Automatic microwave heating uniformity analysis method based on normal distribution and chemical labeling method

Abstract

The invention relates to the technical field of food processing and microwave heating, in particular to an automatic microwave heating uniformity analysis method based on normal distribution and a chemical labeling method. According to the method, the global visualization advantage of a chemical labeling method is combined with the objective quantification capability of mathematical statistics, reliable data support is provided for microwave process optimization and equipment design, and the method capable of automatically, rapidly and accurately carrying out quantitative evaluation on the microwave heating uniformity is provided.

Description

Technical Field

[0001] This invention relates to the field of food processing and microwave heating technology, specifically to an automatic analysis method for microwave heating uniformity based on normal distribution and chemical labeling. Background Technology

[0002] Microwave heating is widely used in the food industry due to its high efficiency and speed. However, its inherent problem of uneven heating has always been a key bottleneck restricting product quality and safety. Therefore, accurately assessing heating uniformity is a prerequisite for optimizing microwave equipment and processes, but existing technologies have significant shortcomings:

[0003] 1. Traditional temperature sensor methods (such as thermocouples and fiber optic sensors): can only obtain temperature data at a limited number of discrete points, cannot reflect the overall (volume) temperature distribution inside the heated material, and are cumbersome to operate and have interference field distribution.

[0004] 2. Visual thermal mapping methods (such as infrared thermal imagers and thermal paper): Although they can provide images of heating distribution in a two-dimensional plane, they are difficult to quantify and evaluate, and cannot obtain information within a three-dimensional volume, making them highly subjective.

[0005] 3. Pure computer simulation method: Although it can provide full-volume temperature data, its results are heavily dependent on the accuracy of the model and must be verified through expensive experiments, resulting in high computational resource consumption.

[0006] 4. Chemical labeling methods (such as those based on Maillard reaction color changes) are a mature technology that can intuitively display the heating status of the entire heated area (including the interior) through color changes, providing global and visualized information on the heating distribution. However, current analysis of chemical labeling results mostly relies on human observation or simple image comparison, lacking precise and objective quantitative evaluation methods.

[0007] Existing technologies, such as Chinese patent CN120379088A, "A Multi-Field Coupled Microwave De-icing Intelligent System and Energy Efficiency Dynamic Control Method," disclose steps for collecting temperature and mechanical data and synchronously labeling them; then, the labeled temperature and mechanical data are input into a feedforward-feedback composite control algorithm to obtain the ice thickness variation trend and compare it with a set interface temperature threshold or adhesion force threshold. This method combines temperature sensor methods and computer simulation methods, but it relies on ice thickness data and is not applicable to ordinary microwave heating of non-ice layers. Therefore, there is an urgent need in the field for a universal microwave heating uniformity assessment method that can fully utilize the global advantages of chemical labeling methods and achieve rapid, objective, and quantitative analysis. Summary of the Invention

[0008] The technical problem to be solved by this invention is that there are still many shortcomings in the current technology for accurately assessing heating uniformity. There is an urgent need for a universal microwave heating uniformity assessment method that can fully utilize the global advantages of chemical labeling and achieve rapid, objective, and quantitative analysis.

[0009] To address the problems of existing technologies, this invention combines the global visualization advantages of chemical labeling with the objective quantitative capabilities of mathematical statistics, providing reliable data support for microwave process optimization and equipment design, and offering a method for automatically, quickly, and accurately quantitatively evaluating the uniformity of microwave heating.

[0010] To achieve the above objectives, the present invention provides an automatic analysis method for microwave heating uniformity based on normal distribution and chemical labeling, comprising the following steps:

[0011] (1) Preparation of samples containing chemical markers;

[0012] (2) Perform microwave heating treatment: Place the sample in the microwave device to be evaluated for heating treatment, and remove the sample after the heating process is completed;

[0013] (3) Obtain volume heating data: Use a three-dimensional scanning device to scan the heated sample and obtain the color value (such as RGB or Lab value) of each pixel or voxel in the entire volume of the sample; according to the pre-established "color-temperature" calibration curve, convert the color value of each pixel into the corresponding temperature value, thereby obtaining the three-dimensional volume temperature distribution dataset of the sample.

[0014] (4) Automatic analysis of uniformity based on normal distribution: Import the three-dimensional volumetric temperature distribution data obtained in S3 into the data processing unit (such as a computer) and execute the following automatic analysis process:

[0015] (4-1) Data preparation: The entire sample is regarded as a collection of a large number of temperature units, i.e., corresponding voxels; the temperature of each unit and the volume it represents, i.e., the pixel volume, are used as the basic data.

[0016] (4-2) Goodness-of-fit test of normal distribution: The correlation coefficient test of the QQ plot is used to determine whether the volume temperature distribution of the whole sample follows a normal distribution;

[0017] (4-3) Quantitative assessment of homogeneity:

[0018] Scenario A (Distribution is accepted as normal): If the test in (4-2) passes, then calculate the two key parameters of the normal distribution: the volume-weighted mean μ and the volume-weighted standard deviation σ; the standard deviation σ can be directly used as the core indicator to measure the uniformity of heating: the smaller the σ value, the more concentrated the temperature distribution and the better the uniformity; at the same time, by using the properties of the normal distribution, the percentage of volume whose temperature falls within a specific interval (such as μ±σ, μ±2σ) can be quickly calculated, and the consistency of product quality can be intuitively evaluated;

[0019] Case B (Distribution does not follow a normal distribution): If the test fails, the volume percentage of the critical temperature range is calculated as the evaluation index.

[0020] (5) Generate analysis report: The data processing unit automatically generates analysis report, including: the conclusion of whether the temperature distribution follows a normal distribution, the uniformity index, i.e., σ value or volume percentage of key temperature range, temperature distribution histogram and fitted normal distribution curve, QQ plot and other visualization charts.

[0021] Furthermore, the chemical marker in step (1) is a Maillard reaction product, with the following formula: 10% whey protein powder, 1% gellan gum, 1% D-ribose, 1% L-lysine, 0.3% CaCl2, and 86.7% distilled water, which changes color upon heating. These chemical markers exhibit the characteristic of undergoing irreversible and quantifiable color changes with increasing temperature.

[0022] Furthermore, the three-dimensional scanning device in step (3) is a high-resolution stereo scanner or a device combining CT and color analysis system.

[0023] Furthermore, the steps for establishing the "color-temperature" calibration curve in step (3) are as follows: (3-1) Take a small amount of sample and heat it at different times and temperatures through experiments; (3-2) Record the time-temperature curve experienced by the sample; (3-3) Calculate the heat treatment degree F value; (3-4) Analyze the color generation concentration by liquid chromatography; (3-5) Establish a standard curve between the product concentration and the F value.

[0024] Further, step (4-2) specifically includes: sorting the volumetric temperature data by temperature value; calculating the cumulative volume percentage; comparing it with the theoretical quantile of the standard normal distribution and plotting a QQ plot; calculating the correlation coefficient r of the QQ plot; comparing the calculated correlation coefficient r with the critical value (e.g., 0.9838) at a given significance level (e.g., 0.05); if r is greater than the critical value, the hypothesis that the temperature distribution follows a normal distribution is accepted; otherwise, the hypothesis is rejected.

[0025] Furthermore, in step (4-3) scenario B, a "low temperature critical value T_cold" related to food safety and a "high temperature critical value T_hot" related to food quality are defined; the volume percentage of temperatures below T_cold (to assess the severity of cold spots) and the volume percentage of temperatures above T_hot (to assess the degree of overheating) are calculated respectively; the smaller these two percentage values ​​are, the better the heating uniformity.

[0026] The beneficial effects of this invention are as follows:

[0027] (1) Global and volumetric: By combining chemical labeling with three-dimensional scanning, the heating status of the entire sample (three-dimensional volume) was accurately quantified for the first time, overcoming the limitations of point measurement and two-dimensional images.

[0028] (2) Objective Quantification and Automation: By introducing statistical tests and parameter calculations based on normal distribution, the uniformity assessment is transformed from subjective "judgment by looking at the graph" to objective "data-driven", thus automating the analysis process and reducing human error.

[0029] (3) Scientific nature and profound insights: Based on the "central limit theorem", the innovative hypothesis that the temperature distribution of microwave heating may follow a normal distribution is proposed, and a rigorous verification method is provided. Once it conforms to a normal distribution, the entire complex temperature field can be concisely described by the two parameters μ and σ, which is rich in connotation and facilitates scientific comparison between different systems.

[0030] (4) High practicality: This method is applicable to quality inspection and process optimization of microwave ovens and industrial microwave equipment, and can also be used as the gold standard for verifying the accuracy of computer simulation models. It has a wide range of applications.

[0031] (5) High efficiency and economy: Compared with traditional methods that require a large number of sensors and repeated experiments, or pure simulation methods that consume a lot of computing resources, the present invention can obtain global data in one experiment, with fast analysis speed and lower overall cost. Attached Figure Description

[0032] Figure 1 : Overall flowchart of the method of the present invention.

[0033] Figure 2 : A schematic diagram of the volumetric temperature distribution generated after the sample is microwave-heated and three-dimensionally scanned.

[0034] Figure 3 Example comparison of volumetric temperature distribution histogram and fitted normal distribution curve.

[0035] Figure 4 An exemplary QQ plot is used to test the normality of the temperature distribution.

[0036] Figure 5 "Color-Temperature" calibration curve. Detailed Implementation

[0037] The following description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0038] The following embodiments can be understood as illustrating only a part of the structure or method of the present invention, or as a combination of embodiments explaining the broader structure or method of the present invention. Unless otherwise specified, all raw materials of the present invention are commercially available.

[0039] In the following examples, the frozen Antarctic krill meat used was purchased from Liaoning Ocean Fishery Co., Ltd.

[0040] Example 1:

[0041] To evaluate the heating uniformity of a new microwave oven, the following steps are included:

[0042] (1) Prepare an agarose gel sample of uniform size, which is uniformly mixed with M-starch chemical marker. 10% whey protein powder, 1% gellan gum, 1% D-ribose, 1% L-lysine, 0.3% CaCl2, 86.7% distilled water.

[0043] (2) Place the sample in the center of the turntable in the microwave oven and heat it at the rated power for a set time (e.g., 500W, 30s).

[0044] (3) After heating, immediately remove the sample and scan it using a high-precision three-dimensional color scanner to obtain the three-dimensional color data of the sample; through Figure 5 The established "color-temperature" calibration curve converts the color of each voxel into a temperature value, forming a three-dimensional matrix containing millions of temperature data points.

[0045] (4) Read the matrix using a data processing unit (such as a computer with dedicated analysis software installed). The software automatically executes the following process: sorting temperature data, calculating cumulative volume percentage, performing QQ plot analysis with the standard normal distribution and calculating the correlation coefficient. Based on the comparison results of the correlation coefficient and the critical value, the software automatically selects the appropriate uniformity index (calculates the standard deviation σ or the volume percentage of the key temperature range) and finally generates a graphical analysis report.

[0046] Furthermore, this invention is not only applicable to static microwave heating systems, but can also be applied to the uniformity evaluation of continuously passing microwave processing systems (such as MATS) through methods such as segmentation marking.

[0047] Example 2:

[0048] The simulation model in this embodiment was built using the commercial electromagnetic solver QuickWave 3D version 7.5 (QW3D, Warsaw, Poland). The coupled electromagnetic field and heat transfer equations were solved numerically using the FDTD method. The microwave heating section of the MATS system was used as the physical model for the simulation. This section was divided into numerous FDTD elements of varying sizes. The element sizes for air, water, and food were 4416, 444, and 441 mm³, respectively, resulting in a total of 106,773,103 elements in the model and 263,716 elements in the food sample. The simulation was run on an HP 2800 workstation equipped with dual X5680 processors (3.33 GHz) and 96 GB of memory. Continuous movement was simulated using 32 discrete movement steps. The simulation took 322.5 hours to complete. In the simulation results, each FDTD element has a temperature element. A total of 263,716 temperature points were analyzed.

[0049] Normal distribution and probability graph

[0050] The normal distribution (Gaussian distribution) is the most well-known and widely used of all distributions. Because it approximates many natural phenomena well, the normal distribution has become a reference standard for many probability problems. The famous bell curve can be easily characterized and described by two parameters: mean and standard deviation. Data derived from the normal distribution are always described using σ² or σ². The probability distribution of a normal distribution is as follows: Figure 1 As shown.

[0051] The normal distribution is symmetrical about the mean. The data are densely distributed around the mean, with 68.27%, 95.45%, and 99.73% of the data falling within 1, 2, and 3 standard deviations to the left and right of the mean, respectively.

[0052] Probability plots are graphical techniques in statistics used to compare two datasets, such as empirical observations and theoretical datasets. There are two basic types of probability plots: quantile-quantile plots (QQ plots) and probability-probability plots (PP plots). A QQ plot is essentially a scatter plot of the ordered observations of a sample against the corresponding quantiles of the hypothesized distribution. A PP plot is a scatter plot of the corresponding cumulative probabilities between the empirical cumulative distribution and the hypothesized cumulative distribution function. They are often used as informal methods to assess the goodness of fit of a hypothesized distribution. However, the correlation coefficient of the probability plot is frequently recommended for formal testing (Looney and Gulledge, 1985; Filliben, 1975; Ryan and Joiner, 1976). In this embodiment, the correlation coefficient of the QQ plot between the volume percentage distribution and the standard normal distribution is used as a formal test for normality. If the temperature volume percentage distribution for each FDTD cell is accepted as a normal distribution, a PP plot will be drawn against the estimated normal distribution to show the difference between the two. The steps for drawing a QQ plot between the volume temperature distribution and the normal distribution are as follows:

[0053] (1) Arrange the temperatures of each FDTD unit of the food in ascending order.

[0054] (2) Select the number of points in the QQ graph. In this invention, n=75 is selected, resulting in temperature intervals.

[0055] (3) Calculate the volume percentage for each point.

[0056] (4) Calculate the corresponding quantile points based on the cumulative distribution function of the standard normal distribution.

[0057] (5) Plot each pair (Ti, xi) and calculate the correlation coefficient r as defined by Johnson and Wichern (2007).

[0058] and These are the highest and lowest temperatures within the food sample, respectively. It is a temperature observation point. It is the total volume; Is the temperature equal to or lower than The volume. and These are the quantiles of the standard normal distribution and the mean of the observed temperature points, respectively. For n=75, at a significance level of 0.05, the critical value for the normality test of the correlation coefficient of the QQ plot is 0.9838. If <0.9838, the normality hypothesis is rejected (Johnson and Wichern, 2007).

[0059] If the temperature distribution is accepted as a normal distribution, the unbiased estimate of the mean and standard deviation, weighted by the volume of each FDTD cell, is calculated using the following formula:

[0060]

[0061]

[0062] in The temperature of an FDTD unit. This is the volume of the unit. This represents the total number of units. All calculations were performed using separate MATLAB scripts (Mathworks, Inc., MA, USA).

[0063] result:

[0064] (1) Normality assessment of the MATS system

[0065] The QQ plot of the temperature distribution of food samples processed by the MATS system versus the standard normal distribution is shown below. Figure 2 As shown in the figure, the power settings for each microwave heating cavity are 6.0, 4.7, 2.8, and 2.7 kW, respectively. The correlation coefficient of this figure is greater than 0.9838. Therefore, the normality assumption cannot be rejected.

[0066] If two sets of data originate from the same distribution, the PP plot will approximately lie on the line y=x. PP plots and cumulative plots are shown below. Figure 3 As shown, a good agreement is observed between the cumulative volume percentage of temperature and the estimated cumulative normal distribution. Scatter plots of these two cumulative distributions confirm a good fit, with slopes of 0.9954 and intercepts of 0.0091, very close to the straight line y=x.

[0067] (2) Temperature distribution in each microwave heating cavity

[0068] The temperature distribution within each microwave heating cavity was analyzed for comparison by setting the internal structure of all four cavities to be identical to that of each individual structure. All other parameters were the same as in the combined structure model. Results using different evaluation criteria are shown in Table 1.

[0069] Table 1 Comparison of heating uniformity of individual microwave heating cavities and combined cavities

[0070] <![CDATA[T max -T min [°C]]]> Mean [°C] Std [°C] <![CDATA[COV a (%)]]> <![CDATA[r b ]]> <![CDATA[Normality c ]]> Cavity1 37.62 126.32 9.39 7.43 0.9697 Rejected Cavity 2 27.83 125.63 5.93 4.72 0.9859 Not Rejected Cavity 3 22.86 127.06 4.48 3.53 0.9868 Not Rejected Cavity 4 53.30 124.88 12.52 10.03 0.9743 Rejected Combination 29.13 126.22 5.43 4.30 0.9931 Not Rejected

[0071] a: COV, Coefficient of Variation, COV = Standard Deviation / Mean

[0072] b: r, QQ graph correlation coefficient

[0073] c: The critical value at a significance level of 0.05 is 0.9838.

[0074] As shown in Table 1, the temperature distributions of the heating modes in cavities 1 and 4 are rejected as normally distributed. The averages of all these heating modes are close to each other, with the largest difference between cavities 4 and 3 being 1.73%. This means that all these heating modes absorb the same amount of energy. The MATS system exhibits stable energy transfer. The combined heating mode has the highest correlation coefficient with the normal distribution. Cavity 3 has the best heating uniformity and can be selected as an option for further improvement. Heating modes that follow a normal distribution (cavities 2, 3, and the combination) have relatively low COV and temperature difference. However, there is a conflict in the evaluation of the cavity 2 and combined models. The temperature difference of cavity 2 is lower than that of the combined model (27.83 < 29.13), but the COV is higher (4.72 > 4.30). This conflict indicates the limitations of these criteria, as they were developed without considering the volumetric temperature distribution. Since both are proven to be normally distributed, the mean and standard deviation should be used as criteria for evaluating heating uniformity. Although the combined model has a higher temperature difference, its temperature distribution uniformity is better than that of cavity 2.

[0075] The cumulative volume percentage distribution of cavities 1 and 4 is as follows: Figure 4 As shown, it clearly displays the volumetric temperature distribution of heating modes rejected by the normal distribution assumption. Critical points can be developed as a standard for evaluating heating modes from different distribution sources. In food heat processing, cold and hot zones are highly valued and are related to food safety and quality, respectively. In the MTAS system, the microwave segment's task is to rapidly raise the food temperature. The heat preservation segment plays a role in inactivating microorganisms and ensuring food safety. Two critical temperature points can be set ( , This is used to assess the volume percentage of low and high temperature levels. These two volume percentage values ​​are good standards for evaluating the heating uniformity of temperature distribution derived from different distribution models.

[0076] in conclusion:

[0077] The MATS system has four microwave heating cavities with different internal structures, which are combined to form a relatively uniform heating pattern. A normal distribution is used to fit the volumetric temperature distribution of the combined heating pattern, and the goodness of fit is evaluated using the correlation coefficient of the QQ plot. The results show that the temperature distribution of food processed by the MATS system follows a normal distribution. The same analysis was performed on the heating pattern of each heating cavity in the MATS system. The heating patterns of cavities 2 and 3 are accepted by the normal distribution assumption, but those of cavities 1 and 4 are not. The mean and standard deviation of the volumetric temperature distribution can be used to evaluate the uniformity of heating patterns that follow the same distribution. Current standards for evaluating heating uniformity have some limitations because they do not consider the volumetric temperature distribution and its distribution model. This invention develops a new standard involving the volume percentage of low and high temperature levels for comparing the uniformity of heating patterns derived from different distribution models. However, for further development of heat treatment models, designing a normally distributed heating pattern is more ideal.

[0078] All aspects, embodiments, and features of this invention should be considered illustrative in all respects and not limiting of the invention; the scope of the invention is defined only by the claims. Other embodiments, modifications, and uses will become apparent to those skilled in the art without departing from the spirit and scope of the invention as claimed.

[0079] In the preparation method of this invention, the order of the steps is not limited to the listed order. For those skilled in the art, variations in the order of the steps without creative effort are also within the scope of protection of this invention. Furthermore, two or more steps or actions can be performed simultaneously.

[0080] Finally, it should be noted that the specific embodiments described herein are merely illustrative examples of the invention and are not intended to limit the implementation of the invention. Those skilled in the art can make various modifications or additions to the described specific embodiments or use similar methods to replace them; it is neither necessary nor possible to exemplify all embodiments here. However, these obvious variations or modifications derived from the essential spirit of the invention still fall within the scope of protection of the invention, and interpreting them as any additional limitation would contradict the spirit of the invention.

Claims

1. An automatic analysis method for microwave heating uniformity based on normal distribution and chemical labeling, characterized in that... Includes the following steps: (1) Preparation of samples containing chemical markers; (2) Perform microwave heating treatment: Place the sample in the microwave device to be evaluated for heating treatment, and remove the sample after the heating process is completed; (3) Obtain volume heating data: Use a three-dimensional scanning device to scan the heated sample and obtain the color value of each pixel or voxel in the entire volume of the sample; according to the pre-established "color-temperature" calibration curve, convert the color value of each pixel into the corresponding temperature value, thereby obtaining the three-dimensional volume temperature distribution dataset of the sample. (4) Automatic analysis of uniformity based on normal distribution: Import the three-dimensional volumetric temperature distribution data obtained in S3 into the data processing unit and execute the following automatic analysis process: (4-1) Data preparation: The entire sample is regarded as a collection of a large number of temperature units, i.e., corresponding voxels; the temperature of each unit and the volume it represents, i.e., the pixel volume, are used as the basic data.

2. (4-2) Goodness-of-fit test of normal distribution: The correlation coefficient test of the QQ plot is used to determine whether the volume temperature distribution of the whole sample follows a normal distribution; (4-3) Quantitative assessment of homogeneity: Case A is accepted as a normal distribution: If the test in (4-2) passes, then calculate the two key parameters of the normal distribution: the volume-weighted mean μ and the volume-weighted standard deviation σ; the standard deviation σ can be directly used as the core indicator to measure the uniformity of heating: the smaller the σ value, the more concentrated the temperature distribution and the better the uniformity; at the same time, by using the properties of the normal distribution, the percentage of volume where the temperature falls within a specific range can be quickly calculated, and the consistency of product quality can be intuitively evaluated; Case B - Distribution does not follow a normal distribution: If the test fails, the volume percentage of the critical temperature range is calculated as the evaluation index; (5) Generate analysis report: The data processing unit automatically generates analysis report, including: the conclusion of whether the temperature distribution follows a normal distribution, the uniformity index, i.e., σ value or volume percentage of key temperature range, temperature distribution histogram and fitted normal distribution curve, QQ plot and other visualization charts.

3. The method as described in claim 1, characterized in that: The chemical marker in step (1) is M-starch.

4. The method as described in claim 1, characterized in that: The three-dimensional scanning device in step (3) is a high-resolution stereo scanner or a device combining CT and color analysis system.

5. The method as described in claim 1, characterized in that: The steps for establishing the "color-temperature" calibration curve in step (3) are as follows: (3-1) Take a small amount of sample and heat it at different times and temperatures through experiments; (3-2) Record the time-temperature curve experienced by the sample; (3-3) Calculate the F value of the degree of heat treatment; (3-4) Analyze the color generation concentration by liquid chromatography; (3-5) Establish a standard curve between the concentration of the product and the F value.

6. The method as described in claim 1, characterized in that: The specific steps of step (4-2) include: sorting the volumetric temperature data by temperature value; calculating the cumulative volume percentage; comparing it with the theoretical quantile of the standard normal distribution and plotting a QQ plot; calculating the correlation coefficient r of the QQ plot; comparing the calculated correlation coefficient r with the critical value at a given significance level, and if r is greater than the critical value, accepting the hypothesis that the temperature distribution follows a normal distribution; otherwise, rejecting the hypothesis.

7. The method as described in claim 1, characterized in that: In step (4-3) scenario B, define the "low temperature critical value T_cold" related to food safety and the "high temperature critical value T_hot" related to food quality; calculate the volume percentage of the temperature below T_cold and the volume percentage of the temperature above T_hot respectively; the smaller these two percentage values ​​are, the better the heating uniformity.

Images