A method and system for analyzing data of a quartz plate production process

Abstract

The application relates to the technical field of quartz plate production, and discloses a quartz plate production process data analysis method and system, which comprises the following steps: acquiring cooling section multi-dimensional time sequence data and flatness values of a quartz plate workpiece; according to the cooling section multi-dimensional time sequence data, a characteristic parameter representing the actual cooling experience of the quartz plate workpiece is extracted, and a cooling thermal history feature vector is generated; based on the cooling thermal history feature vector, a cooling deviation index of the quartz plate workpiece is calculated, and the quartz plate workpiece is divided into different cooling state categories according to the cooling deviation index; based on the flatness values and the cooling state categories of the quartz plate workpiece, flatness statistical characteristics of the quartz plate workpieces in each cooling state category are calculated respectively, and the source of flatness abnormality is determined according to the differences between the flatness statistical characteristics; and according to the source of flatness abnormality, a production process diagnosis result of the quartz plate workpiece is generated, and the cooling process of the quartz plate is quantitatively evaluated by introducing the cooling section multi-dimensional time sequence data and constructing the cooling thermal history feature vector.

Description

Technical Field

[0001] This application relates to the field of quartz plate production technology, and more specifically, to a data analysis method and system for the quartz plate production process. Background Technology

[0002] In the quartz slab production field, to diagnose the stability of the sintering process, data analysis is typically performed on the correlation between the sintering temperature curve and the flatness of the cooled quartz slab. Current technology records flatness data in association with the sintering temperature parameters of that batch, and process engineers analyze this data to evaluate the correspondence between temperature parameters and flatness. However, in actual continuous production workshops, the cooling section often lacks precise closed-loop temperature control, relying instead on a relatively stable workshop environment and preset fan frequencies. When encountering diurnal temperature variations, seasonal temperature changes, or uncontrolled fluctuations in conveyor roller speed due to minor adjustments in the cycle time of preceding and following processes, the actual cooling rate of some slabs in the cooling section deviates from the normal level, resulting in different temperature gradient distributions within the semi-finished quartz slab, generating residual thermal stress that is not measured by conventional sensors. Because existing process records only include the set temperature and holding time of the sintering furnace, lacking variables related to the true thermal history of the cooling section, data analysis systems still forcibly match the same nominal sintering temperature parameter with the resulting discrete flatness detection values. This causes the originally concentrated data cloud to split, resulting in inexplicable wide dispersion or even bimodal patterns in the mapping relationship between process and flatness. Existing conventional analysis methods, because they fail to observe and isolate the hidden thermal stress state of the semi-finished product, cannot distinguish whether the fluctuation is due to fluctuations in the sintering process itself or unknown shifts during cooling. They are prone to mistakenly attributing flatness anomalies to aging of the sintering furnace temperature control components, leading to unnecessary equipment maintenance or sintering parameter adjustments. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this application provides a data analysis method and system for quartz plate production processes. This method and system can solve the technical problem in existing technologies where the lack of real thermal history variables in the cooling section makes it impossible to accurately identify the source of flatness anomalies, thereby interfering with the diagnosis of sintering processes.

[0004] In a first aspect, this application provides a data analysis method for the quartz plate production process, including: Obtain multi-dimensional time-series data of the cooling section and flatness values ​​of the quartz plate workpiece; Based on the multi-dimensional time-series data of the cooling section, feature parameters characterizing the actual cooling experience of the quartz plate workpiece are extracted to generate a cooling thermal history feature vector. Based on the cooling thermal history feature vector, the cooling deviation index of the quartz plate workpiece is calculated, and the quartz plate workpiece is classified into different cooling state categories according to the cooling deviation index. Based on the flatness value and cooling state category of the quartz plate workpiece, the flatness statistical characteristics of the quartz plate workpiece for each cooling state category are calculated, and the source of flatness anomaly is determined based on the difference in flatness statistical characteristics. Based on the source of flatness anomalies, generate production process diagnostic results for quartz plate workpieces.

[0005] Optionally, the steps for obtaining multi-dimensional time-series data of the cooling section and flatness values ​​of the quartz plate workpiece include: The ambient temperature sequence, fan operating frequency sequence, and actual cooling duration of the cooling section are obtained as multi-dimensional time-series data of the cooling section. Obtain the flatness test value as the flatness value; The acquired multidimensional time-series data of the cooling section and the flatness value are spatiotemporally correlated according to the physical flow sequence of the quartz plate workpiece on the production line.

[0006] Optionally, the step of extracting feature parameters characterizing the actual cooling experience of the quartz plate workpiece based on multi-dimensional time-series data of the cooling section and generating a cooling thermal history feature vector includes: Retrieve the theoretical cooling time from the preset process conditions; Calculate the difference between the actual cooling time and the theoretical cooling time, then divide by the theoretical cooling time to obtain the actual cooling time deviation rate; The equivalent cooling intensity of the cooling section is calculated based on the ambient temperature sequence and the fan operating frequency sequence. Based on the ambient temperature sequence and the fan operating frequency sequence, the variance of ambient temperature and the degree of abrupt change in fan frequency are calculated as characteristics of environmental fluctuations. The actual cooling time deviation rate, equivalent cooling intensity, and environmental fluctuation characteristics are combined and dimensionless to generate a cooling thermal history feature vector.

[0007] Optionally, the step of calculating the cooling deviation index of the quartz plate workpiece based on the cooling thermal history feature vector, and classifying the quartz plate workpiece into different cooling state categories according to the cooling deviation index, includes: Quartz plate sample workpieces produced under standard cooling conditions are selected from the historical production process database. A reference feature space is constructed based on the cooling thermal history feature vector of the quartz plate sample workpieces, and the mean vector of the reference feature space is calculated. The Mahalanobis distance between the cooling thermal history eigenvector and the mean vector of the quartz plate workpiece is calculated and used as the cooling offset index. When the cooling offset index is less than the first preset threshold, the quartz plate workpiece is marked as being in normal cooling state. When the cooling offset index is between the first preset threshold and the second preset threshold, the quartz plate workpiece is marked as having a slight cooling offset state; the second preset threshold is greater than the first preset threshold. When the cooling offset index is greater than the second preset threshold, the quartz plate workpiece is marked as having a severe cooling offset state.

[0008] Optionally, the step of calculating the flatness statistical characteristics of the quartz plate workpiece for each cooling state category based on the flatness value and cooling state category of the quartz plate workpiece includes: Calculate the mean and variance of flatness for quartz plate workpieces marked as normal cooling state, slightly cooled, and severely cooled. For quartz plate workpieces of each cooling state category, a one-way ANOVA was conducted with the cooling state category as the independent variable and the flatness value as the dependent variable to obtain the significance level test value, which was used as the statistical characteristic of flatness.

[0009] Optionally, the steps for determining the source of flatness anomalies based on differences in flatness statistical characteristics include: If the statistical characteristics of flatness are less than the preset difference threshold, the following judgment is performed: When the flatness mean and flatness variance of the quartz plate workpiece marked as normal cooling state are both within the corresponding preset normal range, and the flatness mean and flatness variance of the quartz plate workpiece marked as slight cooling deviation state and / or the quartz plate workpiece marked as severe cooling deviation state both exceed the corresponding preset normal range, it is determined that the flatness abnormality originates from the additional influence of the cooling section. When the flatness mean and flatness variance of the quartz plate workpiece marked as being in normal cooling state both exceed the corresponding preset normal range, and the flatness mean and flatness variance of the quartz plate workpiece marked as being in mild cooling deviation state and the quartz plate workpiece marked as being in severe cooling deviation state are both within the corresponding preset normal range, it is determined that the flatness abnormality originates from fluctuations in the sintering process itself. When the flatness mean and flatness variance of the quartz plate workpiece marked as being in normal cooling state both exceed the corresponding preset normal range, and the flatness mean and / or flatness variance of the quartz plate workpiece marked as being in mild cooling deviation state and the quartz plate workpiece marked as being in severe cooling deviation state both exceed the corresponding preset normal range, the flatness abnormality is determined to originate from a mixed abnormality. If the flatness statistical characteristics are not less than the preset difference threshold, the flatness abnormality is determined to originate from fluctuations in the sintering process itself.

[0010] Optionally, the steps for generating production process diagnostic results for quartz plate workpieces based on the source of flatness anomalies include: When it is determined that the flatness abnormality is caused by the additional influence of the cooling section, a production process diagnostic result including suggestions for troubleshooting the cooling section is generated. When it is determined that the flatness abnormality is caused by fluctuations in the sintering process itself, a production process diagnostic result including sintering process diagnostic alarms is generated. When the flatness anomaly is determined to originate from a mixed anomaly, a production process diagnostic result including bidirectional parallel troubleshooting suggestions is generated.

[0011] Optionally, when the anomaly is determined to originate from an additional effect of the cooling section, the method further includes: Retrieve the deviation compensation relationship between the cooling offset index and the flatness value; For quartz plate workpieces marked as slightly cooled and marked as severely cooled, their cooling offset index is substituted into the deviation compensation relationship to calculate the estimated flatness deviation caused by residual thermal stress. The flatness correction value is obtained by subtracting the flatness deviation estimate from the flatness values ​​of the quartz plate workpieces marked as slightly cooled and the quartz plate workpieces marked as heavily cooled. Generate production process diagnostic results that include flatness correction values ​​and cooling section troubleshooting suggestions.

[0012] Optionally, after the step of generating a production process diagnostic result for the quartz plate workpiece based on the source of the flatness anomaly, the following is included: Adjustment instructions are issued to the cooling section control unit and / or sintering furnace based on the production process diagnostic results.

[0013] Secondly, this application also discloses a quartz plate production process data analysis system for performing the aforementioned quartz plate production process data analysis method, comprising: The data acquisition module is used to acquire multi-dimensional time-series data of the cooling section and flatness values ​​of the quartz plate workpiece. The feature extraction module is used to extract feature parameters that characterize the actual cooling experience of the quartz plate workpiece based on the multi-dimensional time series data of the cooling section, and generate a cooling thermal history feature vector. The status assessment module is used to calculate the cooling deviation index of the quartz plate workpiece based on the cooling thermal history feature vector, and classify the quartz plate workpiece into different cooling status categories according to the cooling deviation index. The anomaly location module is used to calculate the flatness statistical characteristics of the quartz plate workpiece for each cooling state category based on the flatness value and cooling state category, and to determine the source of flatness anomaly based on the differences in flatness statistical characteristics. The diagnostic results generation module is used to generate diagnostic results for the production process of quartz plate workpieces based on the source of flatness anomalies.

[0014] This application acquires multi-dimensional time-series data of the cooling section of the quartz plate, constructs a feature vector representing the actual cooling experience, and combines it with flatness statistical features to determine the source of anomalies. This effectively solves the problem of process diagnosis deviation caused by ignoring the thermal history of the cooling section in the prior art. It can decouple environmental fluctuations during the cooling process from sintering process fluctuations, which not only improves the accuracy of flatness anomaly attribution, but also realizes refined management of the production process through deviation compensation and automatic control commands, significantly improving the stability of quartz plate production and product quality. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating a data analysis method for the quartz plate production process provided in an embodiment of this application.

[0016] Figure 2 This is a schematic diagram of the structure of a quartz plate production process data analysis system provided in an embodiment of this application.

[0017] Labeling Explanation: 210, Data Acquisition Module; 220, Feature Extraction Module; 230, Status Assessment Module; 240, Anomaly Localization Module; 250, Diagnostic Result Generation Module. Detailed Implementation

[0018] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0019] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0020] In the quartz slab production industry, after the green quartz slabs are sintered in the sintering furnace according to a preset temperature curve, they need to be conveyed to the cooling section for cooling, and then their surface flatness is scanned at an online inspection station. To diagnose the stability of the sintering process, data analysis is typically performed on the correlation between the sintering temperature curve and the flatness of the cooled quartz slabs. However, in actual continuous production workshops, the cooling section often lacks precise closed-loop temperature control. When encountering diurnal temperature differences, seasonal temperature changes, or uncontrolled fluctuations in the conveyor roller speed due to minor adjustments in the cycle time of preceding and following processes, the actual cooling rate of some slabs in the cooling section deviates from the normal level, resulting in different temperature gradient distributions within the semi-finished quartz slabs, thus generating residual thermal stress. This hidden residual thermal stress directly participates in the formation of the final flatness result. Existing conventional analysis methods, because they fail to observe and isolate the hidden thermal stress state of the semi-finished products, easily and incorrectly attribute flatness anomalies to the aging of the sintering furnace temperature control components, interfering with the accurate diagnosis of the sintering process.

[0021] To provide a clearer and more detailed explanation of the technical solutions in this application, the basic concepts involved are explained below.

[0022] Multidimensional time-series data for the cooling section refers to the environmental or equipment status data recorded over time as the quartz plate workpiece cools in the cooling section. The "multidimensional" aspect here does not refer to temperature alone, but rather to multiple observational dimensions that collectively influence the cooling result, such as the ambient temperature of the cooling section, the fan operating frequency, the actual cooling duration corresponding to the conveyor cycle time, and the passage times at the cooling section inlet and outlet. The "time-series" aspect means that these data are not isolated single-point values, but rather a data sequence formed by continuous recording in the order of acquisition.

[0023] Cooling thermal history feature vector refers to a numerical vector that can comprehensively reflect the real thermodynamic process experienced by the quartz plate workpiece throughout the cooling stage. It is a set of representative feature parameters that is compressed and refined from the original multi-dimensional time series data, so that subsequent classification and comparison no longer directly rely on the original sequences of varying lengths and complex fluctuations, but rely on a unified expression that can summarize the differences in cooling experience.

[0024] The cooling deviation index is a quantitative indicator used to measure the degree of difference between the actual cooling process experienced by a quartz plate workpiece and the standard or ideal cooling process. The larger the cooling deviation index, the more significant the difference between the actual cooling experience and the standard cooling experience of the workpiece; the smaller the cooling deviation index, the closer the workpiece is to the cooling state under normal process conditions.

[0025] Flatness statistical characteristics are characteristic values ​​obtained by statistically calculating the flatness values ​​of multiple quartz plate workpieces under the same category. They are used to reflect the overall distribution law of flatness under that category.

[0026] Firstly, see [the following] Figure 1 This application discloses a data analysis method for the quartz plate production process, including: 1. Obtain multi-dimensional time-series data of the cooling section and flatness values ​​of the quartz plate workpiece; 2. Based on the multi-dimensional time series data of the cooling section, extract the feature parameters that characterize the actual cooling experience of the quartz plate workpiece, and generate a cooling thermal history feature vector. 3. Based on the cooling thermal history feature vector, calculate the cooling deviation index of the quartz plate workpiece, and classify the quartz plate workpiece into different cooling state categories according to the cooling deviation index; 4. Based on the flatness value and cooling state category of the quartz plate workpiece, calculate the flatness statistical characteristics of the quartz plate workpiece for each cooling state category, and determine the source of flatness anomaly based on the differences in flatness statistical characteristics. 5. Generate production process diagnostic results for quartz plate workpieces based on the source of flatness anomalies.

[0027] Specifically, the process begins by acquiring multi-dimensional time-series data of the quartz plate workpiece in the cooling section, along with its final flatness value, from various data acquisition nodes on the production line. Data is jointly output by cooling section sensors, fan controller feedback interfaces, conveyor position detection elements, and flatness detection stations, and then uniformly received and stored by a host computer, manufacturing execution system, or data acquisition server. Next, features are extracted from this raw time-series data, transforming it into a cooling thermal history feature vector that characterizes the actual cooling experience of the workpiece. Subsequently, a quantified cooling deviation index is calculated using this feature vector, and the workpiece is categorized into different cooling state categories based on this index. After classification, flatness statistical characteristics are calculated for each cooling state category, combined with the workpiece's flatness value. By comparing the differences in these statistical characteristics, the true source of flatness anomalies can be logically inferred. Finally, based on the determined source of the anomaly, targeted production process diagnostic results are generated. Thus, quality anomalies, originally mixed in with sintering and cooling fluctuations, are separated into two distinct and traceable sources. The diagnostic conclusions no longer rely solely on experience but are based on workpiece-level cooling experience analysis.

[0028] In one possible example, the ambient temperature of the cooling section can be automatically output by a temperature acquisition device installed inside or on the side wall of the cooling section, the operating status of the fan can be automatically read by the feedback interface of the fan frequency converter or the communication interface of the control cabinet, the time when the workpiece enters and leaves the cooling section can be automatically recorded by the inlet and outlet position detection elements, and the flatness value is automatically uploaded to the data platform by the online flatness detection equipment.

[0029] More specifically, the arithmetic mean of multi-dimensional time-series data of the cooling section over the entire cooling cycle can be calculated, such as the average temperature and average wind speed. These averages can be directly combined as feature parameters characterizing the actual cooling experience, thereby generating a cooling thermal history feature vector. In practice, the original time-series data can be preprocessed before feature parameters are extracted. The preprocessing process may include: removing obviously erroneous collection points, such as null values, duplicate values, or outliers exceeding the normal range of the equipment caused by communication interruptions; rearranging the data of each dimension according to a unified time order; and time-aligning data from different sampling rhythms so that the temperature data and fan frequency data of the same workpiece in the same cooling stage can correspond to similar time segments. After preprocessing, feature parameters reflecting the average level, fluctuation level, and process rhythm are extracted separately. For example, in addition to the average temperature and average fan operating level, the dispersion of ambient temperature, the smoothness of fan operation changes, and the actual residence time of the workpiece in the cooling section can also be extracted. Finally, these feature parameters are arranged in a fixed order to form a cooling thermal history feature vector.

[0030] Specifically, a standard cooling thermal history feature vector can be pre-defined. Then, the Euclidean distance between the current quartz plate workpiece's cooling thermal history feature vector and this standard vector can be calculated, and this Euclidean distance can be used as the cooling deviation index. Subsequently, a fixed distance threshold is set. When the cooling deviation index is greater than the threshold, it is classified as an abnormal cooling state; otherwise, it is classified as a normal cooling state.

[0031] In some implementations, the median flatness values ​​of the quartz plate workpiece under normal cooling state and abnormal cooling state categories can be calculated separately as a statistical feature of flatness. If the median flatness under the abnormal cooling state category is significantly higher than that under the normal cooling state category, the flatness abnormality is determined to originate from the cooling section; if both are high and there is no significant difference, the flatness is determined to originate from the sintering process.

[0032] In a possible scenario, when an anomaly is determined to originate from the cooling section, a text prompt is automatically generated, such as "An anomaly exists in the cooling section; please check," and displayed on the monitoring screen as a production process diagnostic result. To make the diagnostic results more actionable, it is preferable to include supplementary information corresponding to the source of the anomaly in the text prompt, rather than simply providing a general prompt. Specifically, when an anomaly is determined to originate from the cooling section, the diagnostic result can also include the main basis for this determination, such as the flatness statistical characteristics of the abnormal cooling state category being higher than those of the normal cooling state category, and the cooling deviation index being consistently high, and further suggest priority inspection targets, such as environmental fluctuations in the cooling section, fan operation stability, or changes in conveyor cycle time. When an anomaly is determined to originate from the sintering process, the diagnostic result can indicate that the flatness statistical characteristics of each cooling state category are all high, and the differences after cooling classification are not significant, suggesting checking the execution of the sintering temperature curve. When a clear determination cannot be formed temporarily, the result can also be output as "The current sample is insufficient or the category differences are not sufficient," suggesting continued data collection and analysis. The output of the diagnostic results can be a text prompt on the monitoring screen, an alarm record written into the manufacturing execution system, a diagnostic entry generated in the daily report, or pushed to the process management terminal. By outputting the judgment conclusion and its basis together, subsequent investigation work can directly focus on the most likely abnormal links, thereby shortening the problem location time.

[0033] Through the above overall technical solution, it is possible to clearly distinguish whether the flatness abnormality is caused by fluctuations in the sintering process itself or by unknown deviations in the cooling process. This solves the problem of severe dispersion in the flatness detection values ​​after cooling corresponding to the same sintering temperature parameter, restores the interpretability of the process-quality mapping relationship, avoids incorrect attribution, and thus achieves accurate diagnosis of the stability of the sintering process.

[0034] Optionally, the steps for obtaining multi-dimensional time-series data of the cooling section and flatness values ​​of the quartz plate workpiece include: The ambient temperature sequence, fan operating frequency sequence, and actual cooling duration of the cooling section are obtained as multi-dimensional time-series data of the cooling section. Obtain the flatness test value as the flatness value; The acquired multidimensional time-series data of the cooling section and the flatness value are spatiotemporally correlated according to the physical flow sequence of the quartz plate workpiece on the production line.

[0035] The acquisition of the ambient temperature sequence of the cooling section refers to the real-time monitoring of air temperature changes inside or around the cooling section using sensors, and recording the data in chronological order. Specifically, devices such as thermocouples, thermistors, or infrared temperature sensors can be used to convert the collected temperature signals into digital signals and store them at preset time intervals, such as once per second or once per minute, to form a continuous temperature time series.

[0036] The fan operating frequency sequence refers to the recording of the fan's operating frequency at different points in time within the cooling section. The fan operating frequency directly affects the airflow and cooling intensity. Specifically, it can be achieved using the frequency output interface built into the frequency converter, or by indirectly measuring the fan speed through current and voltage sensors and converting it into frequency. These data are then recorded in chronological order to form a time series reflecting the fan's operating status.

[0037] The actual cooling time refers to the actual time it takes for a quartz plate workpiece to travel from entering the cooling section to leaving it. This can be achieved by installing inlet and outlet sensors (such as photoelectric sensors or proximity switches) on the production line. A timer is started when a workpiece passes through the inlet and stopped when it passes through the outlet, thus accurately recording the cooling time for each workpiece. This time data is then associated with a unique identifier for each workpiece.

[0038] The flatness test value refers to the data obtained by measuring the flatness of a cooled quartz plate workpiece using specialized testing equipment. Specifically, high-precision testing equipment such as laser displacement sensors, coordinate measuring machines, or machine vision systems can be used to scan and measure the undulations on the surface of the quartz plate, generating numerical values ​​reflecting its flatness, such as maximum height difference and local curvature.

[0039] The process of spatiotemporally correlating the acquired multidimensional time-series data of the cooling section and the flatness values ​​according to the physical flow sequence of the quartz plate workpieces on the production line involves establishing a one-to-one correspondence between various time-series data (ambient temperature sequence, fan operating frequency sequence, actual cooling duration) of a specific quartz plate workpiece in the cooling section and its final flatness detection value. This correlation ensures a clear correspondence between the cooling process data and the final quality data of each workpiece. Specifically, the workpiece ID or batch number on the production line can be used as a unique identifier to bind all relevant data under the same ID or batch number. For example, when a quartz plate workpiece enters the cooling section, its multidimensional time-series data for the cooling section begins to be recorded, and recording stops and is saved after the workpiece leaves the cooling section. Subsequently, after the flatness detection of the workpiece is completed, its flatness value will be matched and stored with the previously saved multidimensional time-series data of the cooling section through the workpiece ID. This spatiotemporal correlation is the foundation for subsequent data analysis, enabling direct linking of specific parameter changes during the cooling process to the final flatness result, thereby accurately diagnosing the source of flatness anomalies. This precise matching avoids data confusion and misjudgment, enabling subsequent feature extraction, state assessment, and anomaly localization to be based on accurate correspondences, greatly improving the reliability and effectiveness of diagnostic results.

[0040] Optionally, the step of extracting feature parameters characterizing the actual cooling experience of the quartz plate workpiece based on multi-dimensional time-series data of the cooling section and generating a cooling thermal history feature vector includes: Retrieve the theoretical cooling time from the preset process conditions; Calculate the difference between the actual cooling time and the theoretical cooling time, then divide by the theoretical cooling time to obtain the actual cooling time deviation rate; The equivalent cooling intensity of the cooling section is calculated based on the ambient temperature sequence and the fan operating frequency sequence. Based on the ambient temperature sequence and the fan operating frequency sequence, the variance of ambient temperature and the degree of abrupt change in fan frequency are calculated as characteristics of environmental fluctuations. The actual cooling time deviation rate, equivalent cooling intensity, and environmental fluctuation characteristics are combined and dimensionless to generate a cooling thermal history feature vector.

[0041] The theoretical cooling time, which is a preset process condition, refers to the time that the quartz plate workpiece should ideally remain in the cooling section, obtained from the production management system or process database. This can be achieved through database queries, configuration file reading, or manual input. For example, the standard process parameters corresponding to the current production batch, including the theoretical cooling time, can be automatically retrieved from the MES (Manufacturing Execution System).

[0042] The actual cooling time deviation rate is calculated by dividing the difference between the actual and theoretical cooling times by the theoretical cooling time. This deviation rate quantifies the degree of deviation in cooling time by comparing the actual dwell time of the workpiece in the cooling section with the theoretical dwell time. Specifically, real-time sensor data can be used to collect the actual cooling time, followed by mathematical calculations. For example, if the theoretical cooling time is 100 seconds and the actual cooling time is 110 seconds, the deviation rate is (110-100) / 100 = 0.1, or 10%. This deviation rate directly reflects the accuracy of macroscopic time control in the cooling process and is crucial for identifying cooling anomalies caused by fluctuations in conveyor speed, among other reasons.

[0043] The equivalent cooling intensity characterizes the actual heat dissipation capacity of the cooling section for the quartz plate workpiece within a specific time period. Ambient temperature determines the driving force of heat exchange. The lower the ambient temperature, the greater the temperature difference between the quartz plate surface and the ambient temperature, and the faster the heat dissipation. The fan operating frequency determines the airflow velocity across the quartz plate surface; the higher the frequency, the faster the airflow, the larger the convective heat transfer coefficient, and the stronger the heat dissipation effect. The equivalent cooling intensity couples these two factors into a single value, used to uniformly measure the actual cooling performance of the cooling section under different operating conditions.

[0044] For a specific sampling moment within a virtual temperature zone of the cooling section, the fan operating frequency and ambient temperature at that moment are first acquired. Then, the square root of the fan frequency at that moment is taken to obtain a relative value characterizing the current convective heat transfer coefficient. Simultaneously, the difference between the preset plate reference temperature and the ambient temperature at that moment is calculated; this difference reflects the magnitude of the current heat exchange driving force. Finally, the two values ​​are multiplied to obtain the instantaneous cooling driving force value at that moment.

[0045] When a quartz plate passes through a virtual temperature zone, it experiences a continuous time window. Within this time window, the instantaneous cooling driving force value at each sampling moment is calculated. Then, these instantaneous values ​​are accumulated within this time window, i.e., all instantaneous values ​​are integrated over time. Finally, the accumulated result is averaged using the total duration of the time window to obtain the average equivalent cooling intensity of the virtual temperature zone.

[0046] The equivalent cooling intensity is calculated by weighting the average equivalent cooling intensity across all virtual temperature zones traversed by the quartz slab, based on the dwell time of the quartz slab in each zone. The longer the dwell time in a temperature zone, the greater its cooling effect on the entire slab, and the higher its weight in the weighted average.

[0047] This process involves calculating the ambient temperature variance and the degree of abrupt changes in fan frequency based on the ambient temperature sequence and fan operating frequency sequence. These are used as environmental fluctuation characteristics to capture the dynamic instability of the cooling environment. This can be achieved using statistical and signal processing methods. For example, the ambient temperature variance can be obtained by calculating the statistical variance of ambient temperature data over a period of time, reflecting the amplitude of temperature fluctuations. The degree of abrupt changes in fan frequency can be measured by calculating the continuous difference or rate of change in the fan frequency sequence. For instance, when the fan frequency changes significantly within a short period, an abrupt change can be considered to exist.

[0048] The process of combining the actual cooling time deviation rate, equivalent cooling intensity, and environmental fluctuation characteristics, and then performing dimensionless processing to generate a cooling thermal history feature vector, involves integrating the extracted features into a unified vector and eliminating the influence of differences in feature dimensions. This can be achieved using methods such as feature concatenation and standardization (e.g., Z-score standardization, Min-Max standardization). For example, the calculated actual cooling time deviation rate, equivalent cooling intensity, ambient temperature variance, and fan frequency mutation degree can be arranged sequentially into a vector. Then, each element in the vector can be standardized to ensure a uniform numerical range, thus preventing certain features from dominating subsequent analyses due to their large values.

[0049] By combining these multi-dimensional and multi-angle features and performing dimensionless processing, the generated feature vector can comprehensively and objectively reflect the actual thermal history of the quartz plate workpiece in the cooling section, providing a more accurate data basis for subsequent cooling deviation index calculation and anomaly diagnosis.

[0050] Optionally, the step of calculating the cooling deviation index of the quartz plate workpiece based on the cooling thermal history feature vector, and classifying the quartz plate workpiece into different cooling state categories according to the cooling deviation index, includes: Quartz plate sample workpieces produced under standard cooling conditions are selected from the historical production process database. A reference feature space is constructed based on the cooling thermal history feature vector of the quartz plate sample workpieces, and the mean vector of the reference feature space is calculated. The Mahalanobis distance between the cooling thermal history eigenvector and the mean vector of the quartz plate workpiece is calculated and used as the cooling offset index. When the cooling offset index is less than the first preset threshold, the quartz plate workpiece is marked as being in normal cooling state. When the cooling offset index is between the first preset threshold and the second preset threshold, the quartz plate workpiece is marked as having a slight cooling offset state; the second preset threshold is greater than the first preset threshold. When the cooling offset index is greater than the second preset threshold, the quartz plate workpiece is marked as having a severe cooling offset state.

[0051] The selection of quartz plate sample workpieces produced under standard cooling conditions from the historical production process database refers to choosing quartz plate products produced under ideal or stable cooling conditions from existing production data. This can be achieved manually or automatically based on preset conditions (such as the range of ambient temperature fluctuations in the cooling section, the stability of the fan operating frequency, and the deviation rate between actual and theoretical cooling times). The cooling thermal history feature vectors of these sample workpieces are used to construct a benchmark feature space, which represents the feature distribution under standard cooling conditions. The mean vector of the benchmark feature space can be calculated using statistical methods, such as maximum likelihood estimation or sample mean. The mean vector describes the average level of each feature under standard cooling conditions. By constructing such a benchmark model, an objective comparative benchmark is provided for subsequent evaluation of the cooling state of individual workpieces.

[0052] Among them, the Mahalanobis distance between the cooling thermal history feature vector and the mean vector of the quartz plate workpiece is calculated as a cooling offset index. This index quantifies the difference between the current cooling state of the workpiece and the standard cooling state. It can take into account the correlation between different feature dimensions and perform scale-invariant measurement on the data, thereby more accurately reflecting the degree of deviation between the actual cooling experience of the workpiece and the standard cooling experience.

[0053] Specifically, when the cooling offset index is less than a first preset threshold, the quartz plate workpiece is marked as being in a normal cooling state, meaning the cooling process is very close to the standard state. The first preset threshold can be determined based on historical data analysis, expert experience, or statistical methods (such as the 3σ principle); for example, it can be set to a certain percentile of the Mahalanobis distance distribution. When the cooling offset index is between the first and second preset thresholds, the quartz plate workpiece is marked as being in a slightly cooling offset state, indicating a certain degree of deviation in the cooling process. The second preset threshold is greater than the first, and its setting method is similar to the first preset threshold, but the value is higher; for example, it can be set to another higher percentile of the Mahalanobis distance distribution. When the cooling offset index is greater than the second preset threshold, the quartz plate workpiece is marked as being in a severely cooling offset state, indicating a significant deviation in the cooling process. By setting different thresholds, the continuous cooling offset index is discretized into clear cooling state categories, allowing subsequent anomaly localization and diagnosis to be targeted, improving the accuracy and efficiency of the diagnosis.

[0054] By setting different thresholds, the continuous cooling offset index is discretized into clear cooling state categories, enabling subsequent anomaly localization and diagnosis to be targeted, improving the accuracy and efficiency of diagnosis, identifying cooling anomalies of different degrees, and providing a more detailed basis for subsequent production process diagnosis.

[0055] Optionally, the step of calculating the flatness statistical characteristics of the quartz plate workpiece for each cooling state category based on the flatness value and cooling state category of the quartz plate workpiece includes: Calculate the mean and variance of flatness for quartz plate workpieces marked as normal cooling state, slightly cooled, and severely cooled. For quartz plate workpieces of each cooling state category, a one-way ANOVA was conducted with the cooling state category as the independent variable and the flatness value as the dependent variable to obtain the significance level test value, which was used as the statistical characteristic of flatness.

[0056] The calculation of the flatness mean and variance for quartz plate workpieces marked as being in normal cooling, slightly cooled, and severely cooled states respectively refers to a preliminary analysis of the central tendency and dispersion of the flatness data of quartz plate workpieces under different cooling state categories. Specifically, this can be achieved as follows: First, all quartz plate workpieces are divided into normal cooling, slightly cooled, and severely cooled states based on their cooling deviation index. Then, for each cooling state category, the flatness values ​​of all quartz plate workpieces in that category are collected. Next, the arithmetic mean of the flatness values ​​in each category is calculated as the flatness mean, and the variance of the flatness values ​​in each category is calculated as the flatness variance. For example, for quartz plate workpieces in normal cooling, the flatness mean can be obtained by adding the flatness values ​​of all workpieces in normal cooling and then dividing by the number of workpieces; the flatness variance can be obtained by averaging the squares of the differences between each value and the mean.

[0057] For quartz plate workpieces in each cooling state category, a one-way ANOVA is performed with the cooling state category as the independent variable and the flatness value as the dependent variable to obtain a significance level test value. This serves as a statistical characteristic of flatness, quantifying whether the differences in the mean and variance of flatness among different cooling state categories are statistically significant. Specifically, this can be achieved as follows: First, the cooling state category is set as the independent variable, and the flatness value of the quartz plate workpiece is set as the dependent variable. Then, a one-way ANOVA is performed to test whether there are significant differences in the mean flatness under different cooling state categories. During the ANOVA, the between-group variance and within-group variance are calculated, and based on these variances, the F-statistic and the corresponding p-value (significance level test value) are calculated.

[0058] By calculating the cooling deviation index of quartz plate workpieces and classifying them into normal cooling state, slight cooling deviation state, and severe cooling deviation state based on this index, the calculation of flatness statistical characteristics can specifically reflect the flatness performance under different cooling states. This allows for a more accurate revelation of the impact of the cooling process on flatness, enabling more precise identification and quantification of the influence of different cooling states on the flatness of quartz plates. This provides more discriminative data support for subsequent anomaly source determination.

[0059] Optionally, the steps for determining the source of flatness anomalies based on differences in flatness statistical characteristics include: If the statistical characteristics of flatness are less than the preset difference threshold, the following judgment is performed: When the flatness mean and flatness variance of the quartz plate workpiece marked as normal cooling state are both within the corresponding preset normal range, and the flatness mean and flatness variance of the quartz plate workpiece marked as slight cooling deviation state and / or the quartz plate workpiece marked as severe cooling deviation state both exceed the corresponding preset normal range, it is determined that the flatness abnormality originates from the additional influence of the cooling section. When the flatness mean and flatness variance of the quartz plate workpiece marked as being in normal cooling state both exceed the corresponding preset normal range, and the flatness mean and flatness variance of the quartz plate workpiece marked as being in mild cooling deviation state and the quartz plate workpiece marked as being in severe cooling deviation state are both within the corresponding preset normal range, it is determined that the flatness abnormality originates from fluctuations in the sintering process itself. When the flatness mean and flatness variance of the quartz plate workpiece marked as being in normal cooling state both exceed the corresponding preset normal range, and the flatness mean and / or flatness variance of the quartz plate workpiece marked as being in mild cooling deviation state and the quartz plate workpiece marked as being in severe cooling deviation state both exceed the corresponding preset normal range, the flatness abnormality is determined to originate from a mixed abnormality. If the flatness statistical characteristics are not less than the preset difference threshold, the flatness abnormality is determined to originate from fluctuations in the sintering process itself.

[0060] Among them, the statistical characteristics of smoothness refer to the significance level test value calculated by one-way ANOVA, which can be achieved by using the P-value test. The P-value test directly gives the probability of the observed difference occurring under the null hypothesis.

[0061] The preset difference threshold is a critical value used to determine whether the statistical characteristics of flatness are significant. It can be set based on actual production experience and statistical principles; for example, it can be set to 0.05 or 0.01. When the significance level test value is less than this threshold, it indicates that there are significant differences in the statistical characteristics of flatness between different cooling state categories.

[0062] The preset normal range refers to the range in which the mean and variance of the flatness of a quartz plate workpiece should be under normal production conditions. Specifically, it can be determined by statistical analysis based on historical production data, such as calculating the confidence intervals of the mean and variance of the flatness of historical normal batches, or by setting the upper and lower limits based on expert experience.

[0063] Normal cooling state, slight cooling deviation state, and severe cooling deviation state refer to the classification of quartz plate workpieces based on the cooling deviation index. The cooling deviation index is an indicator that measures the degree of deviation between the actual cooling experience and the standard cooling conditions, and can be calculated using Mahalanobis distance. Mahalanobis distance can comprehensively consider multiple cooling thermal history characteristic parameters and eliminate the correlation between these parameters.

[0064] The average flatness value refers to the average flatness value of all quartz plate workpieces under a certain cooling state category. Specifically, it can be calculated by summing all flatness values ​​under that category and then dividing by the number of workpieces.

[0065] Flatness variance refers to the degree of dispersion of flatness values ​​of all quartz plate workpieces under a certain cooling state category. Specifically, it can be obtained by calculating the average of the squares of the differences between each flatness value and the mean.

[0066] When the flatness statistical characteristics are less than a preset difference threshold, it indicates that there are significant differences in the flatness statistical characteristics between different cooling state categories. At this time, by comparing whether the mean flatness and the variance of flatness under different cooling state categories are within a preset normal range, the source of the anomaly can be further refined.

[0067] By classifying workpieces into different cooling state categories and calculating the flatness statistical characteristics of each category, the impact of the cooling process on flatness can be captured more precisely. Thus, when there are significant differences in flatness statistical characteristics, it is possible to more accurately identify whether the anomaly originates from the additional effects of the cooling section, fluctuations in the sintering process itself, or a mixed anomaly.

[0068] When the flatness statistical characteristics are not less than the preset difference threshold, it indicates that the differences in flatness statistical characteristics between different cooling state categories are not significant, that is, flatness anomalies are common regardless of the cooling state.

[0069] By analyzing the differences in the statistical characteristics of the flatness of quartz plate workpieces under different cooling conditions, the source of flatness anomalies can be accurately located, solving the problem in existing technologies that cannot effectively distinguish between sintering process fluctuations and additional effects of cooling sections that cause flatness anomalies.

[0070] Optionally, the steps for generating production process diagnostic results for quartz plate workpieces based on the source of flatness anomalies include: When it is determined that the flatness abnormality is caused by the additional influence of the cooling section, a production process diagnostic result including suggestions for troubleshooting the cooling section is generated. When it is determined that the flatness abnormality is caused by fluctuations in the sintering process itself, a production process diagnostic result including sintering process diagnostic alarms is generated. When the flatness anomaly is determined to originate from a mixed anomaly, a production process diagnostic result including bidirectional parallel troubleshooting suggestions is generated.

[0071] The production process diagnostic results for generating quartz plate workpieces refer to the automatic generation of instructive text or instructions based on identified sources of flatness anomalies. These instructions guide production personnel in troubleshooting and process adjustments. This can be achieved in various ways, such as displaying diagnostic information through a software interface or sending instructions to relevant control units via an automated system.

[0072] When the flatness anomaly is determined to originate from the additional influence of the cooling section, generating a production process diagnostic result that includes troubleshooting suggestions for the cooling section means that if the analysis indicates the flatness problem of the quartz slab is primarily related to abnormal factors in the cooling process, then troubleshooting suggestions for the cooling section will be generated. These suggestions may include checking whether parameters such as the ambient temperature, fan operating status, and conveyor speed of the cooling section meet the set values, or whether there are abnormal fluctuations. For example, it may be suggested to check whether the cooling fan speed is stable, or whether the temperature sensor in the cooling section is working properly.

[0073] When the flatness anomaly is determined to originate from fluctuations in the sintering process itself, the production process diagnostic results, including sintering process diagnostic alarms, will generate such alarms. This means that if the flatness anomaly is primarily related to core process parameters such as sintering furnace temperature control and sintering time, a sintering process diagnostic alarm will be generated. The alarm information can remind process personnel to monitor the sintering furnace's operating status, check whether sintering parameters deviate from set values, or whether there are issues such as equipment aging. For example, the alarm information might indicate abnormal temperature fluctuations in a certain area of ​​the sintering furnace, suggesting a check of the heating elements.

[0074] When a flatness anomaly is determined to originate from a mixed anomaly, generating a production process diagnostic result that includes bidirectional parallel troubleshooting suggestions means that if the analysis suggests the quartz slab's flatness issue may be affected by both the cooling section and the sintering process simultaneously, then troubleshooting suggestions will be generated for both. These suggestions will encourage technicians to investigate both aspects concurrently, avoiding oversights due to a single-direction approach. For example, it might suggest simultaneously checking the cooling section's airflow velocity and the sintering furnace's temperature profile.

[0075] By generating the above different types of diagnostic results, this solution can provide customized solutions based on the nature of the anomaly source, effectively improving the diagnostic accuracy and efficiency of the quartz plate production process and reducing production losses caused by misjudgment or inadequate investigation.

[0076] Optionally, when the anomaly is determined to originate from an additional effect of the cooling section, the method further includes: Retrieve the deviation compensation relationship between the cooling offset index and the flatness value; For quartz plate workpieces marked as slightly cooled and marked as severely cooled, their cooling offset index is substituted into the deviation compensation relationship to calculate the estimated flatness deviation caused by residual thermal stress. The flatness correction value is obtained by subtracting the flatness deviation estimate from the flatness values ​​of the quartz plate workpieces marked as slightly cooled and the quartz plate workpieces marked as heavily cooled. Generate production process diagnostic results that include flatness correction values ​​and cooling section troubleshooting suggestions.

[0077] The process of retrieving the deviation compensation relationship between the cooling offset index and the flatness value refers to obtaining the quantitative relationship between the cooling offset index and the flatness value from a pre-established database or model when the flatness anomaly is determined to originate from the additional influence of the cooling section. Specifically, this can be achieved in the following ways: By using historical production data, a regression model can be established between the cooling offset index and the flatness deviation, such as a linear regression, multinomial regression, or nonlinear regression model. This model can reflect the magnitude and direction of the impact of different degrees of cooling offset on flatness. Alternatively, a physical simulation model can be used to simulate the formation of residual thermal stress inside the quartz plate under different cooling conditions and its impact on flatness, thereby obtaining the mapping relationship between the cooling offset index and the flatness deviation.

[0078] Specifically, for quartz plate workpieces marked as having a slight cooling offset and those marked as having a severe cooling offset, the estimated flatness deviation caused by residual thermal stress is calculated by substituting their cooling offset index into the deviation compensation relationship. This means that after obtaining the deviation compensation relationship, the cooling offset index of the quartz plate workpiece currently diagnosed as having a slight or severe cooling offset is input into the relationship. This can be achieved in the following ways: For example, if the deviation compensation relationship is a mathematical formula, the cooling offset index is directly substituted into the formula for calculation. Alternatively, if the deviation compensation relationship is a lookup table, the corresponding estimated flatness deviation value is found in the table based on the cooling offset index. For instance, when the cooling offset index is 0.5, the estimated flatness deviation value is calculated to be 0.1 mm based on the preset deviation compensation relationship.

[0079] The flatness correction value is obtained by subtracting the estimated flatness deviation from the flatness values ​​of quartz plate workpieces marked as having a slight cooling deviation and those marked as having a severe cooling deviation. This means that after calculating the estimated flatness deviation, this estimated value is subtracted from the original measured flatness value of the workpiece. Specifically, this can be achieved as follows: if the original flatness value is P_original and the estimated flatness deviation value is P_estimate, then the flatness correction value P_corrected = P_original - P_estimate. For example, if the original flatness value of a quartz plate workpiece is 0.3 mm and the calculated estimated flatness deviation value is 0.1 mm, then the flatness correction value is 0.2 mm. This correction value represents the theoretically expected flatness of the workpiece after excluding the additional effects of the cooling section, providing a more accurate benchmark for evaluating the flatness performance of the sintering process itself.

[0080] The generation of production process diagnostic results, including flatness correction values ​​and cooling section troubleshooting suggestions, involves integrating the flatness correction value with the troubleshooting suggestions for the cooling section to form a final diagnostic report. This can be achieved in the following ways: the diagnostic results can be presented to production personnel in the form of text reports, charts, or visualization interfaces. The report will clearly indicate the specific impact of cooling offset on flatness (reflected by the flatness correction value) and provide specific troubleshooting directions, such as checking the operating status of the cooling fan, ambient temperature fluctuations, and conveyor speed stability. For example, the diagnostic results might show that the flatness correction value for quartz plate workpiece A is 0.2mm, suggesting checking whether the cooling section fan frequency is stable and checking the ambient temperature fluctuations in the cooling section.

[0081] When generating the diagnostic report, the cooling thermal history feature vector of the current workpiece is compared dimension by dimension with the mean vector of the reference feature space, and the deviation of each dimension is calculated, i.e., the deviation of the current value relative to the reference value. Then, the deviation of each dimension is normalized to eliminate the differences in the units and numerical ranges of different dimensions, so that the deviations of each dimension can be compared on the same scale. After normalization, one or more dimensions with the largest deviations are the main contributors to the increase in the cooling offset index.

[0082] Based on the dimension type to which the main contribution item belongs, the corresponding investigation direction will be automatically matched: When the main contributing factor is a low average global equivalent cooling intensity, it indicates that the overall cooling capacity experienced by the quartz plate throughout the cooling section is insufficient. In this case, the investigation should focus on checking whether the exhaust fan is operating within the normal frequency range, whether the fan belt is loose or worn, causing a decrease in actual speed, whether the inverter output frequency matches the control command, and whether the fan filter is clogged, leading to reduced airflow.

[0083] When the main contributing factor is the deviation rate of actual cooling time, which deviates from the normal range, it indicates an abnormal residence time of the quartz plate in the cooling section. If the actual cooling time is significantly longer than the theoretical time, the investigation should focus on whether the conveyor roller drive motor is working properly, whether the roller speed is reduced due to mechanical jamming, and whether there are uncontrolled changes in the cycle time of the preceding and following processes. If the actual cooling time is significantly shorter than the theoretical time, the investigation should focus on whether the conveyor roller speed is too fast, resulting in insufficient cooling of the plate.

[0084] When the main contributing factor is excessively large ambient temperature variance, it indicates that the quartz plate experienced severe or uneven ambient temperature fluctuations during its flow through the cooling section. The investigation should focus on checking whether workshop doors and windows were abnormally opened, allowing cold air to enter; whether the sealing cover of the cooling section was intact; and whether the workshop's ambient temperature control system was operating normally during day and night cycles.

[0085] When the main contributing factor is an excessively large fluctuation in fan frequency, it indicates that the cooling fan experienced frequent or unstable frequency changes during the passage of the quartz plate. The investigation should focus on checking the stability of the inverter output, the presence of power supply voltage fluctuations, whether wear on the fan bearings caused changes in rotational resistance leading to frequent inverter adjustments, and whether there is signal interference in the fan control circuit.

[0086] When the main contributing factor is an excessively large spatial gradient variance, it indicates that the quartz plate experienced drastically different cooling conditions between different virtual temperature zones, meaning that cooling was severely uneven along the conveying direction. The investigation should focus on checking for individual differences or partial malfunctions in the fans of each temperature zone in the cooling section, checking for abnormalities in the separation or duct design between temperature zones, and checking for areas along the conveying direction where localized temperature accumulation or poor ventilation exists.

[0087] The above solution can effectively solve the problem of insufficient practicality and guidance of the diagnostic results in the existing technology, enabling the production line to compensate or correct the flatness deviation caused by cooling offset, thereby improving the flatness of the quartz plate.

[0088] Optionally, after the step of generating a production process diagnostic result for the quartz plate workpiece based on the source of the flatness anomaly, the following is included: Adjustment instructions are issued to the cooling section control unit and / or sintering furnace based on the production process diagnostic results.

[0089] Specifically, issuing adjustment commands to the cooling section control unit and / or sintering furnace based on production process diagnostic results means that after diagnosing the source of the flatness anomaly, corresponding adjustment commands are automatically or semi-automatically generated according to the type of diagnostic result, and these commands are sent to the control unit responsible for the cooling section or sintering furnace. This can be achieved in the following ways: when the diagnostic result indicates that the flatness anomaly originates from additional influences of the cooling section, adjustment commands can be issued to the cooling section control unit, such as adjusting the fan operating frequency or cooling duration to optimize the cooling process. When the diagnostic result indicates that the flatness anomaly originates from fluctuations in the sintering process itself, adjustment commands can be issued to the sintering furnace, such as adjusting the sintering temperature or holding time to correct the sintering process.

[0090] Specifically, this mechanism, which issues adjustment instructions to relevant control units based on the diagnostic results of the production process, enables the diagnostic results to be directly translated into actual production operations, thereby correcting abnormalities in the production process in a timely manner, improving the flatness of the quartz slab, and optimizing overall production efficiency and product quality.

[0091] Secondly, see Figure 2 This application also proposes a quartz plate production process data analysis system for performing the above-described quartz plate production process data analysis method. The system includes: The data acquisition module 210 is used to acquire multi-dimensional time-series data of the cooling section and flatness values ​​of the quartz plate workpiece. The feature extraction module 220 is used to extract feature parameters that characterize the actual cooling experience of the quartz plate workpiece based on the multi-dimensional time series data of the cooling section, and generate a cooling thermal history feature vector. The state assessment module 230 is used to calculate the cooling deviation index of the quartz plate workpiece based on the cooling thermal history feature vector, and classify the quartz plate workpiece into different cooling state categories according to the cooling deviation index. The anomaly location module 240 is used to calculate the flatness statistical characteristics of the quartz plate workpiece for each cooling state category based on the flatness value and cooling state category of the quartz plate workpiece, and determine the source of flatness anomaly based on the difference in flatness statistical characteristics. The diagnostic result generation module 250 is used to generate diagnostic results for the production process of quartz plate workpieces based on the source of flatness anomalies.

[0092] The above technical solutions can automatically and intelligently remove the hidden interference of cooling process fluctuations on the final flatness result, accurately identify the true stability state of the sintering process, effectively solve the problem of quality misjudgment caused by the coupling of multiple processes in complex continuous production lines, and provide a reliable guarantee for the high-quality and high-yield production of quartz plates.

[0093] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A data analysis method for the quartz plate production process, characterized in that, include: Obtain multi-dimensional time-series data of the cooling section and flatness values ​​of the quartz plate workpiece; Based on the multi-dimensional time-series data of the cooling section, feature parameters characterizing the actual cooling experience of the quartz plate workpiece are extracted to generate a cooling thermal history feature vector. Based on the cooling thermal history feature vector, the cooling deviation index of the quartz plate workpiece is calculated, and the quartz plate workpiece is classified into different cooling state categories according to the cooling deviation index. Based on the flatness value and cooling state category of the quartz plate workpiece, the flatness statistical characteristics of the quartz plate workpiece for each cooling state category are calculated respectively, and the source of flatness anomaly is determined according to the difference of the flatness statistical characteristics. Based on the source of the flatness anomaly, a production process diagnostic result for the quartz plate workpiece is generated.

2. The data analysis method for the quartz plate production process according to claim 1, characterized in that, The steps for obtaining multi-dimensional time-series data of the cooling section and flatness values ​​of the quartz plate workpiece include: The ambient temperature sequence, fan operating frequency sequence, and actual cooling duration of the cooling section are obtained as multi-dimensional time-series data of the cooling section. Obtain the flatness test value as the flatness value; The acquired multidimensional time-series data of the cooling section and the flatness value are spatiotemporally correlated according to the physical flow sequence of the quartz plate workpiece on the production line.

3. The data analysis method for the quartz plate production process according to claim 2, characterized in that, The step of extracting feature parameters characterizing the actual cooling experience of the quartz plate workpiece based on the multidimensional time-series data of the cooling section, and generating a cooling thermal history feature vector, includes: Retrieve the theoretical cooling time from the preset process conditions; The difference between the actual cooling time and the theoretical cooling time is calculated and then divided by the theoretical cooling time to obtain the actual cooling time deviation rate; The equivalent cooling intensity of the cooling section is calculated based on the ambient temperature sequence and the fan operating frequency sequence. Based on the ambient temperature sequence and the fan operating frequency sequence, the ambient temperature variance and the degree of abrupt change in fan frequency are calculated as environmental fluctuation characteristics. The actual cooling duration deviation rate, the equivalent cooling intensity, and the environmental fluctuation characteristics are combined and dimensionless to generate the cooling thermal history feature vector.

4. The data analysis method for the quartz plate production process according to claim 3, characterized in that, The step of calculating the cooling deviation index of the quartz plate workpiece based on the cooling thermal history feature vector, and classifying the quartz plate workpiece into different cooling state categories according to the cooling deviation index, includes: Quartz plate sample workpieces produced under standard cooling conditions are selected from the historical production process database. A reference feature space is constructed based on the cooling thermal history feature vector of the quartz plate sample workpieces, and the mean vector of the reference feature space is calculated. The Mahalanobis distance between the cooling thermal history feature vector of the quartz plate workpiece and the mean vector is calculated and used as the cooling offset index. When the cooling offset index is less than the first preset threshold, the quartz plate workpiece is marked as being in a normal cooling state. When the cooling offset index is between the first preset threshold and the second preset threshold, the quartz plate workpiece is marked as having a slight cooling offset state; the second preset threshold is greater than the first preset threshold. When the cooling offset index is greater than the second preset threshold, the quartz plate workpiece is marked as having a severe cooling offset state.

5. The data analysis method for the quartz plate production process according to claim 4, characterized in that, The step of calculating the flatness statistical characteristics of the quartz plate workpiece for each cooling state category based on the flatness value and cooling state category of the quartz plate workpiece includes: Calculate the mean flatness and variance of the quartz plate workpieces marked as being in the normal cooling state, the slightly cooled state, and the heavily cooled state, respectively. For each of the aforementioned cooling state categories of quartz plate workpieces, a one-way ANOVA is performed with the cooling state category as the independent variable and the flatness value as the dependent variable to obtain the significance level test value, which is used as the statistical feature of flatness.

6. The data analysis method for the quartz plate production process according to claim 5, characterized in that, The step of determining the source of flatness anomalies based on the differences in the statistical characteristics of flatness includes: If the flatness statistical characteristic is less than a preset difference threshold, the following judgment is performed: When the mean flatness and variance of the quartz plate workpiece marked as being in the normal cooling state are both within the corresponding preset normal range, and the mean flatness and variance of the quartz plate workpiece marked as being in the mild cooling deviation state and / or the quartz plate workpiece marked as being in the severe cooling deviation state both exceed the corresponding preset normal range, it is determined that the flatness abnormality originates from the additional influence of the cooling section. When the mean flatness and variance of the quartz plate workpiece marked as being in the normal cooling state both exceed the corresponding preset normal range, and the mean flatness and variance of the quartz plate workpiece marked as being in the slight cooling deviation state and the quartz plate workpiece marked as being in the severe cooling deviation state are both within the corresponding preset normal range, it is determined that the flatness abnormality originates from fluctuations in the sintering process itself. When the mean flatness and variance of the quartz plate workpiece marked as being in the normal cooling state both exceed the corresponding preset normal range, and the mean flatness and / or variance of the quartz plate workpiece marked as being in the mild cooling deviation state and the quartz plate workpiece marked as being in the severe cooling deviation state both exceed the corresponding preset normal range, it is determined that the flatness abnormality originates from a mixed abnormality. If the flatness statistical characteristics are not less than the preset difference threshold, it is determined that the flatness anomaly originates from fluctuations in the sintering process itself.

7. The data analysis method for the quartz plate production process according to claim 6, characterized in that, The step of generating the production process diagnostic results for the quartz plate workpiece based on the source of the flatness anomaly includes: When it is determined that the flatness abnormality originates from the additional influence of the cooling section, a production process diagnostic result including cooling section troubleshooting suggestions is generated. When it is determined that the flatness abnormality originates from fluctuations in the sintering process itself, a production process diagnostic result including sintering process diagnostic alarms is generated. When it is determined that the flatness anomaly originates from a mixed anomaly, a production process diagnostic result including bidirectional parallel troubleshooting suggestions is generated.

8. The data analysis method for the quartz plate production process according to claim 6, characterized in that, When the anomaly is determined to originate from an additional effect of the cooling section, the method further includes: Retrieve the deviation compensation relationship between the cooling offset index and the flatness value; For quartz plate workpieces marked as having a slight cooling offset state and quartz plate workpieces marked as having a severe cooling offset state, their cooling offset index is substituted into the deviation compensation relationship to calculate the estimated flatness deviation caused by residual thermal stress. The flatness correction value is obtained by subtracting the flatness deviation estimate from the flatness values ​​of the quartz plate workpiece marked as the mild cooling offset state and the quartz plate workpiece marked as the severe cooling offset state. Generate production process diagnostic results that include the flatness correction value and cooling section troubleshooting suggestions.

9. The data analysis method for the quartz plate production process according to claim 7, characterized in that, After the step of generating the production process diagnostic results for the quartz plate workpiece based on the source of the flatness anomaly, the following is included: Based on the production process diagnostic results, adjustment commands are issued to the cooling section control unit and / or the sintering furnace.

10. A quartz plate production process data analysis system, used to execute the quartz plate production process data analysis method as described in any one of claims 1 to 9, characterized in that, The system includes: The data acquisition module is used to acquire multi-dimensional time-series data of the cooling section and flatness values ​​of the quartz plate workpiece. The feature extraction module is used to extract feature parameters that characterize the actual cooling experience of the quartz plate workpiece based on the multi-dimensional time series data of the cooling section, and generate a cooling thermal history feature vector. The state assessment module is used to calculate the cooling deviation index of the quartz plate workpiece based on the cooling thermal history feature vector, and classify the quartz plate workpiece into different cooling state categories according to the cooling deviation index. An anomaly location module is used to calculate the flatness statistical characteristics of the quartz plate workpiece for each cooling state category based on the flatness value and cooling state category of the quartz plate workpiece, and determine the source of flatness anomaly based on the differences in the flatness statistical characteristics. The diagnostic result generation module is used to generate a production process diagnostic result for the quartz plate workpiece based on the source of the flatness anomaly.

Images