A factory data intelligent acquisition and monitoring method based on an internet of things

Abstract

The application relates to the technical field of data processing, in particular to a factory data intelligent acquisition and monitoring method based on the Internet of Things. The method acquires monitoring data of any monitoring index in a factory in a day and historical monitoring data of any monitoring index in a historical day, and forms a monitoring data sequence by monitoring data in a preset time range. The method acquires preliminary suspected abnormal data in the day and historical preliminary suspected abnormal data in the historical day, acquires suspected abnormal data according to the time difference between the preliminary suspected abnormal data and the historical preliminary suspected abnormal data and the time interval regularity characteristics among the preliminary suspected abnormal data, acquires the abnormal risk degree of the monitoring data sequence according to the distribution characteristics of the suspected abnormal data in the monitoring data sequence, and further acquires the data acquisition frequency of any monitoring index in a future preset time range. The method acquires the data acquisition frequency of each monitoring index in the future preset time range, and improves the dynamic acquisition efficiency of the factory monitoring data in the future preset time range.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method for intelligent data acquisition and monitoring of factories based on the Internet of Things. Background Technology

[0002] Currently, IoT-based factory data acquisition and monitoring typically relies on an IoT architecture and multi-dimensional sensors. This involves deploying various types of sensing nodes across key factory equipment, the production environment, and processes to achieve automated, continuous, and highly reliable data acquisition. The acquired data is then processed in real-time, with anomaly detection and status assessment performed to enable intelligent monitoring and early warning of the factory's operational status.

[0003] Existing technologies typically use a fixed frequency to collect and monitor data for each type of data. This method has obvious limitations. A high collection frequency can easily cause pressure on the IoT network and insufficient storage space, while a low collection frequency may lead to missed or false detections of abnormal data. Furthermore, during the operation of equipment in the factory, different production processes and adjustments can lead to changes in the factory environment and equipment parameters. A fixed frequency cannot adapt to the needs of all scenarios, resulting in redundant data overflow or insensitivity to abnormal data in factory data collection at a fixed collection frequency.

[0004] Therefore, how to adaptively adjust the collection frequency of factory monitoring data and improve the accuracy of work data monitoring and system operating efficiency has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide an intelligent data acquisition and monitoring method for factories based on the Internet of Things, in order to solve the problem of how to adaptively adjust the acquisition frequency of factory monitoring data and improve the accuracy of work data monitoring and system operating efficiency.

[0006] This invention provides a method for intelligent data acquisition and monitoring of factories based on the Internet of Things (IoT), which includes the following steps: During factory operation, for any monitoring indicator in the factory, the monitoring data of any monitoring indicator up to the current monitoring time within the day is obtained, as well as the historical monitoring data within a preset historical day before the current day, and the monitoring data within the preset time range are combined into a monitoring data sequence. Based on the data fluctuation characteristics of the monitoring data within the day, obtain the preliminary suspected anomaly data within the day, obtain the historical preliminary suspected anomaly data within the preset historical days, and obtain the suspected anomaly data within the day based on the time difference between the preliminary suspected anomaly data within the day and the historical preliminary suspected anomaly data within the preset historical days, as well as the time interval pattern characteristics between the preliminary suspected anomaly data within the day. Based on the distribution characteristics of suspected abnormal data in the monitoring data sequence, the degree of abnormal risk of the monitoring data sequence is obtained; based on the degree of abnormal risk of the monitoring data sequence, the data collection frequency of any monitoring indicator within a future preset time range is obtained. The data collection frequency of each monitoring indicator is obtained within a preset time range in the future. Using the data collection frequency of each monitoring indicator within a preset time range in the future, the factory is intelligently monitored and collected for data within a preset time range in the future.

[0007] Preferably, the step of obtaining preliminary suspected abnormal data for the day based on the data fluctuation characteristics of the monitoring data within the day includes: For any monitoring data within the same day except for the last monitoring data, obtain the absolute value of the difference between the any monitoring data and the next monitoring data of the same monitoring data to obtain the change difference value of the any monitoring data; Obtain the change difference value of each monitoring data except the last monitoring data, obtain the maximum change difference value and the minimum change difference value of any monitoring indicator, and obtain the preliminary suspected abnormal data for the day based on the change difference value of each monitoring data except the last monitoring data, the degree of difference between the maximum change difference value and the minimum change difference value, and the difference between the change difference values ​​of each monitoring data except the last monitoring data.

[0008] Preferably, the step of obtaining preliminary suspected abnormal data for the day based on the change difference value of each monitoring data other than the last monitoring data, the degree of difference between the maximum change difference value and the minimum change difference value, and the difference between the change difference values ​​of each monitoring data other than the last monitoring data, includes: For any monitoring data within the day, excluding the last two monitoring data, the difference between the change difference value of the any monitoring data and the minimum change difference value is obtained and recorded as the first difference value. The difference between the maximum change difference value and the minimum change difference value is obtained and recorded as the second difference value. The ratio between the first difference value and the second difference value is obtained to obtain the first degree of change of the any monitoring data. Obtain the absolute value of the difference between any monitoring data and the previous monitoring data, and record it as the previous degree of change difference. Obtain the absolute value of the difference between any monitoring data and the next monitoring data, and record it as the subsequent degree of change difference. Normalize the sum of the previous degree of change difference and the subsequent degree of change difference to obtain the second degree of change of any monitoring data. The sum of the first mutation degree and the second mutation degree is obtained to obtain the preliminary suspected abnormality degree of any of the monitoring data; Obtain the preliminary suspected anomaly level of each monitoring data point within the day, excluding the last two monitoring data points. Based on the preliminary suspected anomaly level of each monitoring data point within the day, excluding the last two monitoring data points, obtain the preliminary suspected anomaly data for the day.

[0009] Preferably, the step of obtaining preliminary suspected anomaly data for the day based on the preliminary suspected anomaly level of each monitoring data point excluding the last two monitoring data points includes: For any monitoring data other than the last two monitoring data points within the same day, if the preliminary suspected abnormality level of any monitoring data is greater than or equal to the preset preliminary suspected abnormality level threshold, then the monitoring data is confirmed as preliminary suspected abnormal data.

[0010] Preferably, the step of obtaining suspected anomaly data for the current day based on the time difference between the preliminary suspected anomaly data within the current day and the historical preliminary suspected anomaly data within a preset historical period, as well as the time interval patterns between the preliminary suspected anomaly data within the current day, includes: Any preliminary suspected anomaly data within the same day, excluding the first and last preliminary suspected anomaly data, is recorded as the target data. For any historical day within a preset historical day, a historical monitoring time that is the same as the monitoring time of the target data is obtained within that historical day and is recorded as the target historical monitoring time. Reference data of the target data is obtained within that historical day, and the interval between the target historical monitoring time and the historical monitoring time of the reference data is obtained is recorded as the time difference value between the target data and the reference data within that historical day. Obtain the time difference value between the target data and the reference data for each historical day, and obtain the mean time difference value. Normalize the mean time difference value to obtain the first suspected anomaly degree of the target data. The time interval between the target data and its previous preliminary suspected abnormal data is obtained and recorded as the previous interval time. The time interval between the target data and its next preliminary suspected abnormal data is obtained and recorded as the next interval time. The absolute value of the difference between the previous interval time and the next interval time is normalized to obtain the second suspected abnormality level of the target data. The first suspected anomaly level and the second suspected anomaly level of the target data are added together to obtain the suspected anomaly level of the target data; Obtain the suspected anomaly level for each target data point, and based on the suspected anomaly level for each target data point, obtain the suspected anomaly data for the day.

[0011] Preferably, the step of obtaining suspected abnormal data for the day based on the suspected abnormality level of each target data includes: For any target data, if the suspected abnormality level of the target data is greater than or equal to a preset suspected abnormality level threshold, then the target data is confirmed as suspected abnormal data.

[0012] Preferably, the reference data for obtaining the target data within any historical day includes: Obtain the interval between the historical monitoring time of each preliminary suspected anomaly data within any historical day and the target historical monitoring time, and record the preliminary suspected anomaly data corresponding to the minimum interval time as the reference data of the target data.

[0013] Preferably, obtaining the degree of abnormal risk of the monitoring data sequence based on the distribution characteristics of suspected abnormal data in the monitoring data sequence includes: The number of suspected abnormal data in the monitoring data sequence is obtained, and the number of suspected abnormal data in the monitoring data sequence is normalized to obtain the first abnormal risk level. For any suspected abnormal data in the monitoring data sequence, the minimum time interval between any suspected abnormal data and every other suspected abnormal data is obtained, and recorded as the nearest abnormal time distance of any suspected abnormal data. The nearest abnormal time distance of each suspected abnormal data in the monitoring data sequence is obtained, and the cumulative value of the nearest abnormal time distance is obtained. The reciprocal of the cumulative value of the nearest abnormal time distance is normalized to obtain the second abnormal risk level. The sum of the first abnormal risk level and the second abnormal risk level is obtained to obtain the abnormal risk level of the monitoring data sequence.

[0014] Preferably, the step of obtaining the data collection frequency of any monitoring indicator within a preset time range based on the degree of abnormal risk of the monitoring data sequence includes: The data collection frequency of any monitoring indicator within a preset time range is obtained, and the product of the abnormal risk level of the monitoring data sequence and the data collection frequency of any monitoring indicator within the preset time range is rounded up to obtain the data collection frequency of any monitoring indicator within a future preset time range.

[0015] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows: In this invention, preliminary suspected abnormal data is initially screened based on the data fluctuation characteristics of the monitoring data within the day. To prevent large fluctuations in environmental monitoring data or equipment parameters caused by changes in fixed processes of factory equipment from being falsely detected as abnormal data, more accurate suspected abnormal data is further obtained based on the historical temporal similarity and spatial interval regularity of the preliminary suspected abnormal data. Based on the distribution characteristics of suspected abnormal data in the monitoring data sequence, the degree of abnormal risk of the monitoring data sequence is obtained, and the degree of abnormal risk within a preset time range is assessed. Then, based on the degree of abnormal risk, the data acquisition frequency of any monitoring indicator is adaptively adjusted so that the data acquisition frequency matches the data change characteristics. While ensuring that key abnormal features are not lost, the growth of invalid data caused by sensor noise and environmental interference is effectively suppressed, thereby reducing the false alarm rate and improving the accuracy of data monitoring and system operating efficiency. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of a factory data intelligent acquisition and monitoring method based on the Internet of Things provided in Embodiment 1 of the present invention. Detailed Implementation

[0018] Embodiments of this disclosure are described in detail below, with examples of these embodiments illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting it.

[0019] It should be noted that the terms "first," "second," etc., used in this disclosure and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure.

[0020] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0021] See Figure 1This is a flowchart of a factory data intelligent acquisition and monitoring method based on the Internet of Things provided in Embodiment 1 of the present invention. Figure 1 As shown, the method may include: Step S101: During factory operation, for any monitoring indicator in the factory, acquire the monitoring data of any monitoring indicator up to the current monitoring time within the day, as well as the historical monitoring data within a preset historical day before the current day, and form a monitoring data sequence from the monitoring data within the preset time range.

[0022] When using IoT systems to monitor and collect various types of data in a factory, a fixed collection frequency can easily lead to missed anomalies during low-frequency collection and network and storage pressure during high-frequency collection. This results in low efficiency of IoT system data collection and monitoring, and does not meet the factory's real-time needs for process equipment adjustments and environmental changes.

[0023] Therefore, in this embodiment of the invention, suspected abnormal data is first screened by analyzing the characteristics of historical data. Then, the degree of abnormal risk in the current period is assessed based on the distribution characteristics of suspected abnormal data in the current period. Subsequently, the data collection frequency of the monitoring indicators is adaptively adjusted based on the degree of abnormal risk, so that the data collection frequency matches the data change characteristics. While ensuring that key abnormal features are not lost, the growth of invalid data caused by sensor noise and environmental interference is effectively suppressed, thereby reducing the false alarm rate and improving the accuracy of data monitoring and the efficiency of system operation.

[0024] During factory operation, sensor nodes deployed at key locations in factory equipment, production lines, and the environment collect monitoring data for each indicator. These indicators include equipment operating parameters and the factory environment. Equipment operating parameter monitoring data includes current, voltage, speed, and vibration data, while factory environment monitoring data includes temperature, humidity, dust concentration, and gas concentration data. The collected data is transmitted in real time to the factory data monitoring and processing terminal via the Internet of Things. By processing the collected data in real time, identifying anomalies, and assessing the status, intelligent monitoring and early warning of the factory's operating status are achieved.

[0025] Because different production processes and adjustments during equipment operation within the factory can lead to changes in the factory environment and equipment parameters, this embodiment differentiates the data collection frequency for each type of monitoring data. Specifically, the data collection frequency for each monitoring indicator is adjusted individually, while the adjustment method is the same for each indicator. Taking any given monitoring indicator as an example, this embodiment acquires the monitoring data for each monitoring moment up to the current monitoring moment, as well as historical monitoring data from the previous preset historical days. The monitoring data within the preset time range are combined into a monitoring data sequence for subsequent analysis and adjustment of the data collection frequency for any given monitoring indicator. Since the data collection frequency for each monitoring indicator is adjusted in real time, this embodiment collects data for each indicator according to its real-time data collection frequency. In this embodiment, the adjustment cycle for the data collection frequency is set to 10 minutes, meaning the data collection frequency is adjusted every 10 minutes. This is not a limitation and can be set according to the specific implementation scenario. In order to better adjust and analyze the data collection frequency of any monitoring indicator, the preset time range in this embodiment is set to include 10 minutes of the current monitoring time, that is, the current adjustment cycle. The preset historical days are set to 5 days before the current day. There is no restriction here, and it can be set according to the specific implementation scenario.

[0026] Step S102: Based on the data fluctuation characteristics of the monitoring data within the day, obtain the preliminary suspected abnormal data within the day, obtain the historical preliminary suspected abnormal data within the preset historical days, and obtain the suspected abnormal data within the day based on the time difference between the preliminary suspected abnormal data within the day and the historical preliminary suspected abnormal data within the preset historical days, as well as the time interval pattern characteristics between the preliminary suspected abnormal data within the day.

[0027] Since abnormal data is usually accompanied by fluctuations that differ from those of regular data, this embodiment obtains preliminary suspected abnormal data for the day based on the data fluctuation characteristics of the monitoring data within the day. This data is then used to analyze the degree of abnormal risk in the current monitoring period and adjust the data collection frequency accordingly.

[0028] The method for obtaining preliminary suspected abnormal data for the day based on the data fluctuation characteristics of the monitoring data is as follows: (1) Obtain the variation value of each monitoring data except the last monitoring data.

[0029] For any monitoring data point within the day, excluding the last monitoring data point, obtain the absolute value of the difference between that monitoring data point and the next monitoring data point. This yields the change difference value for that monitoring data point. Taking the q-th monitoring data point within the day (which is not the last monitoring data point of the day) as an example, the change difference value for the q-th monitoring data point is denoted as... ,but ,in, This is the qth monitoring data point of the day. This refers to the (q+1)th monitoring data point within the day (i.e., the next monitoring data point after the qth monitoring data point within the day). The absolute value symbol is used. Similarly, the change difference value of each monitoring data point except the last monitoring data point is obtained.

[0030] (2) Obtain the preliminary suspected abnormality level of each monitoring data except the last two monitoring data within the day.

[0031] The greater the difference in the value of the change in the monitoring data relative to the overall monitoring data, the stronger the fluctuation characteristics of the monitoring data relative to the overall monitoring data. At the same time, the greater the difference in the value of the change in the monitoring data with its adjacent monitoring data, the greater the fluctuation amplitude of the monitoring data, that is, the more specific the abrupt change characteristics.

[0032] Therefore, in this embodiment, the maximum value of the change difference of any monitoring indicator within a historical month is obtained as the maximum change difference value of the monitoring indicator, and the minimum value of the change difference of any monitoring indicator within a historical month is obtained as the minimum change difference value of the monitoring indicator. Based on the change difference value of each monitoring data except the last monitoring data, the degree of difference between the maximum change difference value and the minimum change difference value, and the difference between the change difference values ​​of each monitoring data except the last monitoring data, the preliminary suspected abnormality degree of each monitoring data except the last two monitoring data within the day is obtained, which is used to obtain the preliminary suspected abnormal data within the day.

[0033] For any monitoring data point other than the last two monitoring data points within the same day, the method for obtaining the preliminary suspected anomaly level of that monitoring data point is as follows: The difference between the change difference value of any monitoring data and the minimum change difference value is obtained and recorded as the first difference value. The difference between the maximum change difference value and the minimum change difference value is obtained and recorded as the second difference value. The ratio between the first difference value and the second difference value is obtained to obtain the first degree of change of any monitoring data. Obtain the absolute value of the difference between any monitoring data and the previous monitoring data, and record it as the previous degree of change difference. Obtain the absolute value of the difference between any monitoring data and the next monitoring data, and record it as the subsequent degree of change difference. Normalize the sum of the previous degree of change difference and the subsequent degree of change difference to obtain the second degree of change of any monitoring data. The sum of the first mutation degree and the second mutation degree is obtained to obtain the preliminary suspected abnormality degree of any of the monitoring data.

[0034] In one implementation, taking the q-th monitoring data point of the day (which is not the last two monitoring data points of the day) as an example, the formula for calculating the preliminary suspected anomaly level of the q-th monitoring data point is as follows: ;

[0035] in, The preliminary suspected anomaly level of the q-th monitoring data; This represents the difference in value of the q-th monitoring data. The value representing the maximum difference in change; The minimum variation value; This represents the change difference value of the (q-1)th monitoring data. This represents the change difference value of the (q+1)th monitoring data. It is the absolute value symbol; It is a normalization function for maximum and minimum values.

[0036] It should be noted that, The degree of the first mutation of the q-th monitoring data. The larger the value, the stronger the fluctuation characteristic of the q-th monitoring data, and the more likely it is to be an anomaly with a relatively large fluctuation range. The larger it is; The degree of mutation for the q-th monitoring data point. The larger the value, the greater the difference between the q-th monitoring data and its adjacent monitoring data. This indicates that the q-th monitoring data has more abrupt change characteristics and is more likely to be an anomaly with large fluctuations. The larger it is.

[0037] Similarly, obtain the preliminary suspected anomaly level for each monitoring data point within the day, excluding the last two monitoring data points.

[0038] (3) Based on the preliminary suspected abnormality level of each monitoring data except the last two monitoring data within the day, obtain the preliminary suspected abnormality data within the day.

[0039] The higher the initial suspected anomaly level, the more likely the monitoring data is to be anomalous. Therefore, for any monitoring data within the day, excluding the last two monitoring data points, if the initial suspected anomaly level of any monitoring data is greater than or equal to the preset initial suspected anomaly level threshold, then the monitoring data is confirmed as initially suspected anomalous data. The value range of the initial suspected anomaly level is [0, 2]. After calculation using the above formula for the initial suspected anomaly level, the initial suspected anomaly levels of monitoring data with anomalous characteristics and normal data will be divided into the two ends of the value range. That is, the initial suspected anomaly level of monitoring data with anomalous characteristics will be close to 2, and the initial suspected anomaly level of normal data without anomalous characteristics will be close to 0. Therefore, in this embodiment, the preset initial suspected anomaly level threshold is set to the middle value of 1 in the value range of the initial suspected anomaly level. This is not limited here and can be set according to the specific implementation scenario.

[0040] Since changes in fixed processes of factory equipment may also cause significant fluctuations in environmental data or equipment parameters, the preliminary suspected anomaly data may also contain normal monitoring data generated by changes in fixed processes of factory equipment. Therefore, it is necessary to further analyze the obtained preliminary suspected anomaly data to obtain more accurate suspected anomaly data.

[0041] The method for obtaining suspected abnormal data is as follows: (1) Obtain the degree of suspected abnormality of preliminary suspected abnormal data.

[0042] Considering that changes in fixed processes of factory equipment usually occur at fixed nodes (i.e., within a fixed time range) and are regular, this embodiment obtains preliminary suspected abnormal data within a preset historical period according to the above-mentioned method for obtaining suspected abnormal data. Based on the time difference between the preliminary suspected abnormal data within the current day and the preliminary suspected abnormal data within the preset historical period, as well as the regularity of the time interval between the preliminary suspected abnormal data within the current day, the degree of suspected abnormality of the preliminary suspected abnormal data is obtained, which is used to obtain suspected abnormal data.

[0043] The method for obtaining the degree of suspected anomaly is as follows: Any preliminary suspected anomaly data within the same day, excluding the first and last preliminary suspected anomaly data, is recorded as the target data. For any historical day within a preset historical period, a historical monitoring time that is the same as the monitoring time of the target data is obtained within that historical period and recorded as the target historical monitoring time (e.g., if the monitoring time of the target data is 12:00:00, then 12:00:00 within that historical period is recorded as the target historical monitoring time). The interval between the historical monitoring time of each preliminary suspected anomaly data within that historical period and the target historical monitoring time is obtained respectively. The preliminary suspected anomaly data corresponding to the minimum interval time is recorded as the reference data of the target data (if the minimum interval time corresponds to multiple preliminary suspected anomaly data, i.e., there are two or more preliminary suspected anomaly data within that historical period with the same interval time as the target historical monitoring time, all of which are minimum interval times, then any preliminary suspected anomaly data corresponding to the minimum interval time is recorded as the reference data of the target data). The interval between the target historical monitoring time and the historical monitoring time of the reference data is obtained and recorded as the time difference value between the target data and the reference data within that historical period; Obtain the time difference value between the target data and the reference data for each historical day, and obtain the mean time difference value. Normalize the mean time difference value to obtain the first suspected anomaly degree of the target data. The time interval between the target data and its previous preliminary suspected abnormal data is obtained and recorded as the previous interval time. The time interval between the target data and its next preliminary suspected abnormal data is obtained and recorded as the next interval time. The absolute value of the difference between the previous interval time and the next interval time is normalized to obtain the second suspected abnormality level of the target data. The first suspected anomaly level and the second suspected anomaly level of the target data are added together to obtain the suspected anomaly level of the target data.

[0044] In one implementation, taking the p-th target data as an example, the formula for calculating the suspected anomaly level of the p-th target data is: ; in, The degree of suspected anomaly in the p-th target data; Let p be the time difference between the p-th target data and the reference data within the i-th historical day; The number of days in the preset historical period; The interval is the time between the p-th target data and its previous preliminary suspected anomaly data. This is the interval time, which is the time between the p-th target data and its next preliminary suspected anomaly data. It is the absolute value symbol; This is the normalization function.

[0045] It should be noted that, Let the first suspected anomaly level be the p-th target data. The mean of time differences. The smaller the value, the closer the temporal position of the initially suspected anomaly data in the historical data is to the target data, and the stronger the historical time similarity. In other words, the p-th target data is more likely to have the same factory process characteristics and show fixed changes at fixed nodes every day. The smaller it is, the more... The smaller it is; The second suspected anomaly level of the p-th target data. The smaller the value, the higher the similarity of the intervals between the initial suspected anomalies on the left and right sides of the p-th target data. This indicates that the p-th target data is more likely to be an initial suspected anomaly caused by regular changes in fixed processes of factory equipment, and the stronger the regularity of its spatial intervals. The smaller it is, the more... The smaller it is.

[0046] Similarly, the degree of suspected anomalies for each target data is obtained.

[0047] (2) Based on the degree of suspected abnormality of the preliminary suspected abnormal data, obtain suspected abnormal data.

[0048] Specifically, for any target data, if the suspected anomaly level of the target data is greater than or equal to a preset suspected anomaly level threshold, then the target data is confirmed as suspected anomaly data. In this embodiment, the preset suspected anomaly level threshold is set to the middle boundary value of 1 in the range of suspected anomaly levels. This is not limited here and can be set according to the specific implementation scenario.

[0049] At this point, we obtained the suspected abnormal data for the day.

[0050] Step S103: Based on the distribution characteristics of suspected abnormal data in the monitoring data sequence, obtain the degree of abnormal risk of the monitoring data sequence; based on the degree of abnormal risk of the monitoring data sequence, obtain the data collection frequency of any monitoring indicator within a future preset time range.

[0051] Because equipment operation generates heat and heat, personnel movement, ventilation, and changes in the properties of raw materials can all lead to changes in temperature and humidity data, local emissions and uneven air flow can also affect gas concentration changes. Vibration during equipment operation and micro-oscillations generated by PID self-adjustment can also cause some fluctuations. Therefore, normal environmental monitoring data and equipment parameters in a factory may have some fluctuations. Such normal fluctuations are in a discrete and random distribution state and cannot indicate the presence of abnormal risks. However, when there is a potential risk trend, abnormal data will appear to aggregate and increase.

[0052] Therefore, in this embodiment, based on the distribution characteristics of suspected abnormal data in the monitoring data sequence, the degree of abnormal risk of the monitoring data sequence (that is, the degree of abnormal risk in the current adjustment period) is obtained, and then based on the degree of abnormal risk of the monitoring data sequence, the data collection frequency of any monitoring indicator within a future preset time range is obtained.

[0053] The method for obtaining the degree of anomaly risk of the monitoring data sequence based on the distribution characteristics of suspected abnormal data in the monitoring data sequence is as follows: The number of suspected abnormal data in the monitoring data sequence is obtained, and the number of suspected abnormal data in the monitoring data sequence is normalized to obtain the first abnormal risk level. For any suspected abnormal data in the monitoring data sequence, the minimum time interval between any suspected abnormal data and every other suspected abnormal data is obtained, and recorded as the nearest abnormal time distance of any suspected abnormal data. The nearest abnormal time distance of each suspected abnormal data in the monitoring data sequence is obtained, and the cumulative value of the nearest abnormal time distance is obtained. The reciprocal of the cumulative value of the nearest abnormal time distance is normalized to obtain the second abnormal risk level. The sum of the first abnormal risk level and the second abnormal risk level is obtained to obtain the abnormal risk level of the monitoring data sequence.

[0054] In one embodiment, the formula for calculating the degree of anomaly risk of the monitored data sequence is: ; Where E represents the degree of abnormal risk in the monitored data sequence; To monitor the number of suspected abnormal data in the data sequence; To monitor the closest time distance to the j-th suspected abnormal data in the data sequence; It is a normalization function for maximum and minimum values.

[0055] It should be noted that, The first level of abnormal risk. The larger the value, the more suspected abnormal data there are within the preset time range of the monitored data sequence, and the stronger the possibility of anomaly risk within that preset time range. The larger it is, the larger E will be; The second level of abnormal risk. This is the cumulative value based on the distance from the most recent anomaly. The smaller the value, the denser the distribution of suspected abnormal data within the preset time range of the monitored data sequence, and the stronger the possibility of anomaly risk within that preset time range. The larger it is, the larger E will be.

[0056] Furthermore, based on the degree of abnormal risk of the monitored data sequence, the method for obtaining the data collection frequency of any monitoring indicator within a future preset time range is as follows: The data collection frequency of any monitoring indicator within a preset time range is obtained, and the product of the abnormal risk level of the monitoring data sequence and the data collection frequency of any monitoring indicator within the preset time range is rounded up to obtain the data collection frequency of any monitoring indicator within a future preset time range.

[0057] In one embodiment, the formula for calculating the data collection frequency of any monitoring indicator within a preset future time range is as follows: ; in, The data collection frequency for any of the monitoring indicators within a preset time range in the future; The data collection frequency for any of the monitoring indicators within a preset time range; E represents the degree of abnormal risk in the monitoring data sequence; The rounding up symbol.

[0058] It should be noted that the larger the value of E, the stronger the probability of anomalies within the preset time range of the monitored data sequence. This means a higher likelihood of anomalies occurring during subsequent monitoring, and consequently, a higher probability of anomalies within the future preset time range. To prevent missed or false detections of abnormal data, the frequency of subsequent data collection should be increased. The larger E is, the lower the probability of anomalies within the preset time range of the monitored data sequence. This means a lower probability of anomalies in subsequent monitoring, and consequently, a lower probability of anomalies within the future preset time range. To reduce network and storage pressure on high-frequency monitoring data, the frequency of subsequent data collection should be reduced. The smaller it is.

[0059] Thus, the data collection frequency of any of the monitoring indicators within a future preset time range is obtained.

[0060] Step S104: Obtain the data collection frequency of each monitoring indicator within a future preset time range, and use the data collection frequency of each monitoring indicator within a future preset time range to conduct intelligent data collection and monitoring of the factory within a future preset time range.

[0061] According to the method described above for obtaining the data acquisition frequency of any monitoring indicator within a future preset time range, the data acquisition frequency of each monitoring indicator within the future preset time range is obtained. Then, using the data acquisition frequency of each monitoring indicator within the future preset time range, intelligent data acquisition and monitoring of the factory is carried out within the future preset time range. This ensures that the data acquisition frequency matches the data change characteristics, effectively suppressing the growth of invalid data caused by sensor noise and environmental interference while ensuring that key abnormal features are not lost. This reduces the false alarm rate and improves the accuracy of data monitoring and the efficiency of system operation.

[0062] It is worth noting that the key point of this embodiment of the invention is to obtain the data collection frequency of the monitoring indicators within a future preset time range based on the degree of abnormal risk of the monitoring data sequence (i.e., the degree of abnormal risk within a preset time range); using the data collection frequency of each monitoring indicator within a future preset time range to intelligently collect and monitor the factory's data within a future preset time range is existing technology and will not be elaborated here.

[0063] In summary, in this embodiment of the invention, preliminary suspected abnormal data for the day are initially screened based on the data fluctuation characteristics of the monitoring data within the day. To prevent large fluctuations in environmental monitoring data or equipment parameters caused by changes in fixed processes of factory equipment from being falsely detected as abnormal data, more accurate suspected abnormal data are further obtained based on the historical temporal similarity and spatial interval regularity of the preliminary suspected abnormal data. Based on the distribution characteristics of suspected abnormal data in the monitoring data sequence, the degree of abnormal risk of the monitoring data sequence is obtained, and the degree of abnormal risk within a preset time range is assessed. Then, the data acquisition frequency of any monitoring indicator is adaptively adjusted based on the degree of abnormal risk, so that the data acquisition frequency matches the data change characteristics. While ensuring that key abnormal features are not lost, the growth of invalid data caused by sensor noise and environmental interference is effectively suppressed, thereby reducing the false alarm rate and improving the accuracy of data monitoring and system operating efficiency.

[0064] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A factory data intelligent acquisition and monitoring method based on the Internet of Things, characterized in that, The IoT-based intelligent data acquisition and monitoring method for factories includes: During factory operation, for any monitoring indicator in the factory, the monitoring data of any monitoring indicator up to the current monitoring time within the day is obtained, as well as the historical monitoring data within a preset historical day before the current day, and the monitoring data within the preset time range are combined into a monitoring data sequence. Based on the data fluctuation characteristics of the monitoring data within the day, obtain the preliminary suspected anomaly data within the day, obtain the historical preliminary suspected anomaly data within the preset historical days, and obtain the suspected anomaly data within the day based on the time difference between the preliminary suspected anomaly data within the day and the historical preliminary suspected anomaly data within the preset historical days, as well as the time interval pattern characteristics between the preliminary suspected anomaly data within the day. Based on the distribution characteristics of suspected abnormal data in the monitoring data sequence, the degree of abnormal risk of the monitoring data sequence is obtained; based on the degree of abnormal risk of the monitoring data sequence, the data collection frequency of any monitoring indicator within a future preset time range is obtained. Obtain the data collection frequency of each monitoring indicator within a future preset time range, and use the data collection frequency of each monitoring indicator within a future preset time range to intelligently collect and monitor the factory's data within a future preset time range. The step of obtaining suspected anomaly data for the current day based on the time difference between preliminary suspected anomaly data within the current day and historical preliminary suspected anomaly data within a preset historical period, as well as the regularity of time intervals between preliminary suspected anomaly data within the current day, includes: Any preliminary suspected anomaly data within the same day, excluding the first and last preliminary suspected anomaly data, is recorded as the target data. For any historical day within a preset historical day, a historical monitoring time that is the same as the monitoring time of the target data is obtained within that historical day and is recorded as the target historical monitoring time. Reference data of the target data is obtained within that historical day, and the interval between the target historical monitoring time and the historical monitoring time of the reference data is obtained is recorded as the time difference value between the target data and the reference data within that historical day. Obtain the time difference value between the target data and the reference data for each historical day, and obtain the mean time difference value. Normalize the mean time difference value to obtain the first suspected anomaly degree of the target data. The time interval between the target data and its previous preliminary suspected abnormal data is obtained and recorded as the previous interval time. The time interval between the target data and its next preliminary suspected abnormal data is obtained and recorded as the next interval time. The absolute value of the difference between the previous interval time and the next interval time is normalized to obtain the second suspected abnormality level of the target data. The first suspected anomaly level and the second suspected anomaly level of the target data are added together to obtain the suspected anomaly level of the target data; Obtain the suspected anomaly level for each target data, and based on the suspected anomaly level for each target data, obtain the suspected anomaly data for the day; The reference data for obtaining the target data within any historical day includes: Obtain the interval between the historical monitoring time of each preliminary suspected anomaly data within any historical day and the target historical monitoring time, and record the preliminary suspected anomaly data corresponding to the minimum interval time as the reference data of the target data.

2. The method for intelligent data acquisition and monitoring of factories based on the Internet of Things according to claim 1, characterized in that, The step of obtaining preliminary suspected abnormal data for the day based on the data fluctuation characteristics of the monitoring data includes: For any monitoring data within the same day except for the last monitoring data, obtain the absolute value of the difference between the any monitoring data and the next monitoring data of the same monitoring data to obtain the change difference value of the any monitoring data; Obtain the change difference value of each monitoring data except the last monitoring data, obtain the maximum change difference value and the minimum change difference value of any monitoring indicator, and obtain the preliminary suspected abnormal data for the day based on the change difference value of each monitoring data except the last monitoring data, the degree of difference between the maximum change difference value and the minimum change difference value, and the difference between the change difference values ​​of each monitoring data except the last monitoring data.

3. The method for intelligent data acquisition and monitoring of factories based on the Internet of Things according to claim 2, characterized in that, The preliminary suspected abnormal data for the day is obtained based on the variation difference value of each monitoring data except the last monitoring data, the degree of difference between the maximum variation difference value and the minimum variation difference value, and the difference between the variation difference values ​​of each monitoring data except the last monitoring data, including: For any monitoring data within the day, excluding the last two monitoring data, the difference between the change difference value of the any monitoring data and the minimum change difference value is obtained and recorded as the first difference value. The difference between the maximum change difference value and the minimum change difference value is obtained and recorded as the second difference value. The ratio between the first difference value and the second difference value is obtained to obtain the first degree of change of the any monitoring data. Obtain the absolute value of the difference between any monitoring data and the previous monitoring data, and record it as the previous degree of change difference. Obtain the absolute value of the difference between any monitoring data and the next monitoring data, and record it as the subsequent degree of change difference. Normalize the sum of the previous degree of change difference and the subsequent degree of change difference to obtain the second degree of change of any monitoring data. The sum of the first mutation degree and the second mutation degree is obtained to obtain the preliminary suspected abnormality degree of any of the monitoring data; Obtain the preliminary suspected anomaly level of each monitoring data point within the day, excluding the last two monitoring data points. Based on the preliminary suspected anomaly level of each monitoring data point within the day, excluding the last two monitoring data points, obtain the preliminary suspected anomaly data for the day.

4. The method for intelligent acquisition and monitoring of factory data based on the Internet of Things according to claim 3, characterized in that, The preliminary suspected anomaly data for the day is obtained based on the preliminary suspected anomaly level of each monitoring data point excluding the last two monitoring data points, including: For any monitoring data other than the last two monitoring data points within the same day, if the preliminary suspected abnormality level of any monitoring data is greater than or equal to the preset preliminary suspected abnormality level threshold, then the monitoring data is confirmed as preliminary suspected abnormal data.

5. The method for intelligent data acquisition and monitoring of factories based on the Internet of Things according to claim 1, characterized in that, The step of obtaining suspected abnormal data for the day based on the degree of suspected abnormality of each target data includes: For any target data, if the suspected abnormality level of the target data is greater than or equal to a preset suspected abnormality level threshold, then the target data is confirmed as suspected abnormal data.

6. The method for intelligent data acquisition and monitoring of factories based on the Internet of Things according to claim 1, characterized in that, The step of obtaining the degree of abnormal risk of the monitoring data sequence based on the distribution characteristics of suspected abnormal data in the monitoring data sequence includes: The number of suspected abnormal data in the monitoring data sequence is obtained, and the number of suspected abnormal data in the monitoring data sequence is normalized to obtain the first abnormal risk level. For any suspected abnormal data in the monitoring data sequence, the minimum time interval between any suspected abnormal data and every other suspected abnormal data is obtained, and recorded as the nearest abnormal time distance of any suspected abnormal data. The nearest abnormal time distance of each suspected abnormal data in the monitoring data sequence is obtained, and the cumulative value of the nearest abnormal time distance is obtained. The reciprocal of the cumulative value of the nearest abnormal time distance is normalized to obtain the second abnormal risk level. The sum of the first abnormal risk level and the second abnormal risk level is obtained to obtain the abnormal risk level of the monitoring data sequence.

7. The method for intelligent data acquisition and monitoring of factories based on the Internet of Things according to claim 1, characterized in that, The step of obtaining the data collection frequency of any monitoring indicator within a preset time range based on the degree of abnormal risk of the monitoring data sequence includes: The data collection frequency of any monitoring indicator within a preset time range is obtained, and the product of the abnormal risk level of the monitoring data sequence and the data collection frequency of any monitoring indicator within the preset time range is rounded up to obtain the data collection frequency of any monitoring indicator within a future preset time range.

Images