A method for monitoring state data during the preparation process of aerogel thermal insulation material

Abstract

The present invention relates to the field of data processing technology, and more specifically, the present invention relates to a method for monitoring the state data of the preparation process of aerogel thermal insulation materials, including: obtaining a temperature sequence in the preparation process of aerogel thermal insulation materials; constructing a coordinate system of a preset local range of each data point in the temperature sequence, and performing curve fitting on the data points within the preset local range of each data point; determining the degree of noise performance of each data point according to the difference between the fitting value and the temperature value of the data point within the preset local range of each data point; obtaining the final neighborhood range of each data point; obtaining the preferred degree of the neighborhood range of each data point; obtaining the local distance outlier factor of each data point, so as to complete the state data monitoring of the preparation process of aerogel thermal insulation materials. The present invention analyzes the data distribution characteristics within the local range of the data points of the temperature sequence to adaptively use the neighborhood range of the LDOF algorithm to better adapt to the temperature changes in the preparation process.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and more specifically, to a method for monitoring state data during the preparation process of an aerogel thermal insulation material. Background Art

[0002] Aerogel is a low-density three-dimensional nanoporous solid material composed of nano-colloidal particles or polymer molecules that aggregate into a nano-porous network structure with gaseous media in the pores. Aerogel has physical properties such as low density, high specific surface area, high porosity, low dielectric constant, and low thermal conductivity due to its three-dimensional nano-network porous structure. It has broad application prospects in thermal insulation, fire retardant, sound insulation and noise reduction, optics, electricity and other fields.

[0003] In the preparation process of aerogel insulation materials, temperature data is not only related to the performance and quality of the product, but also directly affects the optimization of the preparation process and the commercial application of the product. Therefore, the monitoring of temperature data should be highly valued during the preparation process. In order to ensure the accuracy and reliability of the data, the collected temperature data needs to be denoised. The Local Density Outlier Factor (LDOF) algorithm is an effective method for identifying noise points in time series data. The algorithm identifies noise points in temperature data by calculating the local distance outlier factor of the data point, thereby denoising the temperature data and accurately monitoring the temperature data.

[0004] The patent document with the announcement number CN109981630B discloses an intrusion detection method and system based on the chi-square test and the LDOF algorithm. Among them, an intrusion detection method based on the chi-square test and the LDOF algorithm includes: using the chi-square test algorithm to perform feature selection on the original intrusion detection data set; normalizing the intrusion detection data set after feature selection; using the LDOF algorithm to find the normalized intrusion detection data set. The largest LDOF factor, The LDOF factor corresponds to data objects are judged as abnormal data; among them, Is a positive integer.

[0005] However, the above patent document is not aimed at the preparation of aerogel thermal insulation materials, and does not solve the problem that the existing LDOF algorithm uses a fixed neighborhood size to calculate the local distance outlier factor when analyzing data points. However, since the preparation process of aerogel thermal insulation materials includes different stages such as heating, cooling, reaction, etc., that is, the distribution changes of temperature data are different, if a fixed neighborhood size is adopted, the best detection effect may not be achieved, resulting in inaccurate identification of noise data points and affecting the denoising effect. Summary of the invention

[0006] In order to solve the problem that the existing LDOF algorithm uses a fixed neighborhood size for local distance outlier factor when analyzing data points, resulting in inaccurate identification of subsequent noise data points and affecting the denoising effect, the present invention proposes a method for monitoring state data of aerogel thermal insulation material preparation process, the method comprising the following steps: obtaining a temperature sequence in the aerogel thermal insulation material preparation process; constructing a coordinate system of a preset local range of each data point in the temperature sequence, and performing curve fitting on the data points within the preset local range of each data point; determining the noise performance degree of each data point according to the difference between the fitting value and the temperature value of the data point within the preset local range of each data point, and the noise performance degree satisfies , For the The noise level of each data point is For the The number of data points in the local range preset by data points, For the The local range of the preset data points The fitted value of the data points, Indicates The local range of the preset data points temperature value of each data point; iteratively obtain the preferred degree of the neighborhood range of each data point, and use the neighborhood range corresponding to the preferred degree obtained in the previous round of iteration when the iteration terminates as the final neighborhood range of each data point. In the process of each round of iteration of each data point: sort the data points in the neighborhood range of each data point from large to small according to the noise performance degree, and divide them into two categories in turn, and use the difference in the mean of the noise performance degree of each data point in the two categories as the inter-class gap, and determine the sub-preferred degree of the neighborhood range of each data point at each division according to the inter-class gap and the variance of the noise performance degree of the data points in the two categories. The sub-preferred degree satisfies , For the During the iteration process The neighborhood of the data point The suboptimal degree of the second split, For the During the iteration process The neighborhood of the data point The inter-class gap between the two categories at the second segmentation, , Respectively During the iteration process The neighborhood of the data point The average of the noise performance of one category and another category in the split, , Respectively During the iteration process The neighborhood of the data point The variance of the noise representation of one class and another class at the time of segmentation, is a normalized function; the maximum value of the sub-optimal degree is taken as the preferred degree of the neighborhood range of each data point obtained in this round of iteration; according to the final neighborhood range of each data point, the local distance outlier factor of each data point is obtained by the LDOF algorithm to complete the state data monitoring of the aerogel thermal insulation material preparation process.

[0007] The beneficial effects are: dynamically adjusting the neighborhood range according to the noise performance of the data point can better adapt to different data distribution characteristics and enhance the robustness of the LDOF algorithm; by iteratively obtaining the preferred degree of the neighborhood range of each data point and determining the optimal final neighborhood range, the LDOF algorithm can more accurately reflect the local density outlier factor of the data point, improving the accuracy of subsequent noise data point identification; by dynamically adjusting the neighborhood range instead of using a fixed-size neighborhood range, the denoising effect is improved, thereby improving the accuracy of subsequent aerogel insulation material preparation process state data monitoring. Among them, the noise performance degree is quantified by calculating the difference between the fitted value and the actual temperature value, and by averaging the differences in the local range, the noise performance degree is made more representative, improving the robustness and adaptability of the algorithm, and ultimately improving the accuracy of monitoring.

[0008] Furthermore, the abscissa of the coordinate system is time, and the ordinate is temperature.

[0009] Furthermore, the curve fitting is performed by least squares method.

[0010] The beneficial effect is that the least squares method finds the best fitting curve by minimizing the residual sum of squares, thereby providing a more accurate model for a given data set. At the same time, the least squares principle is simple to use and has strong adaptability.

[0011] Furthermore, the data points within the neighborhood of each data point are sorted from large to small according to the noise performance degree, and are divided into two categories in sequence, including: The first to the second split data points into one category, and the remaining data points into another category, until the other category contains only one data point, the segmentation ends; The value range is to , is the number of data points in the neighborhood.

[0012] The beneficial effects are: starting from the first data point, the segmentation is gradually carried out, covering various possible segmentation situations, comprehensively analyzing the potential structure of the data, adapting to different data distributions, and obtaining the segmentation method that best reflects the characteristics of the data.

[0013] Furthermore, in the process of each round of iteration for each data point, it also includes: in response to the preference degree obtained in the current round of iteration for the data point being not less than the preference degree obtained in the previous round of iteration, a number of neighboring data points outside the neighborhood range of the data point are added to the neighborhood range of the data point, and the next round of iteration for the data point is started; in response to the preference degree obtained in the current round of iteration for the data point being less than the preference degree obtained in the previous round of iteration, the iteration of the data point is terminated.

[0014] Furthermore, the method for completing the status data monitoring of the aerogel thermal insulation material preparation process includes: in response to a local distance outlier factor being greater than a preset noise threshold, identifying the data point as a noise data point, and obtaining a denoised temperature value for the temperature value of the noise data point using a linear interpolation method; in response to a temperature value exceeding a preset temperature range, identifying the aerogel thermal insulation material preparation as abnormal, and issuing an early warning prompt, thereby completing the status data monitoring of the aerogel thermal insulation material preparation process.

[0015] The present invention has the following beneficial effects:

[0016] By analyzing the data distribution characteristics in the local range of the data points of the temperature sequence during the preparation of aerogel thermal insulation materials, the neighborhood range of the LDOF algorithm is adaptively used to better adapt to the temperature changes in the preparation process of aerogel thermal insulation materials. The specific situation of each data point can be optimized, which significantly improves the accuracy and efficiency of the subsequent LDOF algorithm in identifying the noise data points of the temperature sequence, thereby improving the accuracy of the status data monitoring of the aerogel thermal insulation material preparation process. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0018] Figure 1 The present invention is a flowchart of a method for monitoring state data during the preparation of an aerogel thermal insulation material according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0020] Because the preparation process of aerogel thermal insulation materials includes different stages such as heating, cooling, reaction, etc., that is, the distribution changes of temperature data are different. When using the LDOF algorithm to identify noise points, the fixed neighborhood method in the traditional algorithm may not be able to adapt to the changes in temperature data, resulting in inaccurate noise point identification and affecting the denoising effect. Therefore, the present invention analyzes the data distribution characteristics within the local range of the temperature data point and adapts the neighborhood range of the LDOF algorithm to accurately identify the noise points using the LDOF algorithm with an adaptive neighborhood, thereby improving the subsequent denoising effect of the temperature data and thus improving the monitoring efficiency.

[0021] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0022] See also Figure 1 , which shows a flow chart of the steps of a method for monitoring state data of an aerogel thermal insulation material preparation process provided by an embodiment of the present invention, the method comprising the following steps:

[0023] S1: Obtain the temperature sequence during the preparation of aerogel insulation materials.

[0024] The temperature data during the preparation of aerogel thermal insulation materials is collected in real time by temperature sensors to obtain a temperature sequence.

[0025] Implementers can set the acquisition frequency according to specific implementation conditions, for example, 1 Hz.

[0026] S2: constructing a coordinate system of a preset local range of each data point in the temperature sequence, and performing curve fitting on the data points within the preset local range of each data point.

[0027] It should be noted that the regularity of data points within their local range can be represented by the fitting error when curve fitting is performed on the data points within the local range. The larger the fitting error, the smaller the regularity. This is because the more regular the temperature sequence is when curve fitting is performed, the more accurately the internal structure of the data can be captured, making the fitting error smaller. On the contrary, if the temperature sequence is irregular, the fitting error will be larger.

[0028] Implementers can set the size of the local range according to the specific implementation situation. For example, with the current data point as the center, 10 data points are taken on each side. If there are less than 10 data points on either side, they are supplemented on the other side, and finally 21 data points are reached as the local range of the current data point.

[0029] Specifically, the abscissa of the coordinate system is time, and the ordinate is temperature.

[0030] Specifically, the curve fitting is performed by least square method.

[0031] S3: Determine the noise performance level of each data point according to the difference between the fitting value and the temperature value of the data point within the local range preset for each data point.

[0032] It should be noted that in the preparation process of aerogel thermal insulation materials, problems such as electronic noise and environmental interference may occur, resulting in noise in the collected temperature data sequence. Such noise will cause irregular fluctuations or deviations in the temperature data. Therefore, the regularity of the temperature data points within their local range can be used to calculate the degree of noise performance, where the lower the regularity, the higher the degree of noise performance.

[0033] The noise performance level satisfies the following relationship:

[0034] ;

[0035] In the formula, For the The noise level of each data point is For the The number of data points in the local range preset by data points, For the The local range of the preset data points The fitted value of the data points, Indicates The local range of the preset data points The temperature value of a data point.

[0036] in, Indicates The local range of the preset data points The fitting error of the data points is That is the The mean of the fitting errors of all data points within the local range preset by the data points; because the mean of the fitting error is negatively correlated with the regularity, and the regularity is negatively correlated with the degree of noise performance, then the mean of the fitting error is positively correlated with the regularity, so it can be used Directly represent temperature data points The noise performance level.

[0037] S4: Obtain the final neighborhood range of each data point.

[0038] It should be noted that, because the LDOF algorithm is a ratio of an average distance, indicating the degree of deviation between a data point and its nearest neighbor, it also indicates the degree of outlier of a data point in its neighborhood. Therefore, in order to improve the accuracy of subsequent LDOF algorithm calculations, it should be ensured as much as possible that the data points with high noise performance and low noise performance in the neighborhood of the data point have obvious differences, so that the algorithm can more accurately calculate its local distance outlier factor. Therefore, the present invention determines the final neighborhood range by iterating the size of the neighborhood range of the temperature data point and calculating the preferred degree of each size neighborhood range during the iteration process.

[0039] The optimization degree of the neighborhood range of each data point is obtained iteratively, and the neighborhood range corresponding to the optimization degree obtained in the previous round of iteration when the iteration is terminated is used as the final neighborhood range of each data point.

[0040] S5: Obtain the degree of preference of the neighborhood range of each data point.

[0041] It should be noted that the preference of the data point neighborhood in the iteration process can be achieved by dividing the temperature data points in the neighborhood into two categories according to their noise performance levels, represented by the inter-class gap and the intra-class gap. The larger the inter-class gap and the smaller the intra-class variance, the more obvious the difference between the data points with high noise performance and the data points with low noise performance in the neighborhood, and the higher the preference level.

[0042] During each round of iteration of each data point, the data points in the neighborhood of each data point are sorted from large to small according to the noise performance degree, and are divided into two categories in turn. The difference in the mean of the noise performance degree of each data point in the two categories is taken as the inter-class gap. According to the inter-class gap and the variance of the noise performance degree of the data points in the two categories, the sub-preferred degree of the neighborhood of each data point at each division is determined; the maximum value of the sub-preferred degrees is taken as the preferred degree of the neighborhood of each data point obtained in this round of iteration.

[0043] Specifically, the data points within the neighborhood of each data point are sorted from large to small according to the noise performance degree, and are divided into two categories in sequence, including:

[0044] The first split is to classify the first data point into one category and the remaining data points into another category. The second split is to classify the first to second data points into one category and the remaining data points into another category. And so on. The first to the second split data points into one category, and the remaining data points into another category, until the other category contains only one data point, the segmentation ends; The value range is to , is the number of data points in the neighborhood.

[0045] The suboptimal degree satisfies the following relationship:

[0046] ;

[0047] In the formula, For the During the iteration process The neighborhood of the data point The suboptimal degree of the second split, For the During the iteration process The neighborhood of the data point The inter-class gap between the two categories at the second segmentation, , Respectively During the iteration process The neighborhood of the data point The average of the noise performance of one category and another category in the split, , Respectively During the iteration process The neighborhood of the data point The variance of the noise representation of one class and another class at the time of segmentation, is the normalization function.

[0048] Implementers can set the size of the neighborhood range according to the specific implementation situation. For example, the 10 data points closest to the current data point are used as the neighborhood range of the current data point. If two data points are at the same distance from the current data point, the data point on the right is selected first.

[0049] Specifically, in each round of iteration of each data point, the process further includes:

[0050] In response to the preference degree of the data point obtained in the current iteration being not less than the preference degree obtained in the previous iteration, a number of neighboring data points outside the neighborhood range of the data point are added to the neighborhood range of the data point, and the next iteration of the data point is started;

[0051] In response to the preference level of a data point obtained in this round of iteration being less than the preference level obtained in the previous round of iteration, the iteration of the data point is terminated.

[0052] Implementers can set the number of neighboring data points, for example, 5, based on specific implementation circumstances.

[0053] S6: Obtain the local distance outlier factor of each data point to complete the state data monitoring of the aerogel thermal insulation material preparation process.

[0054] According to the final neighborhood range of each data point, the local distance outlier factor of each data point is obtained through the LDOF algorithm to complete the state data monitoring of the aerogel thermal insulation material preparation process.

[0055] Specifically, the state data monitoring of the aerogel thermal insulation material preparation process is completed, including:

[0056] In response to the local distance outlier factor being greater than a preset noise threshold, the data point is identified as a noise data point, and a denoised temperature value is obtained by using a linear interpolation method for the temperature value of the noise data point;

[0057] In response to the temperature value exceeding the preset temperature range, the aerogel thermal insulation material preparation is determined to be abnormal, and an early warning prompt is issued to complete the status data monitoring of the aerogel thermal insulation material preparation process.

[0058] Implementers can set the noise threshold and temperature range according to the specific implementation situation. For example, the noise threshold is 1; the temperature range for primary drying in the preparation process is 60~70℃, the temperature range for intermediate drying in the preparation process is 90~100℃, and the temperature range for final drying in the preparation process is 120~150℃.

[0059] The present invention denoises the data points of the temperature sequence during the preparation of the aerogel thermal insulation material, and then determines whether the preparation process is abnormal based on the accurate temperature value. Once the parameter deviates from the set temperature range, an early warning prompt is issued. Statistical analysis can also be performed based on the denoised temperature data to evaluate the stability of the indicator process and the product quality.

[0060] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for monitoring state data during the preparation process of an aerogel thermal insulation material, characterized in that: include: The temperature sequence in the process of preparing the aerogel thermal insulation material is obtained; a coordinate system of a preset local range of each data point in the temperature sequence is constructed, and a curve fitting is performed on the data points within the preset local range of each data point; the noise performance degree of each data point is determined according to the difference between the fitting value and the temperature value of the data point within the preset local range of each data point, and the noise performance degree satisfies , For the The noise level of each data point is For the The number of data points in the local range preset by data points, For the The local range of the preset data points The fitted value of the data points, Indicates The local range of the preset data points The temperature value of the data point; Iteratively obtain the preferred degree of the neighborhood range of each data point, and use the neighborhood range corresponding to the preferred degree obtained in the previous round of iteration when the iteration terminates as the final neighborhood range of each data point. In the process of each round of iteration of each data point: sort the data points in the neighborhood range of each data point from large to small according to the noise performance degree, and divide them into two categories in turn, and use the difference in the mean of the noise performance degree of each data point in the two categories as the inter-class gap, and determine the sub-preferred degree of the neighborhood range of each data point at each division according to the inter-class gap and the variance of the noise performance degree of the data points in the two categories. The sub-preferred degree satisfies , For the During the iteration process The neighborhood of the data point The suboptimal degree of the second split, For the During the iteration process The neighborhood of the data point The inter-class gap between the two categories at the second segmentation, , Respectively During the iteration process The neighborhood of the data point The average of the noise performance of one category and another category in the split, , Respectively During the iteration process The neighborhood of the data point The variance of the noise representation of one class and another class at the time of segmentation, is a normalized function; the maximum value of the sub-optimal degrees is used as the optimal degree of the neighborhood range of each data point obtained in this round of iteration; According to the final neighborhood range of each data point, the local distance outlier factor of each data point is obtained through the LDOF algorithm to complete the state data monitoring of the aerogel thermal insulation material preparation process.

2. The method for monitoring state data of aerogel thermal insulation material preparation process according to claim 1, characterized in that: The horizontal axis of the coordinate system is time, and the vertical axis is temperature.

3. The method for monitoring state data of aerogel thermal insulation material preparation process according to claim 1, characterized in that: The curve fitting is performed by least square method.

4. The method for monitoring state data of aerogel thermal insulation material preparation process according to claim 1, characterized in that: The data points within the neighborhood of each data point are sorted from large to small according to the noise performance degree, and divided into two categories in sequence, including: No. The first to the second split data points into one category, and the remaining data points into another category, until the other category contains only one data point, the segmentation ends; The value range is to , is the number of data points in the neighborhood.

5. The method for monitoring state data of aerogel thermal insulation material preparation process according to claim 1, characterized in that: The process of each round of iteration of each data point also includes: In response to the preference degree of the data point obtained in the current iteration being not less than the preference degree obtained in the previous iteration, a number of neighboring data points outside the neighborhood range of the data point are added to the neighborhood range of the data point, and the next iteration of the data point is started; In response to the preference level of a data point obtained in this round of iteration being less than the preference level obtained in the previous round of iteration, the iteration of the data point is terminated.

6. The method for monitoring state data of aerogel thermal insulation material preparation process according to claim 1, characterized in that: The method for completing the state data monitoring of the aerogel thermal insulation material preparation process includes: In response to the local distance outlier factor being greater than a preset noise threshold, the data point is identified as a noise data point, and a denoised temperature value is obtained by using a linear interpolation method for the temperature value of the noise data point; In response to the temperature value exceeding the preset temperature range, the aerogel thermal insulation material preparation is determined to be abnormal, and an early warning prompt is issued to complete the status data monitoring of the aerogel thermal insulation material preparation process.

Images