A digital environment control method and system for photosensitive drum production

Abstract

The present application relates to the field of environmental control technology, and in particular to a digital environmental control method and system for photosensitive drum production; the method comprises the steps of: collecting environmental data during the photosensitive drum production process; analyzing the correlation between the environmental data to obtain each pair of theoretical related data, calculating the trend correlation coefficient based on the difference in data distribution form between the environmental data in the theoretical related data, using a causal relationship test to obtain the test results of the theoretical related data, calculating the causal coefficient between the theoretical related data, calculating the causal legitimacy coefficient, obtaining an updated local reachable density, and then using the LOF abnormal data detection algorithm in combination with the updated local reachable density to control the environmental data, thereby improving the efficiency of environmental control in photosensitive drum production and improving the effect of environmental control.

Description

Technical Field

[0001] The present application relates to the field of environmental control technology, and in particular to a digital environmental control method and system for photosensitive drum production. Background Art

[0002] With the development of science and technology, users' requirements for printers and copiers have also increased. The requirements for photosensitive drums, the core components of printers and copiers, have also become higher. Defects in photosensitive drums will affect imaging quality. The photoreceptor drum (Photoreceptor Drum or Imaging Drum) is the core imaging component of laser printers, digital copiers and multifunction printers. The core function of the photoreceptor drum is to convert light signals into visible images through photoconductive materials (such as selenium and organic photoconductive materials OPC). Its function is equivalent to the "negative" in traditional film cameras, but it realizes digital imaging through the photoelectric effect and electrostatic principles. Therefore, in the process of photosensitive drum production, traditional abnormal data detection methods usually perform anomaly detection based on data density, and do not consider the relationship between different types of environmental data for anomaly detection, such as temperature and humidity. In the actual production process of photosensitive drums, abnormal alarms are triggered based on density abnormalities, thereby controlling the environment.

[0003] The LOF abnormal data detection algorithm identifies anomalies by quantifying the density deviation of a data point relative to its neighborhood, and focuses more on local density. In the process of photosensitive drum production, there are many environmental factors that need to be controlled. For each environmental factor, the traditional LOF abnormal data detection algorithm only identifies anomalies based on the local density of a single environmental factor, without considering the correlation between each environmental factor. Although it is relatively simple and easy to implement, this method has limitations and has poor detection effect in more complex production environments. Summary of the Invention

[0004] In order to solve the problem of poor detection of abnormal environmental data in the photosensitive drum production environment due to failure to consider the correlation between different types of environmental data, the present application provides a digital environment control method and system for photosensitive drum production.

[0005] In a first aspect, the present application provides a digital environment control method for photosensitive drum production, which adopts the following technical solution:

[0006] Collect environmental data during the photosensitive drum production process, including: light intensity data, temperature data, relative humidity data, concentration of particulate matter in the air, and electrostatic voltage data;

[0007] Analyze the correlation between environmental data to obtain each pair of theoretical correlation data. Calculate the trend correlation coefficient based on the differences in data distribution patterns between environmental data in the theoretical correlation data. Use the Granger causality test to obtain the causal relationship of the theoretical correlation data and determine the causal coefficient. The causal coefficient is assigned a value of 1 when the temperature and humidity data or the light and temperature data show a unidirectional causal relationship, and the electrostatic voltage data and the air particulate matter concentration data show a bidirectional causal relationship; otherwise, it is assigned a value of 0.

[0008] The sum of the trend correlation coefficient and the causal coefficient is used as an exponent and an exponential operation is performed with a natural constant as the base to construct a causal legitimacy coefficient; after the causal legitimacy coefficient is normalized, it is multiplied with the original local reachable density calculated by the LOF anomaly data detection algorithm to generate an updated local reachable density, anomaly detection is performed on the environmental data based on the updated local reachable density, and the environment is controlled based on the anomaly detection result.

[0009] The beneficial effect is: the traditional LOF algorithm only relies on the local density of a single environmental parameter to detect anomalies, ignoring the physical correlation between parameters, resulting in a high false alarm rate. This application quantifies the distribution morphology differences between parameters through the trend correlation coefficient, combines the causal coefficient to verify the logical rationality of the physical relationship, and significantly improves the accuracy of anomaly discrimination.

[0010] Furthermore, the method for acquiring the theoretically relevant data is: taking the temperature data and humidity data, the electrostatic voltage data and the air particle concentration data, and the light intensity data and the temperature data as a pair of theoretically relevant data.

[0011] Furthermore, the trend correlation coefficient is calculated based on the difference in data distribution patterns between environmental data in the theoretical correlation data, including: for each pair of theoretical correlation data in each sub-time period, calculating the skewness and kurtosis of each data in each pair of theoretical correlation data; dividing the theoretical correlation data into theoretical correlation data with negative correlation change relationships and theoretical correlation data with positive correlation change relationships, and applying different calculation logics to obtain the trend correlation coefficient based on the change relationship of the theoretical correlation data.

[0012] The beneficial effects are: this application uses skewness to characterize the consistency of trend direction, tolerates reasonable fluctuations through kurtosis differences, and is more sensitive to complex morphological changes.

[0013] Furthermore, the dividing of the theoretical related data into theoretical related data of negative correlation change relationship and theoretical related data of positive correlation change relationship includes: the theoretical related data of negative correlation change relationship is temperature data and humidity data;

[0014] The theoretical related data of the positive correlation relationship are electrostatic voltage data, particle concentration data, light intensity data and temperature data.

[0015] Furthermore, the calculation formula of the trend correlation coefficient is: ;

[0016] Where, is the trend correlation coefficient of the theoretical correlation data of each sub-time period; are the normalized skewness and kurtosis of any kind of theoretical related data in each sub-time period, It is the normalized skewness and kurtosis of another type of theoretical related data in each sub-time period.

[0017] The beneficial effects are: this application dynamically quantifies the trend consistency of environmental parameters. In the negative correlation scenario, the skewness difference value (numerator) amplifies the reverse trend characteristics, and the kurtosis difference value (denominator) suppresses the asynchronous fluctuation noise; in the positive correlation scenario, the skewness sum value strengthens the same direction drift signal, combined with the kurtosis difference to filter out occasional spike interference, reflecting the strength of trend correlation.

[0018] Furthermore, the Granger causality test is used to obtain the causal relationship of the theoretically relevant data, including: for each type of theoretically relevant data, using the causal relationship test to obtain the significant relationship between the theoretically relevant data;

[0019] For temperature data and humidity data, when temperature data is significant to humidity data, or humidity data is significant to temperature data, there is a one-way causal relationship between temperature data and humidity data; when temperature data is significant to humidity data and humidity data is significant to temperature data, there is a two-way causal relationship between temperature data and humidity data; when temperature data is not significant to humidity data and humidity data is not significant to temperature data, there is no causal relationship between temperature data and humidity data; similarly, the same method as temperature data and humidity data is used to obtain the causal relationship of other theoretically related data.

[0020] The beneficial effect is: by clearly defining the judgment rules of unidirectional causality and bidirectional causality, the statistical results of the Granger test are converted into executable process logic, and combined with the trend correlation coefficient, the humidification system failure rather than temperature fluctuation can be accurately identified.

[0021] Furthermore, the use of causal relationship test to obtain the significant relationship between theoretically related data includes: assuming that the temperature data is not the cause of the humidity data, outputting a p-value, and if the p-value is less than the significance level, rejecting the null hypothesis, that is, the temperature data is the cause of the humidity data, and the temperature data is significant to the humidity data; conversely, the temperature data is not the cause of the humidity data;

[0022] Assume that humidity data is not the cause of temperature data and output the p-value. If the p-value is less than the significance level, reject the null hypothesis, that is, humidity data is the cause of temperature data and humidity data is significant to temperature data; on the contrary, humidity data is not the cause of temperature data.

[0023] Furthermore, before adopting the Granger causality test, the method further includes: performing a stationarity test on the theoretically relevant data, and when the theoretically relevant data is non-stationary data, performing differential processing on the theoretically relevant data to obtain stationary data.

[0024] The beneficial effect is: this application solves the Granger causality false determination problem caused by the non-stationarity of industrial environment data through forced stationarity test and differential preprocessing.

[0025] Furthermore, the analysis, calculation and construction steps are all performed based on the environmental data within a sub-time period of a preset length before each collection moment.

[0026] In a second aspect, the present application provides a digital environmental control system for photosensitive drum production, which adopts the following technical solutions:

[0027] A digital environment control system for photosensitive drum production includes a processor and a memory. The memory stores computer program instructions. When the computer program instructions are executed by the processor, the digital environment control method for photosensitive drum production is implemented.

[0028] The above-mentioned digital environment control method for photosensitive drum production is generated into a computer program and stored in a memory so as to be loaded and executed by a processor. Thus, a system is made based on the memory and the processor for easy use.

[0029] This application has the following technical effects:

[0030] The present application collects each type of environmental data that is mutually correlated in the photosensitive drum production process, analyzes the correlation between different types of environmental data, obtains the mutually correlated data, constructs the trend correlation coefficient of the mutually correlated data of each sub-time period, and reflects the changing trend of the different types of environmental data; further, in order to reflect the correctness of the causal relationship between different types of data in the mutually correlated data, a causal test is performed on the mutually correlated data, and the causal legitimacy coefficient of the mutually correlated data of each sub-time period is obtained in combination with the trend correlation coefficient; finally, the local reachable density in the LOF abnormal data detection algorithm is updated based on the causal legitimacy coefficient, which solves the problem of poor abnormal environmental data detection effect due to failure to consider the correlation between different types of environmental data, thereby improving the efficiency of environmental control in photosensitive drum production and improving the effect of environmental control. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a method flow chart of a digital environment control method for photosensitive drum production in the present application;

[0032] Figure 2This is a causal relationship verification flow chart of an embodiment of the present application. DETAILED DESCRIPTION

[0033] An embodiment of the present application discloses a digital environmental control method for photosensitive drum production, which collects environmental data during the photosensitive drum production process; analyzes the correlation between the environmental data to obtain each pair of theoretical related data, calculates the trend correlation coefficient based on the difference in data distribution form between the environmental data in the theoretical related data, uses a causal relationship test to obtain the test results of the theoretical related data, calculates the causal coefficient between the theoretical related data, calculates the causal legitimacy coefficient, obtains the updated local reachable density, and then uses the LOF abnormal data detection algorithm in combination with the updated local reachable density to control the environmental data, thereby improving the efficiency of environmental control in photosensitive drum production and improving the effect of environmental control.

[0034] Reference Figure 1 A digital environment control method for photosensitive drum production includes steps S1 to S4.

[0035] S1: Collect environmental data during the photosensitive drum production process.

[0036] Environmental data at each collection moment in the photosensitive drum production process is collected. The environmental data selected in the photosensitive drum production process in this application include: light intensity data, temperature data, relative humidity data, particulate matter concentration in the air, and electrostatic voltage data. This application monitors and controls the above environmental data in the coating area during the photosensitive drum production process. The implementer can select environmental data in other areas of the photosensitive drum production process, such as the curing area, assembly area, etc. The electrostatic voltage data obtained in this application is the electrostatic voltage data of the work surface. The implementer can select other objects based on actual conditions, such as the base, roller or operator's clothing.

[0037] In this application, the interval between each collection moment in the photosensitive drum production process is 2s, and the implementer can select other time lengths based on actual conditions.

[0038] S2: Analyze the correlation between the environmental data to obtain each pair of theoretical correlation data, and calculate the trend correlation coefficient based on the difference in data distribution patterns between the environmental data in the theoretical correlation data.

[0039] In traditional LOF abnormal data detection algorithms, the degree of abnormality of each data is usually evaluated based on the local reachable density of each data. When the local reachable density of the data is larger, the value of the local abnormality factor obtained from the data is smaller, and the degree of abnormality of the data is smaller. The local abnormality factor is an important parameter that reflects whether the data is abnormal. When the value of the local abnormality factor is greater than the preset detection threshold, the data is abnormal data; however, the local reachable density of the data is usually only obtained based on the average reachable distance of the nearest neighbor of the data, and does not consider the correlation between the various environmental data in the process of photosensitive drum production. In the production process of photosensitive drums, different types of environmental data affect each other and have a certain correlation. When the local density of the data is small, it may be affected by other environmental data. At this time, the data is not abnormal, but because the local density of the data is small, it may still be judged as abnormal, resulting in poor environmental control effect during the production of photosensitive drums.

[0040] Based on the above analysis, this application analyzes the changes in various environmental data and the correlation of environmental data during the production process based on the actual situation of the photosensitive drum production process, improves the local reachable density of the LOF abnormal data detection algorithm, and improves the abnormality detection accuracy of environmental data during the photosensitive drum production process, thereby improving the environmental control effect of the photosensitive drum production process.

[0041] Based on the above analysis, the environmental data in the photosensitive drum production process is analyzed. In the photosensitive drum production process, for each of the above environmental data, the temperature data and the humidity data have a certain physical coupling relationship. Specifically, in the photosensitive drum production process, when the temperature rises, the air saturation water content increases, and the relative humidity naturally decreases; at the same time, there is a certain positive feedback loop relationship between the electrostatic voltage and the concentration of particulate matter in the air. When the electrostatic voltage on the workbench is stronger, the ability of the workbench to adsorb particulate matter is enhanced, resulting in the deterioration of the cleanliness in the photosensitive drum production process, thereby generating more frictional electricity and further increasing the static electricity; light has a direct impact on the temperature in the air, and the size of the light will affect the change in temperature in the air; therefore, the present application performs abnormal detection and processing on the environmental data based on the correlation between the environmental data in the photosensitive drum production process, thereby controlling the environment in the photosensitive drum production process.

[0042] For each type of environmental data, a time period of preset duration before each collection moment is used as each sub-time period. In one embodiment of the present application, the preset duration is 60s, and the implementer can select other durations based on actual conditions. Based on the above analysis, the temperature data and the humidity data are negatively correlated, while the electrostatic voltage data and the air particle concentration data, and the light intensity data and the temperature data are positively correlated. Moreover, in the changes of the temperature data and the humidity data, and the light intensity data and the temperature data, the change of one of the data affects and drives the change of the other data, while the electrostatic voltage and the air particle concentration data are in a positive feedback loop relationship that affects each other.

[0043] Therefore, the negatively correlated relationship between temperature data and humidity data, the electrostatic voltage data and the concentration data of particulate matter in the air, and the light intensity data and temperature data are all regarded as a pair of theoretical related data.

[0044] For each pair of theoretically relevant data in each sub-time period, the skewness and kurtosis of each data in each pair of theoretically relevant data are calculated. The calculation method of skewness and kurtosis is a well-known technique and is not described in detail in this application. The magnitude of the skewness reflects the degree of skewness of the data distribution. When the skewness is positive, the distribution of the data is right-skewed. When the skewness is negative, the distribution of the data is left-skewed. The magnitude of the kurtosis reflects the sharpness of the peak. When the kurtosis value is larger, the shape of the peak is sharper.

[0045] That is, for each pair of theoretically related data, when there is a positive feedback relationship between the two data in the theoretically related data, the positive and negative values ​​of the skewness of the two data in the theoretically related data are consistent; when there is a negative feedback relationship between the two data in the theoretically related data, the positive and negative values ​​of the skewness of the two data in the theoretically related data are opposite; therefore, the skewness and kurtosis of the theoretically related data of all sub-time periods are normalized, and the trend correlation coefficient of the theoretically related data in each sub-time period is constructed. The calculation formula is: ;

[0046] Where, is the trend correlation coefficient of the theoretical correlation data of each sub-time period; are the normalized skewness and kurtosis of any kind of theoretical related data in each sub-time period, It is the normalized skewness and kurtosis of another type of theoretical related data in each sub-time period.

[0047] It should be noted that, in each pair of theoretical correlation data in each sub-time period, when the theoretical correlation data is a negative correlation relationship, that is, in this application, when the theoretical correlation data is temperature data and humidity data, under normal circumstances, when the positive and negative values ​​of the skewness of the two data in the theoretical correlation data are opposite, the obtained The larger the value is, the more consistent it is with the changing trend of the negative correlation between the two data in the theoretical correlation data. At the same time, the change of temperature data affects the change of humidity data. The degree of change of temperature data is the same as that of humidity data. Therefore, the difference between the peak values ​​obtained is small, and the obtained The value of is small, and the value of the trend correlation coefficient is large; otherwise, the value of the trend correlation coefficient is small; secondly, when the theoretical correlation data in the sub-time period is a positive correlation relationship, that is, in this application, the theoretical correlation data is the electrostatic voltage data and the particle concentration data, the light intensity data and the temperature data, under normal circumstances, the positive and negative values ​​of the skewness of the two data in the theoretical correlation data are the same, and the obtained The larger the value is, the more it conforms to the changing trend of the positive correlation between the two data in the theoretical correlation data. At the same time, the electrostatic voltage data and the particle concentration data affect each other, and the light intensity data affects the change of the temperature data. The degree of change is the same. Therefore, the difference between the obtained peak values ​​is small, and the obtained The smaller the value of , the larger the trend correlation coefficient is; otherwise, the smaller the trend correlation coefficient is.

[0048] At this point, the trend correlation coefficient of each pair of theoretical correlation data in the sub-time period of each acquisition moment is obtained.

[0049] S3: Use Granger causality test to obtain the causal relationship of theoretically relevant data, determine the causal coefficient, use the sum of the trend correlation coefficient and the causal coefficient as an exponent, perform exponential operation with a natural constant as the base, and construct a causal legitimacy coefficient.

[0050] Based on the above steps, the trend correlation coefficient of each pair of theoretical related data in each sub-time period is obtained. At the same time, the trend changes between each pair of theoretical related data have a certain causal relationship, that is, in this application, when the theoretical related data are temperature data and humidity data, light intensity data and temperature data, then, under normal circumstances, the causal relationship of the theoretical related data is unidirectional, and the change of temperature data affects and drives the change of humidity data, and the change of light intensity affects and drives the change of temperature data; when the theoretical related data are electrostatic voltage data and particulate matter concentration data in the air, then, under normal circumstances, the causal relationship of the theoretical related data is bidirectional.

[0051] For the temperature data and humidity data of the sub-time period of each collection moment, the Granger causality test is used to process the temperature data and humidity data of the sub-time period of the collection moment. Before testing the causal relationship, the temperature data and humidity data are tested for stationarity. The method of the stationarity test is a well-known technology and is not described in detail in this application. If the temperature data and humidity data are non-stationary data, they are differentiated to be stationary, and then the temperature data and humidity data are tested for whether there is a causal relationship, and the causal relationship is determined to be unidirectional or bidirectional. Specifically:

[0052] First, assume that "temperature data is not the Granger cause of humidity data" and output the p-value. If the p-value is less than the significance level, the null hypothesis is rejected, that is, temperature data is the Granger cause of humidity data, and temperature data is significant to humidity data. Otherwise, temperature data is not the Granger cause of humidity data. Further, assume that "humidity data is not the Granger cause of temperature data" and output the p-value. If the p-value is less than the significance level, the null hypothesis is rejected, that is, humidity data is the Granger cause of temperature data, and humidity data is significant to temperature data. Otherwise, humidity data is not the Granger cause of temperature data. In one embodiment of the present application, the significance level is represented by α, with a value of 0.05. The implementer can select other values ​​based on actual conditions. Among them, the process of outputting the p-value using the Granger causality test is a well-known technology and will not be elaborated in this application.

[0053] Based on the above analysis, when the temperature data is significant to the humidity data, or the humidity data is significant to the temperature data, the temperature data and humidity data have a unidirectional causal relationship; when the temperature data is significant to the humidity data and the humidity data is significant to the temperature data, the temperature data and humidity data have a bidirectional causal relationship; when the temperature data is not significant to the humidity data and the humidity data is not significant to the temperature data, the temperature data and humidity data have no causal relationship.

[0054] Similarly, the same method as that used for temperature and humidity data is used to analyze the causal relationship between other theoretically related data. The causal relationship test flow chart is as follows: Figure 2 As shown in the causal relationship test flow chart, data X and data Y respectively represent two types of environmental data in the theoretical related data.

[0055] The causal coefficient is assigned based on the causal relationship. Specifically, when the theoretical related data are temperature data and humidity data, or light intensity data and temperature data, and the Granger causality test is used to test that the theoretical related data is a unidirectional causal relationship, the causal coefficient of the theoretical related data is assigned to 1, otherwise, the causal coefficient of the theoretical related data is assigned to 0; when the theoretical related data are electrostatic voltage data and air particulate matter concentration data, and the Granger causality test is used to test that the theoretical related data is a bidirectional causal relationship, the causal coefficient of the mutual correlation coefficient is assigned to 1, otherwise, the causal coefficient of the theoretical related data is assigned to 0; and the trend correlation coefficient of the theoretical related data of each sub-time period is combined to construct the causal legitimacy coefficient. The formula is: Where, is the causal justification coefficient of the theoretical relevant data of the sub-time period at each acquisition moment; e is a natural constant, is the trend correlation coefficient of the theoretical correlation data of the sub-time period at each acquisition moment, is the causal coefficient of the theoretical correlation data of the sub-time period at each acquisition moment.

[0056] It should be noted that for the theoretical related data of the sub-time period of each collection moment, when the theoretical related data are temperature data and humidity data, light intensity data and temperature data, and the relationship between the theoretical related data conforms to the trend change and causal relationship under normal circumstances, the corresponding trend correlation coefficient value is larger, and at the same time, the causal relationship is a unidirectional causal relationship. At this time, the value of the causal coefficient obtained is 1, and the larger the value of the causal legitimacy coefficient obtained at this time, otherwise, the smaller the value of the causal legitimacy coefficient obtained; when the theoretical related data are electrostatic voltage data and particulate matter concentration data in the air, and the relationship between the theoretical related data conforms to the trend change and causal relationship under normal circumstances, the corresponding trend correlation coefficient value is larger, and at the same time, the causal relationship is a bidirectional causal relationship. At this time, the larger the value of the causal legitimacy coefficient obtained, otherwise, the smaller the value of the causal legitimacy coefficient obtained.

[0057] At this point, the causal legitimacy coefficients of the theoretically relevant data for each sub-time period are obtained.

[0058] S4, after normalizing the causal legitimacy coefficient, multiply it with the original local reachable density calculated by the LOF anomaly data detection algorithm to generate an updated local reachable density, perform anomaly detection on the environmental data based on the updated local reachable density, and control the environment based on the anomaly detection results.

[0059] Based on the above-mentioned causal legitimacy coefficient, the local reachability density of each data is improved, and the LOF abnormal data detection algorithm is used to process each type of environmental data in the photosensitive drum production process. In one embodiment of the present application, the k value in the LOF abnormal data detection algorithm is 10, and the implementer can select other values ​​based on actual conditions. For each type of environmental data, the causal legitimacy coefficient of all sub-time periods of the acquisition time is normalized. The LOF abnormal data detection algorithm is used to obtain the local reachability density of the environmental data at each acquisition time. The product of the local reachability density and the normalized causal legitimacy coefficient is calculated as the updated local reachability density, and the LOF factor is obtained based on the updated local reachability density. When the causal legitimacy coefficient is larger, the degree of abnormality of the data is smaller, and the value of the obtained local abnormality factor is smaller; conversely, the value of the obtained local abnormality factor is larger. In one embodiment of the present application, the value of the abnormality threshold is 2, and the implementer can select other values ​​based on actual conditions. When the value of the LOF factor of the environmental data is greater than the abnormality threshold, the environmental data is determined to be abnormal data.

[0060] At this point, a more accurate local anomaly factor is obtained based on the causal legitimacy coefficient and the relationship between each type of environmental data, and abnormal data in the environmental data is obtained based on the local anomaly factor. The method of obtaining abnormal data based on the LOF abnormal data detection algorithm is well known and will not be elaborated here. When abnormal environmental data occurs, the environmental parameters are controlled.

[0061] For example, taking electrostatic voltage data as an example, when the LOF abnormal data detection algorithm is used and combined with the updated local reachable density to detect abnormal data of electrostatic voltage, in order to control the environment of the photosensitive drum, an ion fan, ion rod or other equipment is used on the work surface to neutralize the charge. At the same time, check whether the work surface is well grounded, and use an anti-static work table mat to reduce the impact of static electricity. The specific control method is a well-known technology and will not be elaborated in this application.

[0062] An embodiment of the present application also discloses a digital environment control system for photosensitive drum production, including a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, a digital environment control method for photosensitive drum production according to the present application is implemented.

[0063] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and will not be described in detail here.

[0064] The above are all preferred embodiments of the present application, and are not intended to limit the scope of protection of the present application. Therefore, any equivalent changes made based on the structure, shape, and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A digital environment control method for photosensitive drum production, characterized in that: The method comprises the following steps: collecting environmental data during the production process of the photosensitive drum, wherein the environmental data includes light intensity data, temperature data, relative humidity data, concentration of particulate matter in the air, and electrostatic voltage data; Analyze the correlation between environmental data to obtain each pair of theoretically relevant data; Calculating the trend correlation coefficient based on the difference in data distribution patterns between environmental data in the theoretical correlation data, including: for each pair of theoretical correlation data in each sub-time period, calculating the skewness and kurtosis of each data in each pair of theoretical correlation data; dividing the theoretical correlation data into theoretical correlation data with negative correlation change relationships and theoretical correlation data with positive correlation change relationships, specifically: the theoretical correlation data with negative correlation change relationships are temperature data and humidity data, and the theoretical correlation data with positive correlation change relationships are electrostatic voltage data and particulate matter concentration data, light intensity data and temperature data; based on the change relationship of the theoretical correlation data, applying different calculation logics to obtain the trend correlation coefficient; The calculation formula of the trend correlation coefficient is: , is the trend correlation coefficient of the theoretical correlation data of each sub-time period; are the normalized skewness and kurtosis of any kind of theoretical related data in each sub-time period, is the normalized skewness and kurtosis of another data in the theoretical related data of each sub-time period; The Granger causality test was used to obtain the causal relationship of theoretically relevant data and determine the causal coefficient. The assignment rule is that when the temperature and humidity data or the light and temperature data show a unidirectional causal relationship, and the electrostatic voltage data and the air particulate matter concentration data show a bidirectional causal relationship, the causal coefficient is assigned a value of 1, otherwise it is 0; The sum of the trend correlation coefficient and the causal coefficient is used as an exponent and an exponential operation is performed with a natural constant as the base to construct a causal legitimacy coefficient; after the causal legitimacy coefficient is normalized, it is multiplied with the original local reachable density calculated by the LOF anomaly data detection algorithm to generate an updated local reachable density, anomaly detection is performed on the environmental data based on the updated local reachable density, and the environment is controlled based on the anomaly detection result.

2. The digital environment control method for photosensitive drum production according to claim 1, characterized in that: The method for acquiring the theoretical related data is: taking temperature data and humidity data, electrostatic voltage data and air particle concentration data, and light intensity data and temperature data as a pair of theoretical related data.

3. The digital environment control method for photosensitive drum production according to claim 1, characterized in that: The Granger causality test is used to obtain the causal relationship of the theoretically relevant data, including: for each type of theoretically relevant data, using the causal relationship test to obtain the significant relationship between the theoretically relevant data; For temperature data and humidity data, when temperature data is significant to humidity data, or humidity data is significant to temperature data, there is a one-way causal relationship between temperature data and humidity data; when temperature data is significant to humidity data and humidity data is significant to temperature data, there is a two-way causal relationship between temperature data and humidity data; when temperature data is not significant to humidity data and humidity data is not significant to temperature data, there is no causal relationship between temperature data and humidity data; similarly, the same method as temperature data and humidity data is used to obtain the causal relationship of other theoretically related data.

4. The digital environment control method for photosensitive drum production according to claim 3, characterized in that: The use of causal relationship test to obtain the significant relationship between theoretically related data includes: assuming that the temperature data is not the cause of the humidity data, outputting a p-value, and if the p-value is less than the significance level, rejecting the null hypothesis, that is, the temperature data is the cause of the humidity data, and the temperature data is significant to the humidity data; conversely, the temperature data is not the cause of the humidity data; Assume that humidity data is not the cause of temperature data and output the p-value. If the p-value is less than the significance level, reject the null hypothesis, that is, humidity data is the cause of temperature data and humidity data is significant to temperature data; on the contrary, humidity data is not the cause of temperature data.

5. The digital environment control method for photosensitive drum production according to claim 1, characterized in that: Before adopting the Granger causality test, the method further includes: performing a stationarity test on the theoretical related data, and when the theoretical related data is non-stationary data, performing differential processing on the theoretical related data to obtain stationary data.

6. The digital environment control method for photosensitive drum production according to claim 1, characterized in that: The analysis, calculation and construction steps are all performed based on the environmental data in a sub-time period of a preset length before each collection moment.

7. A digital environmental control system for photosensitive drum production, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a digital environment control method for photosensitive drum production according to any one of claims 1 to 6 is implemented.

Images