Abnormal monitoring method for chemical production process
By deploying distributed sensors and preprocessing data, combined with a two-way evaluation process, the problem of insufficient monitoring accuracy in chemical production processes has been solved, and precise anomaly monitoring in chemical production processes has been achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-10
- Publication Date
- 2026-04-07
AI Technical Summary
Traditional chemical production process monitoring methods suffer from false alarms or missed alarms due to the reliance on a single threshold. They are unable to adapt to fluctuations in operating conditions and equipment aging, and are difficult to distinguish between instantaneous anomalies and normal production fluctuations, thus affecting monitoring accuracy.
The system employs a process for collecting abnormal pollution source characteristics, a data preprocessing process, a routine data entry process, and a two-way evaluation process for abnormal monitoring parameters. Through distributed sensor deployment, data preprocessing, and two-way evaluation, it distinguishes between obvious anomalies and normal fluctuations and dynamically adjusts the treatment intensity.
It improves monitoring coverage and real-time performance, avoids monitoring blind spots, enables precise control and adapts to changes in the production environment, and enhances the accuracy of the monitoring system.
Smart Images
Figure CN121810029A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application belongs to the technical field of chemical monitoring, and particularly relates to a chemical production process anomaly monitoring method. BACKGROUND
[0002] The core purpose of monitoring the abnormality of a chemical production process is to ensure safety, prevent pollution, stabilize production, improve efficiency and meet regulatory requirements. Chemical production itself has characteristics such as high temperature and high pressure, flammability and explosiveness, toxicity and harm, complex process, and strong continuity. Any slight parameter deviation can trigger a chain reaction, leading to serious consequences. Since traditional monitoring uses fixed threshold values to judge abnormalities, it cannot adapt to dynamic factors such as process fluctuations, equipment aging, and raw material changes in chemical production processes, which can easily lead to false positives or false negatives. The problem of poor adaptability of static threshold monitoring causes each monitoring point to be difficult to distinguish between normal floating data and abnormal data, and the constant abnormal threshold set in actual operation can easily affect the accuracy of the system. That is, it is difficult to distinguish between transient abnormalities and normal production fluctuations relying on a single threshold value, which can lead to over-reaction or neglect of potential risks. Therefore, the adaptability of traditional abnormal monitoring methods is not ideal when applied to multiple monitoring points with variable monitoring data. For this reason, we propose a chemical production process anomaly monitoring method. SUMMARY
[0003] The application provides a chemical production process anomaly monitoring method to solve the problems in the background art.
[0004] The application provides the following technical scheme: a chemical production process anomaly monitoring method, comprising a pollution source abnormality feature collection process, a data preprocessing process, a conventional data entry process, and an abnormal monitoring parameter bidirectional evaluation process. Pollution source abnormality feature collection: select a sensor corresponding to a pollution source, arrange the sensor at a chemical production end collection surface through a distributed coverage method, and perform abnormality analysis on the pollution source parameters generated by the chemical equipment during production. The sensor nodes are connected through a wireless network to realize reading and uploading of collected data. The distributed coverage method arranges the sensor at the chemical production end collection surface, which can preliminarily screen and separately extract features from the abnormal chemical region that is easily affected by the environment and has variable monitoring, thereby reducing the work pressure of the subsequent data preprocessing process and improving the analysis efficiency during subsequent preprocessing.
[0005] Data preprocessing: distinguish the feature data that exceeds the set abnormal threshold value in the abnormality feature data, so that the abnormality feature data is divided into over-limit abnormal data and non-over-limit real-time data, and the over-limit abnormal data and non-over-limit real-time data are numerically integrated.
[0006] The conventional data is entered according to the corresponding pollution source exceeding value, the conventional parameter is used for comparison and analysis of the abnormal data exceeding the limit, after the conventional data is entered, the preprocessed real-time data not exceeding the limit is set as a conventional floating parameter, and the conventional data is modified twice according to the variable generated by the conventional floating parameter and used for comparison and analysis of the abnormal data exceeding the limit; after the conventional data is entered, the conventional floating parameter is set based on the real-time data not exceeding the limit, and the conventional data is modified twice based on the floating parameter, so that the judgment value is changed in real time, and in the subsequent abnormal monitoring parameter bidirectional evaluation process, the accuracy is avoided from being affected due to deviation of a single reference through comparison of two references. In the conventional data acquisition module, the reading unit receives data from the acquisition module, first sets the abnormal threshold value, then receives data from the acquisition module, and can quickly perform hard screening, and clearly distinguifies the data into two categories, namely the abnormal data exceeding the limit in violation of the rules or high risk, and the real-time data not exceeding the limit in normal production within the legal threshold value, and the risk value of the abnormal data is initially qualitatively separated, so that the subsequent monitoring system can quickly take corresponding response measures according to the abnormal data reacted by each collection point after acquiring the abnormal data.
[0007] Abnormal monitoring parameter bidirectional evaluation: the unmodified conventional data and the modified conventional data are compared with the obtained abnormal data exceeding the limit, and the processing degree of the external pollution source treatment system is set according to the difference items after comparison.
[0008] The further improvement of the present application is that the pollution source abnormal feature acquisition is realized through a pollution source abnormal feature acquisition module, the pollution source abnormal feature acquisition module includes a sensor for abnormal data monitoring and acquisition, a communication machine and a single-chip microcomputer for abnormal analysis, the sensor is signal connected with the single-chip microcomputer, and the single-chip microcomputer is signal connected with the external monitoring system through the communication machine; after the sensor set in the pollution source abnormal feature acquisition process monitors that the pollution parameter falls back to the safe range, the signal transmitter set in the abnormal monitoring parameter bidirectional evaluation process automatically triggers a reset signal, and reactivates the operation of the pollution source abnormal feature acquisition module, so that the last obtained real-time data not exceeding the limit continues to work as a new monitoring frequency, and continuously collects data for subsequent process monitoring and processing.
[0009] The further improvement of the present application is that the conventional data acquisition is realized by a conventional data acquisition module, the conventional data acquisition module includes a reading unit connected with the communication machine, an HDD memory for storing the abnormal data collected by the sensor, and a data writing unit for entering the pollution source exceeding value.
[0010] A further improvement of the present invention is that the bidirectional evaluation process of the abnormal monitoring parameters is implemented by the bidirectional evaluation module of the abnormal monitoring parameters. The bidirectional evaluation of the abnormal monitoring parameters includes a microprocessor, a time-series data reader / writer, and a signal transmitter. The microprocessor completes the comparison and analysis of the uncorrected normal data and the corrected normal data with the number of out-of-limit anomalies. After the data is entered into the data storage and traceability through the time-series data reader / writer, it is synchronously sent to the monitoring terminal through the signal transmitter.
[0011] A further improvement of the present invention is that it also includes a secondary monitoring and reset process, which is used to continuously control the operation of the pollution source abnormal feature acquisition module when the bidirectional evaluation process of abnormal monitoring parameters ends and the abnormal value is obtained.
[0012] Compared with existing technologies, the beneficial effects of this invention are as follows: By incorporating the pollution source anomaly feature acquisition process, distributed monitoring deployment is implemented, improving monitoring coverage and real-time performance, avoiding monitoring blind spots, and increasing the data acquisition efficiency of each monitoring point. Furthermore, after data preprocessing, the data acquired by the pollution source anomaly feature acquisition process can clearly distinguish between obvious anomalies and normal fluctuations. Subsequently, after routine data entry, routine floating parameters are set based on real-time data that does not exceed limits, and routine data is corrected a second time based on these floating parameters to adapt to changes in the production environment and to make real-time changes in the judgment values. In the subsequent bidirectional evaluation process of anomaly monitoring parameters, the accuracy is avoided due to deviations from a single benchmark by comparing two benchmarks. Moreover, after evaluation and comparison, the treatment intensity of the pollution treatment system can be dynamically adjusted directly based on specific differences, achieving precise control. Attached Figure Description
[0013] Fig. 1 This is a flowchart of a method for monitoring anomalies in a chemical production process according to the present invention.
[0014] Fig. 2 This is a structural component used to implement the monitoring process in a chemical production process anomaly monitoring method of the present invention. Detailed Implementation
[0015] The present invention will be further described below with reference to specific embodiments. The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the present invention. In order to better illustrate the specific embodiments of the present invention, some parts in the drawings may be omitted, enlarged or reduced, and do not represent the actual product size. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings. Based on the specific embodiments of the present invention, all other specific embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0016] To address the technical problem in existing technologies where a fixed anomaly threshold can negatively impact system accuracy—specifically, relying solely on a single threshold makes it difficult to distinguish between instantaneous anomalies and normal production fluctuations, potentially leading to over-response or ignoring potential risks—this invention solves this problem by establishing a pollution source anomaly feature acquisition process, a data preprocessing process, a routine data entry process, and a two-way evaluation process for anomaly monitoring parameters. The inclusion of the pollution source anomaly feature acquisition process enables distributed monitoring deployment, improving monitoring coverage and real-time performance, avoiding monitoring blind spots, and increasing the efficiency of data acquisition at each monitoring point. Furthermore, data preprocessing allows for a clear distinction between obvious anomalies and normal fluctuations in the data acquired through the pollution source anomaly feature acquisition process. Following routine data entry, a routine floating parameter is set based on real-time data that does not exceed limits, and this floating parameter is used to further correct the routine data, adapting to changes in the production environment and adjusting the judgment values accordingly. In the subsequent two-way evaluation process for anomaly monitoring parameters, comparison between two benchmarks avoids accuracy issues caused by deviations from a single benchmark. Moreover, after evaluation and comparison, the treatment intensity of the pollution treatment system can be dynamically adjusted directly based on specific differences, achieving precise control.
[0017] Please see Figs. 1-2 A method for monitoring anomalies in chemical production processes includes a process for collecting abnormal characteristics of pollution sources, a data preprocessing process, a routine data entry process, and a two-way evaluation process for abnormal monitoring parameters.
[0018] Pollution source anomaly feature acquisition: Sensors corresponding to the pollution source are selected and deployed at the chemical production end of the acquisition surface through a distributed coverage method. Anomaly analysis is performed based on the pollution source parameters generated by the chemical equipment during production. The sensor nodes are connected through a wireless network to realize the reading and uploading of the acquired data. By adding the pollution source anomaly feature acquisition process, distributed monitoring deployment is carried out to improve monitoring coverage and real-time performance, avoid monitoring blind spots, and improve the data acquisition efficiency of each monitoring point.
[0019] Data preprocessing: Distinguish the abnormal feature data that exceeds the set abnormal threshold, so that the abnormal feature data is divided into out-of-limit abnormal data and non-out-of-limit real-time data, and then the out-of-limit abnormal data and non-out-of-limit real-time data are numerically integrated; After data preprocessing, the data obtained by the pollution source abnormal feature collection process can be clearly distinguished from obvious abnormalities and normal fluctuations.
[0020] Routine data entry: Routine data is set based on the corresponding pollution source exceedance values. This routine parameter is used for comparative analysis of exceedance anomaly data. After routine data entry, the preprocessed real-time data that does not exceed the limit is set as the routine floating parameter. The routine data used for comparative analysis of exceedance anomaly data is then corrected based on the variables generated by the routine floating parameter. After routine data entry, the routine floating parameter is set based on the real-time data that does not exceed the limit, and the routine data is corrected based on the floating parameter to adapt to changes in the production environment and to make real-time changes in the judgment value. In the subsequent two-way evaluation process of anomaly monitoring parameters, the accuracy is avoided due to the deviation of a single benchmark by comparing two benchmarks.
[0021] Two-way evaluation of abnormal monitoring parameters: Uncorrected and corrected routine data are compared separately to obtain abnormal data exceeding limits. The treatment level of the external pollution source treatment system is set based on the differences in the compared data. After evaluation and comparison, the treatment intensity of the pollution treatment system can be dynamically adjusted directly based on specific differences, achieving precise control.
[0022] In one optional embodiment of this example, the abnormal features of pollution sources are acquired through a pollution source abnormal feature acquisition module. The pollution source abnormal feature acquisition module includes a sensor for abnormal data monitoring and acquisition, a communication unit, and a microcontroller for abnormal analysis. The sensor and the microcontroller are connected by signals, and the microcontroller is connected by signals to an external monitoring system through the communication unit.
[0023] In this embodiment, due to the large area, numerous equipment, and dispersed pollution sources in chemical production areas, multiple specific pollutants are assigned independent sensor nodes (such as VOCs, specific toxic gases, and particulate matter, depending on the type of chemical plant) for distributed data collection. This arrangement allows for precise correlation between abnormal emissions and specific production equipment, process steps, and operating teams, providing direct evidence for subsequent precise control and accountability. Furthermore, after the sensor nodes self-organize into a network using 5G wireless technology, they transmit the collected parameters such as concentration, flow rate, and temperature to the local microcontroller in real time. Since the uploaded data has already undergone preliminary screening and individual feature extraction based on the corresponding abnormal sensor layout, the workload of the backend data preprocessing process is reduced, improving the efficiency of subsequent preprocessing analysis.
[0024] In an optional embodiment of this example, routine data acquisition is implemented by a routine data acquisition module, which includes a reading unit connected to a communication device, an HDD memory for storing abnormal data collected by the sensor, and a data writing unit for recording the pollution source exceeding the standard value.
[0025] In this embodiment, before the reading unit in the conventional data acquisition module receives data from the acquisition module, it first sets an abnormal threshold. Then, after receiving data from the acquisition module, it can quickly perform hard screening and clearly distinguish the data into two categories: clearly illegal or high-risk out-of-limit abnormal data and normal production real-time data that fluctuates within the legal threshold and does not exceed the limit. The risk value of the abnormal data is initially qualitatively separated. This processing method is used so that after the subsequent monitoring system obtains abnormal data, it can quickly take corresponding response measures based on the abnormal data reported by each acquisition point.
[0026] In one optional embodiment of this example, the bidirectional evaluation process of abnormal monitoring parameters is implemented by the bidirectional evaluation module of abnormal monitoring parameters. The bidirectional evaluation of abnormal monitoring parameters includes a microprocessor, a time-series data reader / writer, and a signal transmitter. The microprocessor completes the comparison and analysis of uncorrected regular data and corrected regular data with the number of out-of-limit anomalies, and after the data is entered into the data storage and traceability through the time-series data reader / writer, it is synchronously sent to the monitoring terminal through the signal transmitter.
[0027] In this embodiment, since the preprocessed real-time data that does not exceed the limit is obtained, the continuously obtained real-time data that does not exceed the limit after preprocessing can be set as a regular floating parameter in the bidirectional evaluation process of abnormal monitoring parameters. This parameter is the natural fluctuation of equipment, process, and raw materials within the allowable range during actual operation, as well as the reference value during actual operation. In this embodiment, the setting method is as follows: the time series data reader reads the real-time data that does not exceed the limit in one week as the regular floating parameter. When catalyst activity is used as a chemical anomaly indicator, if the baseline value of a certain parameter rises slowly in the following week due to the natural decay of catalyst activity, but still does not exceed the standard, the microprocessor in the bidirectional evaluation process of abnormal monitoring parameters will correct the regular data a second time based on this variable and appropriately raise the reasonable upper limit of its comparison benchmark. The microprocessor is used to store the entire process, results, and recommended handling degree of the comparison analysis in an immutable decision log formed by the time series data reader. At the same time, the graded control command is sent to the external pollution source treatment system and monitoring terminal through the signal transmitter for them to use as reference data.
[0028] In an optional embodiment of this example, a secondary monitoring reset process is also included. This secondary monitoring reset process is used to continuously control the operation of the pollution source abnormal feature acquisition module after the bidirectional evaluation process of abnormal monitoring parameters ends and the abnormal value is obtained. Specifically, after the bidirectional evaluation process of abnormal monitoring parameters is completed and the external processing system performs the corresponding operation, if the sensor set in the pollution source abnormal feature acquisition process detects that the pollution parameter has fallen back to within the safe range, the signal transmitter set in the bidirectional evaluation process of abnormal monitoring parameters will automatically trigger a reset signal and reactivate the operation of the pollution source abnormal feature acquisition module. This allows the previously acquired real-time data that did not exceed the limit to continue working as the new monitoring frequency, and to continuously collect data for monitoring and processing in subsequent processes.
[0029] The workflow of this invention includes the following steps: Step S1: Obtain the pollution source parameters generated in a single production run of chemical equipment through distributed data acquisition, and perform anomaly analysis. Step S2: Perform data preprocessing on the pollution source parameters after anomaly analysis. First, distinguish the feature data that exceeds the set anomaly threshold in the anomaly feature data to obtain the out-of-limit anomaly data and the real-time data that does not exceed the limit. Then, perform numerical integration on the out-of-limit anomaly data and the real-time data that does not exceed the limit. Step S3: After the numerical integration of out-of-limit abnormal data and non-out-of-limit real-time data is completed, the preprocessed non-out-of-limit real-time data is set as a regular floating parameter, and the regular data used for comparison and analysis of out-of-limit abnormal data is corrected a second time based on the variables generated by the regular floating parameter, in order to monitor the machine learning of the data model. Step S4: Compare the uncorrected regular data and the corrected regular data with the obtained out-of-limit abnormal data to determine whether the current data is abnormal; Step S5: When data is abnormal, set the treatment level of the external pollution source treatment system according to the data difference items after comparison; Step S6: When the data is not abnormal, control the sensor to continue operating for real-time data acquisition.
[0030] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.
Claims
1. A method for monitoring anomalies in a chemical production process, characterized in that: This includes the process for collecting abnormal characteristics of pollution sources, the data preprocessing process, the routine data entry process, and the two-way evaluation process for abnormal monitoring parameters. Pollution source anomaly feature acquisition: Select sensors corresponding to the pollution source and deploy them on the chemical production end acquisition surface in a distributed coverage manner. Perform anomaly analysis based on the pollution source parameters generated by the chemical equipment during production. The sensor nodes are connected through a wireless network to realize the reading and uploading of the acquired data. Data preprocessing: Distinguish the abnormal feature data that exceed the set abnormal threshold, so that the abnormal feature data are divided into out-of-limit abnormal data and non-out-of-limit real-time data, and then integrate the out-of-limit abnormal data and non-out-of-limit real-time data. Routine data entry: Routine data is set according to the corresponding pollution source exceedance value. This routine parameter is used for comparative analysis of exceedance abnormal data. After routine data entry, the preprocessed non-exceedance real-time data is set as routine floating parameters, and the routine data used for comparative analysis of exceedance abnormal data is corrected a second time based on the variables generated by the routine floating parameters. Two-way evaluation of abnormal monitoring parameters: The abnormal data exceeding the limit are obtained by comparing the uncorrected regular data and the corrected regular data respectively, and the treatment level of the external pollution source treatment system is set according to the differences in the compared data.
2. The method for monitoring anomalies in a chemical production process according to claim 1, characterized in that: The abnormal feature acquisition of pollution sources is achieved through a pollution source abnormal feature acquisition module. The pollution source abnormal feature acquisition module includes a sensor for abnormal data monitoring and acquisition, a communication unit, and a microcontroller for abnormal analysis. The sensor and the microcontroller are connected by signal, and the microcontroller is connected by signal to an external monitoring system through the communication unit.
3. The method for monitoring anomalies in a chemical production process according to claim 1, characterized in that: The routine data acquisition is implemented by the routine data acquisition module, which includes a reading unit connected to the communication device, an HDD memory for storing abnormal data collected by the sensor, and a data writing unit for recording the pollution source exceeding the standard value.
4. The method for monitoring anomalies in a chemical production process according to claim 1, characterized in that: The bidirectional evaluation process of the abnormal monitoring parameters is implemented by the bidirectional evaluation module of the abnormal monitoring parameters. The bidirectional evaluation of the abnormal monitoring parameters includes a microprocessor, a time-series data reader / writer, and a signal transmitter. The microprocessor completes the comparison and analysis of the uncorrected normal data and the corrected normal data with the number of out-of-limit anomalies. After the data is entered into the data storage and traceability through the time-series data reader / writer, it is synchronously sent to the monitoring terminal through the signal transmitter.
5. The method for monitoring anomalies in a chemical production process according to claim 1, characterized in that: It also includes a secondary monitoring and reset process, which is used to continuously control the operation of the pollution source abnormal feature acquisition module when the bidirectional evaluation process of abnormal monitoring parameters ends and the abnormal value is obtained.
6. The method for monitoring anomalies in a chemical production process according to claim 1, characterized in that: Implement anomaly monitoring by following these steps: Step S1: Obtain the pollution source parameters generated in a single production run of chemical equipment through distributed data acquisition, and perform anomaly analysis. Step S2: Perform data preprocessing on the pollution source parameters after anomaly analysis. First, distinguish the feature data that exceeds the set anomaly threshold in the anomaly feature data to obtain the out-of-limit anomaly data and the real-time data that does not exceed the limit. Then, perform numerical integration on the out-of-limit anomaly data and the real-time data that does not exceed the limit. Step S3: After the numerical integration of out-of-limit abnormal data and non-out-of-limit real-time data is completed, the preprocessed non-out-of-limit real-time data is set as a regular floating parameter, and the regular data used for comparison and analysis of out-of-limit abnormal data is corrected a second time based on the variables generated by the regular floating parameter, in order to monitor the machine learning of the data model. Step S4: Compare the uncorrected regular data and the corrected regular data with the obtained out-of-limit abnormal data to determine whether the current data is abnormal; Step S5: When data is abnormal, set the treatment level of the external pollution source treatment system according to the data difference items after comparison; Step S6: When the data is not abnormal, control the sensor to continue operating for real-time data acquisition.