An intelligent roasting method based on feedback adjustment of internal environment parameters of a roasting oven

Abstract

The application discloses an intelligent baking method based on feedback adjustment of internal environment parameters of a baking oven, and belongs to the technical field of data processing. The method comprises the following steps: generating a smooth temperature state estimator according to a temperature distribution matrix, and calculating a temperature uniformity index; if the temperature uniformity index exceeds a first preset threshold, generating a heating power ratio based on the smooth temperature state estimator; calculating a humidity influence coefficient based on the heating power ratio and a humidity change sequence; if a fluctuation amplitude of the humidity change sequence exceeds a second preset threshold, generating a humidification amount adjustment value based on the humidity influence coefficient; processing the humidification amount adjustment value and a gas concentration value to generate heating power and ventilation rate set values; and generating the power of each heating element and the rotating speed of a fan based on the heating power and ventilation rate set values. The intelligent baking method based on feedback adjustment of internal environment parameters of the baking oven solves the problem that the existing baking oven adjustment system is difficult to realize multi-parameter coordinated adjustment, thereby leading to unstable quality of baked food.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to an intelligent baking method based on internal environment parameter feedback adjustment of a baking oven. BACKGROUND

[0002] Food baking, as a core link of modern food processing, directly determines the taste, nutritional value and commercial quality of food, and occupies a pivotal position in the food industry. With the continuous improvement of consumers' requirements for food quality and the urgent needs of the development of food industry, the accurate adjustment of baking process has become a key factor affecting the competitiveness of enterprises. Traditional baking methods mainly rely on the experience judgment and manual adjustment of operators, which not only has high labor intensity, but also is difficult to ensure the consistency of baking parameters.

[0003] Although the existing automatic baking equipment reduces manual intervention to some extent, it mainly adopts preset program adjustment and lacks dynamic response ability to actual environmental changes in the baking process, resulting in large fluctuations in baking quality. The adjustment of environmental parameters in the baking process faces complex technical challenges. The temperature distribution, humidity change and gas composition concentration in the internal environment of the baking oven interact with each other, forming a complex system that changes dynamically. Non-uniform distribution of temperature will directly affect the consistency of food heating, and then cause differences in baking effect. Fluctuations in humidity level will change the evaporation rate of water on the surface of food, affecting the texture and color formation of food. Changes in gas composition not only relate to the progress of baking chemical reactions, but also affect the development of food flavor. There is a complex coupling relationship between these environmental parameters, and the change of a single parameter often causes a chain reaction of other parameters.

[0004] The current adjustment system is difficult to realize the coordinated adjustment of these interrelated parameters, and cannot adjust in real time according to the actual baking state, resulting in that the environmental parameters deviate from the optimal range in the baking process, and finally affect the quality stability of the baked food. SUMMARY

[0005] In order to overcome the defects existing in the prior art, the present application provides an intelligent baking method based on internal environment parameter feedback adjustment of a baking oven to solve the above problems.

[0006] The technical scheme adopted by the present application to solve its technical problems is: an intelligent baking method based on internal environment parameter feedback adjustment of a baking oven, comprising the following steps:

[0007] S1: Collect a set of baking environment data from a sensor group, including a temperature distribution matrix, a humidity change sequence and a gas concentration value, and store it as an initial data set;

[0008] S2: generating a smoothed temperature state estimate from the temperature distribution matrix in the initial data set, calculating a temperature uniformity index from the smoothed temperature state estimate; if the temperature uniformity index exceeds a first preset threshold, generating a heating power ratio based on the smoothed temperature state estimate;

[0009] S3: calculating a humidity influence coefficient based on the heating power ratio and the humidity change sequence in the initial data set; if the fluctuation amplitude of the humidity change sequence in the initial data set exceeds a second preset threshold, generating a humidification amount adjustment value based on the humidity influence coefficient;

[0010] S4: processing the humidification amount adjustment value and the gas concentration value in the initial data set to generate heating power and ventilation rate setting values;

[0011] S5: generating a set of adjustment parameters including the power of each heating element and the fan speed based on the heating power and ventilation rate setting values.

[0012] Preferably, in the step S2, a Kalman filtering algorithm is applied to the temperature distribution matrix in the initial data set to generate a smoothed temperature state estimate, and the average of the variances of all smoothed temperature state estimates is calculated as the temperature uniformity index.

[0013] Optionally, the step of applying a Kalman filtering algorithm to the temperature distribution matrix in the initial data set in the step S2 includes:

[0014] performing mean filtering processing on the temperature distribution matrix T_origin using a 3x3 convolution kernel to obtain a smoothed temperature matrix T_smooth;

[0015] constructing an observation vector Z_k of Kalman filtering by taking each element of the smoothed temperature matrix T_smooth as an observation value, setting a state transition matrix F of Kalman filtering as an identity matrix, and setting an observation matrix H of Kalman filtering as an identity matrix;

[0016] performing a prediction-update cycle of Kalman filtering according to the observation vector Z_k, the state transition matrix F, and the observation matrix H to obtain a state estimate matrix X_estimate;

[0017] extracting node temperature values from the state estimate matrix X_estimate, and performing moving average filtering processing on the node temperature values with a window width of a preset value to obtain a smoothed state estimate matrix X_estimate, wherein the elements in the smoothed state estimate matrix X_estimate are smoothed temperature state estimates.

[0018] Specifically, in the step S2, the step of generating a heating power ratio based on a smoothed temperature state estimate includes:

[0019] A linear regression model is established based on the relationship between historical temperature data and historical heating power, and the heating power proportion P_i corresponding to each grid node in the smoothed state estimation matrix X_estimate is calculated for the smoothed temperature state estimation.

[0020] It is worth noting that in the step S3, a temperature-humidity coupling relationship model is established based on the heating power proportion and the humidity change sequence in the initial data set, and the humidity influence coefficient is calculated.

[0021] Specifically, in the step S3, the step of calculating the humidity influence coefficient includes:

[0022] The variance of the humidity change sequence H_seq of each grid node in the time dimension is calculated, and if the variance is greater than a preset threshold T_hum, a moving average filter with a window size of 5 is used to process the humidity change sequence H_seq to obtain a smoothed humidity change sequence H_seq_smooth; the smoothed humidity change sequence H_seq_smooth is mapped to the humidity distribution vector V_hum corresponding to each grid node according to the spatial coordinates of the grid nodes.

[0023] A linear regression model with intercept is used to calculate the slope parameter as the coupling relationship coefficient C_th, with the smoothed temperature state estimation as the independent variable and the humidity distribution vector V_hum as the dependent variable.

[0024] The humidity influence coefficient W_hum corresponding to each grid node is calculated through the coupling relationship coefficient C_th and the heating power proportion P_i corresponding to each grid node, where W_hum=C_th*P_i / sum(P_i), and sum(P_i) is the sum of the heating power proportions of all grid nodes.

[0025] Optionally, in the step S3, the humidity change sequence fluctuation amplitude calculation step includes:

[0026] The humidity time sequence is obtained, and the humidity time sequence data includes humidity measurement values in multiple time windows.

[0027] The standard deviation of the humidity measurement values in each time window is calculated using a numerical calculation method to obtain the corresponding fluctuation amplitude value B in each time window.

[0028] Preferably, in the step S3, the step of generating the humidification amount adjustment value based on the humidity influence coefficient includes:

[0029] The humidity influence coefficient W_hum corresponding to each grid node in the corresponding time window is extracted.

[0030] The grid nodes are sorted in descending order according to the numerical value of the humidity influence coefficient W_hum to obtain a grid node priority sequence;

[0031] The humidification amount adjustment value W_tz corresponding to each grid node is calculated by using a weighted distribution algorithm according to the numerical value of the humidity influence coefficient W_hum, and the weight value in the weighted distribution algorithm is determined by the ratio of the reciprocal of the priority sequence number to the total number of grid nodes.

[0032] Specifically, in the step S4, the humidification amount adjustment value W_tz of each grid node is obtained, and the gas concentration measurement data Q_nd of the corresponding time period is extracted from the initial data set;

[0033] The humidification amount adjustment value W_tz and the gas concentration value Q_nd are aligned according to the grid nodes to construct a double-input parameter matrix with the number of rows being the total number of grids and the number of columns being 2;

[0034] For the humidification amount adjustment value W_tz in the first column of the double-input parameter matrix, the humidification amount range is set, and the humidification amount range is divided into three humidification amount intervals of low, medium and high;

[0035] For the gas concentration value Q_nd in the second column of the double-input parameter matrix, the gas concentration range is set, and the gas concentration range is divided into three gas concentration intervals of thin, moderate and thick;

[0036] A rule base containing 9 rules is established according to the three humidification amount intervals and the three gas concentration intervals, and the rule format is that if the humidification amount adjustment value belongs to the humidification amount interval A and the gas concentration belongs to the gas concentration interval B, then the output heating power is C and the ventilation rate setting value is D;

[0037] After inputting the humidification amount adjustment value W_tz and the gas concentration measurement data Q_nd corresponding to the grid node, the corresponding heating power and ventilation rate setting value are output from the rule base.

[0038] Optionally, in the step S5, according to the heating elements in the pre-stored heating element number mapping table, the heating power value is split into a power distribution array composed of the operating power of multiple heating elements, and the operating power in the power distribution array is arranged in order according to the heating element number in the heating element number mapping table;

[0039] According to the ventilation rate setting value, the pre-established ventilation rate-fan speed correspondence table is queried to obtain the corresponding fan speed value.

[0040] The beneficial effects of this invention are as follows: In the intelligent baking method based on feedback adjustment of internal environmental parameters of the baking oven, temperature uniformity index is calculated by collecting temperature, humidity, and gas concentration data. When the temperature is uneven, a heating power ratio is generated based on the temperature state estimate. Then, the humidity influence coefficient is calculated. When the humidity fluctuates significantly, a humidification adjustment value is generated. Finally, the humidification adjustment value and gas concentration value are processed to generate heating power and ventilation rate setpoints, and the power distribution of heating elements and fan speed are adjusted accordingly. This invention achieves coordinated regulation of temperature, humidity, and gas concentration in the baking environment, improving the stability of baking quality. Attached Figure Description

[0041] Fig. 1 This is a flowchart of an intelligent baking method based on feedback adjustment of internal environmental parameters of a baking oven, as described in one embodiment of the present invention.

[0042] Fig. 2 This is a flowchart of step S2, which calculates the temperature uniformity index, in one embodiment of the present invention.

[0043] Fig. 3 This is a flowchart of step S3, calculating the humidity influence coefficient, in one embodiment of the present invention. Detailed Implementation

[0044] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings. It should be noted that these descriptions are for the purpose of aiding understanding the present invention, but do not constitute a limitation thereof. Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0045] like Figs. 1-3 As shown, an intelligent baking method based on feedback adjustment of internal environmental parameters of a baking oven includes the following steps:

[0046] S1: Collect a set of baking environment data from the sensor group, including a temperature distribution matrix, humidity change sequence, and gas concentration value, and store it as an initial dataset; In this embodiment, the oven is divided into multiple grid nodes. Each grid node is equipped with a temperature sensor, a humidity sensor, and a gas concentration sensor. The temperature collected by the temperature sensor corresponding to each grid is arranged according to the grid node to form a temperature distribution matrix. The humidity collected by the humidity sensor corresponding to each grid node is arranged according to time to form a humidity change sequence. The data collected by the gas concentration sensor of each grid node is used as the gas concentration value of the corresponding grid node.

[0047] S2: generating a smooth temperature state estimator according to the temperature distribution matrix in the initial data set, calculating a temperature uniformity index according to the smooth temperature state estimator; if the temperature uniformity index exceeds a first preset threshold, generating a heating power ratio based on the smooth temperature state estimator;

[0048] S3: calculating a humidity influence coefficient based on the heating power ratio and the humidity change sequence in the initial data set; if the fluctuation amplitude of the humidity change sequence in the initial data set exceeds a second preset threshold, generating a humidification amount adjustment value based on the humidity influence coefficient;

[0049] S4: processing the humidification amount adjustment value and the gas concentration value in the initial data set to generate heating power and ventilation rate setting values;

[0050] S5: generating a set of adjustment parameters including the power of each heating element and the fan speed based on the heating power and ventilation rate setting values.

[0051] In the intelligent baking method based on baking oven internal environment parameter feedback adjustment, temperature, humidity and gas concentration data are collected, and a temperature uniformity index is calculated. When the temperature is not uniform, a heating power ratio is generated based on the temperature state estimator. Then a humidity influence coefficient is calculated. When the humidity fluctuates greatly, a humidification amount adjustment value is generated. Finally, the humidification amount adjustment value and the gas concentration value are processed to generate heating power and ventilation rate setting values, and the heating element power distribution and fan speed are adjusted accordingly. The present application realizes the coordinated adjustment of temperature, humidity and gas concentration in the baking environment, and improves the stability of the baking quality.

[0052] It is worth noting that in the step S2, the Kalman filtering algorithm is applied to the temperature distribution matrix in the initial data set to generate a smooth temperature state estimator, and the average value of the variances of all smooth temperature state estimators is calculated as the temperature uniformity index.

[0053] Preferably, in the step S2, the step of applying the Kalman filtering algorithm to the temperature distribution matrix in the initial data set comprises:

[0054] A 3x3 convolution kernel is used to perform mean filtering processing on the temperature distribution matrix T_origin to obtain a smooth temperature matrix T_smooth, which represents the smooth temperature distribution;

[0055] Each element of the smooth temperature matrix T_smooth is taken as an observation value to construct an observation vector Z_k of Kalman filtering, which represents a temperature observation sequence; the state transition matrix F of Kalman filtering is set to an identity matrix to represent state transition, and the observation matrix H of Kalman filtering is set to an identity matrix to represent observation relationship;

[0056] According to the observation vector Z_k, the state transition matrix F and the observation matrix H, a prediction-update cycle of Kalman filtering is performed to obtain a state estimation matrix X_estimate;

[0057] Each node temperature value is extracted from the state estimation matrix X_estimate, and a moving average filtering process with a window width of a preset value is performed on the node temperature value to obtain a smoothed state estimation matrix X_estimate, and an element in the smoothed state estimation matrix X_estimate is a smoothed temperature state estimation quantity.

[0058] For the smoothed state estimation matrix X_estimate, the variance σ² of each element is calculated to obtain a variance matrix, and the variance matrix represents the uncertainty of temperature estimation; according to the variance matrix, the average value of all variances is calculated to obtain a temperature uniformity index U, and the temperature uniformity index U represents the uniformity degree of temperature distribution.

[0059] Beneficial effect: when the 3x3 convolution kernel mean filtering process is performed on the temperature distribution matrix T_origin, it can be understood that the average value of the temperature value of each pixel point and its surrounding 8 neighborhood points in the matrix is taken to generate a smoothed temperature matrix T_smooth. The core of this method is to reduce local temperature noise and retain the overall temperature distribution trend. In a possible implementation, assuming that the temperature distribution matrix collected by the sensor in the baking environment is 10x10 dimension, the temperature value of a certain pixel point is 85.0℃, and the temperature values of its 8 neighborhood points are 84.5℃, 85.2℃, 84.8℃, etc., and after mean filtering, the temperature of the point may be adjusted to 84.9℃. After such processing, the smoothed temperature matrix T_smooth can more stably reflect the temperature distribution of the baking environment and reduce abnormal fluctuations caused by sensor noise.

[0060] When constructing the observation vector Z_k of Kalman filtering, each element of the smoothed temperature matrix T_smooth is regarded as an observation value. For example, the T_smooth matrix of 10x10 dimension can be expanded into an observation vector Z_k of 100 dimensions, representing the temperature observation sequence of the entire baking area. The state transition matrix F is set to the unit matrix, which means that the temperature state is assumed to change relatively smoothly over time, which is suitable for the relatively stable scene of the baking environment. The observation matrix H is an identity matrix, which means that the sensor directly observes the temperature value without complex transformation. This setting simplifies the calculation complexity of Kalman filtering, while ensuring the accuracy of temperature state estimation.

[0061] Specifically, when performing the prediction-update cycle of Kalman filtering, the system predicts the temperature state of the next time according to the observation vector Z_k, and updates it combined with the actual observation value to generate the state estimation matrix X_estimate. For example, during the baking process, the initial temperature estimate of a certain area is 84.9°C, and the next time temperature may be 85.0°C through Kalman filtering prediction, but the actual observation value is 85.1°C, and the updated estimate value may be more accurate 85.05°C. This method significantly improves the robustness of temperature estimation by fusing prediction and observation.

[0062] The principle of moving average filtering is to take the average of the temperature values of a node and its adjacent nodes to reduce the influence of local noise. Assuming that the window width is set to 3, in the case of a grid node temperature value of 86.0 degrees Celsius, and the temperatures of its two adjacent nodes are 85.5 degrees Celsius and 86.2 degrees Celsius, respectively, the node temperature may be adjusted to 85.9 degrees Celsius after filtering. This method can effectively smooth abnormal fluctuations and make the temperature data more continuous.

[0063] Optionally, in the step S2, the step of generating the heating power ratio based on the smoothed temperature state estimate includes:

[0064] A linear regression model is used to calculate the heating power ratio P_i corresponding to each grid node in the smoothed state estimation matrix X_estimate based on the relationship between historical temperature data and historical heating power.

[0065] The principle of the linear regression model is to predict the power adjustment ratio required by each node under the current temperature distribution based on the relationship between historical temperature data and historical heating power. Assuming that a node temperature is 84.5 degrees Celsius, the linear regression model will suggest increasing the heating power ratio P_i of the node to 1.2 times to compensate for the temperature deficiency after analysis. This method can adjust the heating distribution in a targeted manner to improve overall temperature consistency. For example, from another side, if the temperatures of multiple nodes in a certain area are too high, the model may reduce the P_i value of these nodes, for example, to 0.8 times, to avoid overheating.

[0066] Specifically, in the step S3, a temperature-humidity coupling relationship model is established based on the heating power ratio and the humidity change sequence in the initial data set to calculate the humidity influence coefficient.

[0067] It is worth noting that in the step S3, the step of calculating the humidity influence coefficient includes:

[0068] The variance of the humidity change sequence H_seq of each grid node in the time dimension is calculated, and if the variance is greater than a preset threshold T_hum, a moving average filter with a window size of 5 is used to process the humidity change sequence H_seq to obtain a smoothed humidity change sequence H_seq_smooth; the smoothed humidity change sequence H_seq_smooth is mapped to the humidity distribution vector V_hum corresponding to each grid node according to the spatial coordinates of the grid node.

[0069] A linear regression model with intercept is used to calculate the slope parameter as the coupling relationship coefficient C_th, with the smoothed temperature state estimator as the independent variable and the humidity distribution vector V_hum as the dependent variable.

[0070] The humidity influence coefficient W_hum corresponding to each grid node is calculated through the coupling relationship coefficient C_th and the heating power ratio P_i corresponding to each grid node, where W_hum = C_th * P_i / sum(P_i), and sum(P_i) is the sum of the heating power ratios of all grid nodes.

[0071] In this embodiment, the linear regression model with intercept is humidity distribution vector = a * smoothed temperature state estimator + b, where a represents the slope parameter and b represents the intercept item. The value of the temperature distribution vector V_temp and the value of the smoothed temperature state estimator corresponding to the grid node are substituted into the linear regression model with intercept to solve the slope parameter a and the intercept parameter b in the regression equation. The slope parameter a is used as the temperature and humidity coupling relationship coefficient C_th.

[0072] For the variance calculation of the humidity change sequence H_seq in the time dimension, the stability of the humidity distribution can be determined. If the variance is large, the humidity fluctuation is obvious and needs to be further processed. Assuming that the humidity value of a certain grid node fluctuates greatly within a period of time, the variance value is 8.5, and the preset threshold T_hum is 5.0, obviously, smoothing needs to be performed. A moving average filter with a window size of 5 can be used to smooth the humidity value of each node in the humidity change sequence H_seq, reducing the influence of short-term fluctuations. For example, the humidity value sequence of a certain node is 62%, 64%, 60%, 63%, and 61%, and after filtering, it may be adjusted to 62%, making the data more stable.

[0073] The smoothed humidity change sequence H_seq reflects the change of humidity in the baking area over time, and by mapping it to the humidity distribution vector V_hum, the humidity distribution state of each grid node can be intuitively displayed. Assuming that in a 10x10 grid, the humidity change sequence H_seq of a certain grid node reflects that the humidity value changes from 60% to 65% within a period of time, and the value of the mapped humidity distribution vector V_hum is the average humidity of the grid node.

[0074] Regarding the analysis of the temperature-humidity coupling relationship, a linear regression model with intercept is constructed by taking the smoothed temperature state estimate as the independent variable and the humidity distribution vector V_hum as the dependent variable, which can quantify the correlation between the two. By substituting the smoothed temperature state estimate and the humidity distribution vector V_hum corresponding to multiple grid nodes into the linear regression model with intercept, a slope parameter a is obtained as the coupling relationship coefficient C_th through regression analysis. This coefficient reflects the degree of influence of temperature change on humidity, providing a basis for subsequent regulation.

[0075] When calculating the humidity influence coefficient W_hum, combined with the coupling relationship coefficient C_th and the heating power ratio P_i, a reasonable humidity adjustment weight can be assigned to each grid node. This method ensures that humidity regulation matches heating power distribution. For example, in a certain area, if the temperature is high and the heating power ratio is large, the corresponding humidity influence coefficient will also be high, indicating that attention should be paid to the humidity change in that area.

[0076] Specifically, in the step S3, the humidity change sequence fluctuation amplitude calculation step includes:

[0077] Obtain the humidity time series, which contains the humidity measurement values of multiple time windows;

[0078] Calculate the standard deviation of the humidity measurement values in each time window using numerical calculation method to obtain the corresponding fluctuation amplitude value B in each time window.

[0079] The humidity time series data is usually collected periodically by sensors in the baking environment, recording the humidity changes in a specific time period. For example, in a 24-hour monitoring period, humidity data is collected every 10 minutes, forming a sequence containing 144 time points. Assuming that the humidity data of a certain baking area in a certain time window is 58%, 60%, 59%, 62%, and 61%, these data reflect the dynamic changes of environmental humidity. By analyzing the time series, we can understand whether the humidity is stable or there are significant fluctuations, providing a basis for subsequent regulation. For example, by calculating the standard deviation of the humidity measurement values in each time window, the fluctuation amplitude value B of the humidity can be quantified. The standard deviation reflects the degree of data deviation from the mean, and the larger the value, the more obvious the humidity fluctuation. In a 10x10 grid baking environment, assuming that the humidity values of a certain node in a certain time window are 60%, 62%, 58%, 61%, and 59%, the B value is calculated to be 1.414 by standard deviation. If the second preset threshold T2 is 1.0, obviously the B value of the node exceeds the threshold, indicating that the humidity fluctuation is large and needs to be further processed.

[0080] Preferably, in the step S3, the step of generating the humidification amount adjustment value based on the humidity influence coefficient comprises:

[0081] The humidity influence coefficient W_hum corresponding to each grid node in the corresponding time window is extracted;

[0082] The grid nodes are sorted in descending order according to the numerical value of the humidity influence coefficient W_hum, to obtain a priority sequence of the grid nodes;

[0083] The humidification amount adjustment value W_tz corresponding to each grid node is calculated by using a weighted distribution algorithm according to the numerical value of the humidity influence coefficient W_hum, and the weight value in the weighted distribution algorithm is determined by the ratio of the reciprocal of the priority sequence number to the total number of grid nodes.

[0084] In an embodiment, if the fluctuation amplitude value B exceeds the second preset threshold T2, the humidity influence coefficient W_hum of each grid node is extracted. Assuming that there are 100 nodes in the grid, after extracting the W_hum of all nodes, a set of coefficients can be obtained, such as 0.8, 0.6, 0.9, etc. Specifically, the grid nodes are sorted in descending order according to the numerical value of W_hum to form a priority sequence. For example, the nodes with W_hum of 0.9, 0.8, and 0.6 are ranked first, second, and third, respectively, and the priority sequence is node A, node B, and node C. This sorting method ensures that the control resources are preferentially allocated to the nodes with greater humidity influence. For example, the W_hum of node A is 0.9, indicating that the humidity change has the greatest impact on environmental stability, and the working state of the humidification equipment should be adjusted first. The humidification amount adjustment value W_tz of each grid node is calculated by using a weighted distribution algorithm, and the weight value is determined by the ratio of the reciprocal of the priority sequence number to the total number of nodes. Assuming that the priority sequence number of node A is 1 and the total number of nodes is 100, the weight of node A is 1 / 1 ÷ 100 = 0.01; the priority sequence number of node B is 2, and the weight of node B is 1 / 2 ÷ 100 = 0.005.

[0085] Then, the humidity influence coefficient W_hum is multiplied by the corresponding weight to obtain the corresponding humidification amount adjustment value W_tz.

[0086] Optionally, in the step S4, the humidification amount adjustment value W_tz of each grid node is obtained, and the gas concentration measurement data Q_nd of the corresponding time period is extracted from the initial data set;

[0087] The humidification amount adjustment value W_tz and the gas concentration value Q_nd are aligned according to the grid nodes to construct a double-input parameter matrix with a row number equal to the total number of grids and a column number equal to 2;

[0088] For the humidification amount adjustment value W_tz in the first column of the double-input parameter matrix, the humidification amount range is set, and the humidification amount range is divided into three humidification amount intervals of low, medium, and high;

[0089] For the gas concentration value Q_nd in the second column of the double-input parameter matrix, a gas concentration range is set, and the gas concentration range is divided into three gas concentration intervals of thin, moderate and thick;

[0090] A rule base containing 9 rules is established according to the three humidification amount intervals and the three gas concentration intervals, and the rule format is that if the humidification amount adjustment value belongs to the humidification amount interval A and the gas concentration belongs to the gas concentration interval B, then the output heating power is C and the ventilation rate set value is D.

[0091] After inputting the humidification amount adjustment value W_tz and the gas concentration measurement data Q_nd corresponding to the grid node, the corresponding heating power and ventilation rate set value are output from the rule base.

[0092] For example, in a 10x10 grid baking environment, the construction of the double-input parameter matrix aligns the humidification amount adjustment value W_tz and the gas concentration value Q_nd by grid node, forming a 100-row 2-column matrix, which facilitates subsequent fuzzy reasoning processing. This matrix structure ensures data space consistency and provides a data basis for multi-parameter joint control.

[0093] For the humidification adjustment value W_tz, in this embodiment, the humidification amount range is 0 to 100%, if the input value is in the low interval (0-33%), the output humidification amount interval A is the low humidification amount interval, if the input value is in the middle interval (34%-66%), the output humidification amount interval A is the middle humidification amount interval, and if the input value is in the high interval (67%-100%), the output humidification amount interval A is the high humidification amount interval.

[0094] For the gas concentration value Q_nd, in this embodiment, the gas concentration range is 0 to 500ppm, if the input value is in the thin interval (0-167ppm), the output gas concentration interval B is the thin gas concentration interval, if the input value is in the moderate interval (168-333ppm), the output gas concentration interval B is the moderate gas concentration interval, and if the input value is in the thick interval (334-500ppm), the output gas concentration interval B is the thick gas concentration interval.

[0095] In this embodiment, one interval is selected from the three humidification amount intervals and one interval is selected from the three gas concentration intervals each time, and 9 different combinations can be obtained, and 9 different combinations all correspond to a preset heating power C and ventilation rate set value D. According to the 9 pre-established rule base, if the humidification adjustment value W_tz is the humidification amount interval and the gas concentration Q_nd is the gas concentration interval, the output is the corresponding heating power and ventilation rate set value.

[0096] It is worth mentioning that in the step S5, the heating power value is split into a power distribution array composed of operating powers of multiple heating elements according to the heating elements in the pre-stored heating element number mapping table, wherein the operating powers in the power distribution array are arranged in order of the heating element numbers in the heating element number mapping table.

[0097] According to the ventilation rate setting value, a pre-established ventilation rate-fan speed correspondence table is queried to obtain a corresponding fan speed value.

[0098] In a possible implementation, the heating power and the ventilation rate are, for example, 60W and 1.5m / s respectively for a grid node in a certain baking environment. After obtaining these data, the system needs to convert them into executable instructions of the device to ensure accurate regulation and control. For example, the heating power needs to be distributed to multiple heating elements, and the ventilation rate setting value needs to be converted into a fan speed.

[0099] For the distribution of the heating power value, the pre-stored heating element number mapping table plays a key role. Assuming that there are 5 heating elements in the baking environment, numbered H1 to H5, the mapping table records the maximum power and position information of each element. The heating power value 60W needs to be split into an initial power distribution array of each element. In one embodiment, the system assigns an operating power to each heating element according to the preset rated weight of each element according to the element position and power capacity. For example, H1 and H2 are close to the baking center, and are assigned higher weights of 25W and 20W respectively; H3 to H5 are farther away, and are assigned 8W, 4W and 3W respectively, forming the initial array [25, 20, 8, 4, 3]. In this embodiment, this distribution takes into account the heat distribution requirements of the baking area, ensuring uniform temperature in the central area. Each heating element operates according to its corresponding operating power.

[0100] For the processing of the ventilation rate setting value, the system queries the ventilation rate-fan speed correspondence table to convert the ventilation rate setting value of 1.5m / s into a specific speed. Assuming that the correspondence table shows that 1.5m / s corresponds to 3000rpm, the fan operates accordingly.

[0101] The embodiments of the present application are described in detail above with reference to the accompanying drawings, but the present application is not limited to the described embodiments. For those skilled in the art, various changes, modifications, replacements and variations of the embodiments can be made without departing from the principles and spirits of the present application, and still fall within the protection scope of the present application.

Claims

1. A smart baking method based on feedback adjustment of internal environment parameters of a baking oven, characterized in that, The method comprises the following steps: S1: Collecting a set of roasting environment data from a sensor group, including a temperature distribution matrix, a humidity change sequence and a gas concentration value, and storing it as an initial data set; S2: Generating a smoothed temperature state estimate from the temperature distribution matrix in the initial data set, calculating a temperature uniformity index based on the smoothed temperature state estimate, and generating a heating power ratio based on the smoothed temperature state estimate if the temperature uniformity index exceeds a first preset threshold; S3: Calculating a humidity influence coefficient based on the heating power ratio and the humidity change sequence in the initial data set, and generating a humidification amount adjustment value based on the humidity influence coefficient if the humidity change sequence in the initial data set fluctuates beyond a second preset threshold; S4: Processing the humidification amount adjustment value and the gas concentration value in the initial data set to generate heating power and ventilation rate set values; S5: Generating a set of adjustment parameters including the power of each heating element and the fan speed based on the heating power and ventilation rate set values.

2. The intelligent baking method based on feedback adjustment of internal environmental parameters of a baking oven according to claim 1, characterized in that: In the step S2, a Kalman filter algorithm is applied to the temperature distribution matrix in the initial data set to generate a smoothed temperature state estimate, and the average of the variances of all smoothed temperature state estimates is calculated as a temperature uniformity index.

3. The intelligent baking method based on feedback adjustment of internal environmental parameters of a baking oven according to claim 2, characterized in that: In the step S2, the step of applying a Kalman filter algorithm to the temperature distribution matrix in the initial data set comprises: Performing mean filtering processing on the temperature distribution matrix T_origin using a 3x3 convolution kernel to obtain a smoothed temperature matrix T_smooth; Taking each element of the smoothed temperature matrix T_smooth as an observation value to construct an observation vector Z_k of the Kalman filter, setting the state transition matrix F of the Kalman filter as an identity matrix, and setting the observation matrix H of the Kalman filter as an identity matrix; Performing a prediction-update cycle of the Kalman filter according to the observation vector Z_k, the state transition matrix F and the observation matrix H to obtain a state estimate matrix X_estimate; Extracting node temperature values from the state estimate matrix X_estimate, and performing moving average filtering processing on the node temperature values with a window width of a preset value to obtain a smoothed state estimate matrix X_estimate, wherein the elements in the smoothed state estimate matrix X_estimate are smoothed temperature state estimates.

4. The intelligent baking method based on feedback adjustment of internal environmental parameters of a baking oven according to claim 3, characterized in that: In the step S2, the step of generating a heating power ratio based on the smoothed temperature state estimate comprises: Using a linear regression model previously established based on the relationship between historical temperature data and historical heating power, calculating the heating power ratio P_i corresponding to each grid node in the smoothed state estimate matrix X_estimate for the smoothed temperature state estimate.

5. The intelligent roasting method based on feedback adjustment of internal environmental parameters of a roasting oven according to claim 4, characterized in that: In the step S3, a temperature-humidity coupling relationship model is established based on the heating power ratio and the humidity change sequence in the initial data set to calculate a humidity influence coefficient.

6. The intelligent roasting method based on feedback adjustment of internal environmental parameters of a roasting oven according to claim 5, characterized in that: In the step S3, the step of calculating the humidity influence coefficient comprises: Calculate the variance of the humidity change sequence H_seq of each grid node in the time dimension, if the variance is greater than a preset threshold T_hum, use a moving average filter with a window size of 5 to process the humidity change sequence H_seq, and obtain the smoothed humidity change sequence H_seq_smooth; Map the smoothed humidity change sequence H_seq_smooth to the humidity distribution vector V_hum corresponding to each grid node according to the spatial coordinates of the grid node; Use a linear regression model with intercept to calculate the slope parameter as the coupling relationship coefficient C_th, with the smoothed temperature state estimator as the independent variable and the humidity distribution vector V_hum as the dependent variable; Through the coupling relationship coefficient C_th and the heating power ratio P_i corresponding to each grid node, calculate the humidity influence coefficient W_hum=C_th*P_i / sum(P_i) corresponding to each grid node, where sum(P_i) is the sum of the heating power ratios of all grid nodes.

7. The intelligent baking method based on feedback adjustment of internal environmental parameters of a baking oven according to claim 6, characterized in that: In the step S3, the humidity change sequence fluctuation amplitude calculation step includes: Obtain the humidity time sequence, which contains the humidity measurement values of multiple time windows; Use a numerical calculation method to calculate the standard deviation of the humidity measurement values in each time window, and obtain the corresponding fluctuation amplitude value B in each time window.

8. The intelligent roasting method based on feedback adjustment of internal environmental parameters of a roasting oven according to claim 7, characterized in that: In the step S3, the step of generating a humidification amount adjustment value based on the humidity influence coefficient includes: Extract the humidity influence coefficient W_hum corresponding to each grid node in the corresponding time window; Sort the grid nodes in descending order according to the numerical value of the humidity influence coefficient W_hum, and obtain the grid node priority sequence; According to the numerical value of the humidity influence coefficient W_hum, use a weighted allocation algorithm to calculate the humidification amount adjustment value W_tz corresponding to each grid node, and the weight value in the weighted allocation algorithm is determined by the ratio of the reciprocal of the priority sequence number to the total number of grid nodes.

9. The intelligent roasting method based on feedback adjustment of internal environmental parameters of a roasting oven according to claim 8, characterized in that: In the step S4, obtain the humidification amount adjustment value W_tz of each grid node, and extract the gas concentration measurement data Q_nd of the corresponding time period from the initial data set; Align the humidification amount adjustment value W_tz and the gas concentration value Q_nd by grid node to construct a double-input parameter matrix with row number equal to total number of grids and column number equal to 2; For the humidification amount adjustment value W_tz in the first column of the double-input parameter matrix, set the humidification amount range, and divide the humidification amount range into three humidification amount intervals: low, medium and high; For the gas concentration value Q_nd in the second column of the double-input parameter matrix, set the gas concentration range, and divide the gas concentration range into three gas concentration intervals: thin, moderate and thick; According to the three humidification amount intervals and the three gas concentration intervals, establish a rule base containing 9 rules, and the rule format is: if the humidification amount adjustment value belongs to the humidification amount interval A and the gas concentration belongs to the gas concentration interval B, then output the heating power as C and the ventilation rate set value as D; After inputting the humidification amount adjustment value W_tz and the gas concentration measurement data Q_nd corresponding to the grid node, output the corresponding heating power and ventilation rate set value from the rule base.

10. The intelligent roasting method based on feedback adjustment of internal environmental parameters of a roasting oven according to claim 9, characterized in that: In the step S5, the heating power value is split into a power distribution array composed of operating powers of multiple heating elements according to the heating elements in a pre-stored heating element number mapping table, wherein the operating powers in the power distribution array are arranged in sequence according to the heating element numbers in the heating element number mapping table. According to the ventilation rate setting value, a pre-established ventilation rate-fan rotating speed correspondence table is inquired to obtain a corresponding fan rotating speed value.

Images