corresponding spatial power allocation parameter set retrieved from the knowledge base. After finding a matching parameter set, the zoned radiant power intensity values, zoned heating duration sequences, and cross-zone power switching timing schemes are loaded. During the loading process, the parameters undergo secondary verification.

[0162] Check if the power intensity exceeds the equipment's maximum power limit;

[0163] Verify whether the heating duration meets the minimum interval constraint.

[0164] If there are parameter conflicts, the parameter fine-tuning mechanism will be automatically activated: the power intensity will be scaled proportionally (e.g., the power of all zones will be reduced by 10%), or the heating time will be adjusted (e.g., the shutdown time will be extended by 0.5 seconds) until the parameters meet the equipment operating requirements, and then the adjusted parameters will be applied to the heating process.

[0165] Rapid Adaptation and Efficient Heating: The heat diffusion knowledge base, through dual matching of morphological characteristics and power parameters, reduces the retrieval time for power allocation schemes of new ingredients to within 1 second, minimizing preheating waiting time. Parameter matching based on geometry and duty cycle ensures a high degree of fit between the heat field distribution of different shaped ingredients and the power adjustment strategy, avoiding local overcooking or undercooking issues. The parameter set accumulated in the knowledge base can serve as historical experience to guide subsequent adjustments, and parameter schemes for new ingredients are automatically added to the database after practical verification, forming a "learning-application-optimization" closed loop. The parameter loading verification and fine-tuning mechanism effectively avoids equipment overload or high-frequency switching, extending hardware lifespan.

[0166] like Figure 2 As shown, embodiments of the present invention also provide an energy-saving and heat-preserving adaptive dynamic control system for grilling equipment, comprising:

[0167] The feature extraction module is used to collect real-time data on surface temperature distribution, thickness gradient, and ambient temperature and humidity of food ingredients, extract three-dimensional heat diffusion morphological features, and generate an initial morphological feature vector.

[0168] The power parameter calculation module is used to calculate the pseudo-partial derivative estimate based on the real-time deviation between the set target temperature and the measured core temperature of the food, combined with the morphological feature vector, and generate periodic intermittent power parameters.

[0169] The dynamic correction module is used to input the morphological feature vector into the dynamic compensation function and output the morphological correction coefficient. Based on the periodic intermittent power parameter, it dynamically corrects the pseudo-partial derivative estimate and generates the adaptive partial derivative.

[0170] The parameter optimization module is used to input adaptive partial derivatives into the dynamic power allocator, combine intermittent power constraints and temperature error tolerance, and optimize the step size weight and correction intensity through a directional parameter search algorithm to generate a spatial power allocation parameter set.

[0171] The matching module is used to associate the spatial power allocation parameter set with the corresponding morphological feature vector and periodic power parameters and store them in the heat diffusion knowledge base. When a new food ingredient is detected, the spatial power allocation parameter set is automatically matched.

[0172] It should be noted that this system is a system corresponding to the above method. All implementation methods in the above method embodiments are applicable to this embodiment and can achieve the same technical effect.

[0173] Embodiments of the present invention also provide a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0174] Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0175] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An energy-saving and heat-preserving adaptive dynamic control method for grilling equipment, characterized in that, The method includes: Step 1: Collect real-time data on surface temperature distribution, thickness gradient, and ambient temperature and humidity of the food, extract three-dimensional heat diffusion morphological features, and generate an initial morphological feature vector. Step 2: Based on the real-time deviation between the set target temperature and the measured core temperature of the food, and combined with the morphological feature vector, calculate the pseudo-partial derivative estimate, and generate periodic intermittent power parameters, including duty cycle and period, based on the pseudo-partial derivative estimate. Step 3: Input the morphological feature vector into the dynamic compensation function, output the morphological correction coefficient, and dynamically correct the pseudo-partial derivative estimate based on the periodic intermittent power parameter to generate the adaptive partial derivative. Step 4: Input the adaptive partial derivative into the dynamic power allocator, combine the intermittent power constraint and temperature error tolerance, and optimize the step size weight and correction intensity through the directional parameter search algorithm to generate the spatial power allocation parameter set; Step 5: Associate the spatial power allocation parameter set with the corresponding morphological feature vector and periodic power parameter and store them in the heat diffusion knowledge base. When a new food ingredient is detected, the spatial power allocation parameter set will be automatically matched.

2. The energy-saving and heat-preserving adaptive dynamic control method for grilling equipment according to claim 1, characterized in that, Real-time data collection of food surface temperature distribution, thickness gradient, and ambient temperature and humidity is used to extract three-dimensional heat diffusion morphological features and generate an initial morphological feature vector, including: The infrared thermal imaging array collects real-time data on the surface temperature distribution of food, the laser thickness sensor obtains data on the thickness gradient of food, and the temperature and humidity sensor collects data on the ambient temperature and humidity. Based on surface temperature distribution, thickness gradient and ambient temperature and humidity data, three-dimensional thermal diffusion morphology features are extracted. The three-dimensional thermal diffusion morphological features are fused to generate an initial morphological feature vector.

3. The energy-saving and heat-preserving adaptive dynamic control method for grilling equipment according to claim 2, characterized in that, The three-dimensional thermal diffusion morphological characteristics include: Geometric shape factor, which is the thermal diffusion rate parameter of the surface curvature distribution of food ingredients; Thickness distribution factor, which is the heat conduction efficiency parameter of the thickness difference in different areas of the food. Surface curvature factor is a parameter for the convective heat transfer efficiency of the uneven surface of food.

4. The energy-saving and heat-preserving adaptive dynamic control method for grilling equipment according to claim 3, characterized in that, Based on the real-time deviation between the set target temperature and the measured core temperature of the food, and combined with the morphological feature vector, a pseudo-partial derivative estimate is calculated. Based on this estimate, periodic intermittent power parameters, including duty cycle and period, are generated. Obtain the initial morphological feature vector and calculate the deviation between the set target temperature and the actual core temperature of the food in real time; The deviation value and the initial morphological feature vector are input into the pseudo-partial derivative estimator. The thickness distribution factor in the initial morphological feature vector is processed to generate the thickness weight coefficient. The thickness weight coefficient is then fused with historical deviation data and rolled optimization is performed through a dynamic weighted least squares algorithm to output the pseudo-partial derivative estimate. The pseudo-partial derivative estimate represents the intensity of the effect of unit power change on core temperature. The pseudo-partial derivative estimate is input into the intermittent power generator to calculate the base duty cycle, which is inversely proportional to the pseudo-partial derivative estimate. The geometric shape factor in the initial morphological feature vector is obtained, and the period value is curvature compensated according to the geometric shape factor. The periodic intermittent power parameters are output, including the duty cycle optimized by the pseudo-partial derivative and the period value compensated by the geometric shape factor.

5. The energy-saving and heat-preserving adaptive dynamic control method for grilling equipment according to claim 4, characterized in that, The morphological feature vector is input into the dynamic compensation function, which outputs morphological correction coefficients. Based on the periodic intermittent power parameters, the pseudo-partial derivative estimate is dynamically corrected to generate adaptive partial derivatives, including: The morphological feature vector is input into the dynamic compensation function, and a Gaussian kernel nonlinear mapping is performed on the geometric shape factor and the surface curvature factor to generate the curvature-diffusion coupling coefficient. Based on the interaction between the coupling coefficient and the thickness distribution factor, the morphology correction coefficient is output to quantify the intensity of the food surface's need for compensation for heat diffusion. The periodic intermittent power parameters are obtained, and the power density characteristics in the duty cycle and the switching frequency characteristics in the period value are analyzed. The morphology correction coefficient and the power switching frequency characteristics are fused to construct a dynamic weight matrix. The morphology correction coefficient is used as the main correction factor to perform variable gain correction on the pseudo-partial derivative estimate. The adaptive partial derivative is output to represent the actual influence intensity of unit power on core temperature after compensation.

6. The energy-saving and heat-preserving adaptive dynamic control method for frying and grilling equipment according to claim 5, characterized in that, The adaptive partial derivative is input into the dynamic power allocator. Combined with intermittent power constraints and temperature error tolerance, a directional parameter search algorithm optimizes the step size weights and correction strength to generate a spatial power allocation parameter set, including: The adaptive partial derivative is input into the dynamic power allocator, and the parameter search path is initialized using the adaptive partial derivative as the gradient descent direction reference. Combining intermittent power constraints, including maximum power threshold limits and minimum intermittent duration constraints, and temperature error tolerances, including the allowable fluctuation range of core temperature and the surface temperature uniformity threshold, an axial probe is performed along the adaptive partial derivative direction using an directional parameter search algorithm. The step size weight is mapped to the power adjustment step size, and the correction intensity is mapped to the thermal response inertia compensation coefficient. During the axial detection process, a Pareto final solution set is generated within the constraint boundary, which includes the radiative power intensity values ​​of each zone, the heating duration sequence of each zone, and the power switching timing scheme across zones.

7. The energy-saving and heat-preserving adaptive dynamic control method for grilling equipment according to claim 6, characterized in that, The spatial power allocation parameter set is associated with and stored in the heat diffusion knowledge base along with the corresponding morphological feature vector and periodic power parameters. When a new food ingredient is detected, the spatial power allocation parameter set is automatically matched, including: Establish a primary correlation mapping between the spatial power allocation parameter set and the morphological feature vector; establish a secondary adjustment mapping between the spatial power allocation parameter set and the periodic power parameter. Using the real-time geometric shape factor in the morphological feature vector as the reference key for thermal field distribution and the real-time duty cycle dynamic range in the periodic power parameter as the reference key for power adjustment, and associating it with the power allocation parameter set of the storage space, a thermal diffusion knowledge base storage structure is constructed. When a new food ingredient is detected, the real-time morphological feature vector and periodic power parameters of the new food ingredient are extracted. The thermal diffusion characteristics of the thermal field distribution reference key are calculated based on the real-time geometric shape factor, and the dynamic compatibility of the power adjustment reference key is verified based on the real-time duty cycle dynamic range. The set of spatial power allocation parameters that have passed dual verification is retrieved, and the zone radiation power intensity value, zone heating duration sequence, and cross-zone power switching timing scheme are loaded.

8. An energy-saving and heat-preserving adaptive dynamic control system for grilling equipment, wherein the system implements the method as described in any one of claims 1 to 7, characterized in that, include: The feature extraction module is used to collect real-time data on surface temperature distribution, thickness gradient, and ambient temperature and humidity of food ingredients, extract three-dimensional heat diffusion morphological features, and generate an initial morphological feature vector. The power parameter calculation module is used to calculate the pseudo-partial derivative estimate based on the real-time deviation between the set target temperature and the measured core temperature of the food, combined with the morphological feature vector, and generate periodic intermittent power parameters. The dynamic correction module is used to input the morphological feature vector into the dynamic compensation function and output the morphological correction coefficient. Based on the periodic intermittent power parameter, it dynamically corrects the pseudo-partial derivative estimate and generates the adaptive partial derivative. The parameter optimization module is used to input adaptive partial derivatives into the dynamic power allocator, combine intermittent power constraints and temperature error tolerance, and optimize the step size weight and correction intensity through a directional parameter search algorithm to generate a spatial power allocation parameter set. The matching module is used to associate the spatial power allocation parameter set with the corresponding morphological feature vector and periodic power parameters and store them in the heat diffusion knowledge base. When a new food ingredient is detected, the spatial power allocation parameter set is automatically matched.

9. A computing device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that, when executed by a processor, implements the method as described in any one of claims 1 to 7.

Images