Graphical process early warning method based on chemical production monitoring and tracing

Abstract

The invention discloses a graphical process early warning method based on chemical production monitoring and tracing, and the method achieves the modeling of a product core index through the integrated analysis of a product production condition and a material BOM, the sorting of an index hierarchical relation, and the display in a mind map form. Real-time values of the indexes are obtained in real time, abnormity judgment is carried out, the index values and abnormal states are displayed on the indexes corresponding to the mind map, and graphical monitoring, tracing and early warning of products are achieved. The system is applied to a manufacturing execution system in the chemical industry, the key index relationship of chemical products is analyzed, other modules of the manufacturing execution system are connected to realize visual monitoring, tracing and early warning of product production on the premise of effectively maintaining basic production data, and product production abnormity is quickly responded in combination with message pushing.

Description

Technical Field

[0001] The present invention relates to the field of computer application technology, and in particular to a manufacturing execution system applied to the chemical industry, and specifically to a graphical monitoring and backtracking scenario of key indicators such as production condition parameters and material quality inspection in the production process of chemical products. Background Art

[0002] In complex production fields such as chemical industry, the product production process involves many production parameters and material BOM (bill of materials) information. Traditional monitoring and alarm methods are mostly discrete data display and simple threshold judgment, which lack integrity and intuitiveness. The management of key product indicators is also mostly carried out in isolation. When problems occur, it is difficult to quickly and accurately trace the cause. This situation seriously affects production efficiency and quality control, and cannot meet the needs of modern production for efficient, accurate monitoring and rapid problem tracing. A new technical means is urgently needed to integrate information and realize graphical monitoring, early warning and tracing functions. Summary of the invention

[0003] The purpose of the present invention is to solve the defects in the prior art and provide an innovative early warning method in the production process of chemical products, which can realize graphical monitoring, backtracking and early warning.

[0004] In order to achieve the above object, the technical solution provided by the present invention is as follows: A graphical process early warning method based on chemical production monitoring and traceability includes the following steps: Build product model: According to the chemical production process, determine the key indicators and their subordinate sub-indicators, and associate the production condition parameters and bill of materials to build the product model; Data storage and processing: real-time collection of production data, acquisition of material testing information, real-time values ​​of key indicators and sub-indicators, storage of corresponding time points in the historical database, and processing of each data to determine whether to trigger an alarm; Graphical monitoring: Generate a monitoring mind map based on the product model to display the current values ​​and alarm status of key indicators and sub-indicators; Historical backtracking: Taking any designated historical time as the starting time, according to the set time step frequency and screen refresh frequency, the data of the corresponding time point is extracted from the historical database, and graphically displayed and the alarm status of each time point is synchronously displayed.

[0005] The present invention establishes a product model through key indicators and sub-indicators, and connects the key indicators with production condition parameters and the bill of materials in series in the form of a mind map, so as to monitor the real-time value and alarm status of each indicator in a graphical way, making the monitoring process more intuitive; at the same time, it associates with the historical database, and realizes rapid backtracking of abnormal problems by extracting data at corresponding time points.

[0006] Furthermore, the graphical process early warning method also includes pushing alarm messages: designating responsible persons for each key indicator and sub-indicator in the product model, and pushing the corresponding responsible persons in real time for the indicators that trigger the alarm. By setting the responsible persons, the alarm can be quickly responded to abnormal situations.

[0007] Furthermore, the above-mentioned graphical process warning method also includes product model management: defining the attributes of the product model, including basic model attributes and subordinate indicator attributes; the basic model attributes include model number, model name and start-up and shutdown status; subordinate indicators are key indicators and sub-indicators confirmed in the construction of the product model; the subordinate indicator attributes include indicator code, indicator name, model to which it belongs, parent indicator, indicator type, data source, start-up and shutdown status, and prerequisite judgment formula; the parent indicator refers to the indicator that the current indicator will affect or the operating condition indicator and quality inspection indicator of the downstream equipment in the process production process; the indicator types include three types, namely real-time indicators, test indicators, and qualitative analysis indicators; real-time indicators are production condition data collected in real time; test indicators are obtained material test information data; qualitative analysis indicators are sub-indicators that are directly related to the abnormal state of the parent indicator, and they have no related Data source, whose main function is to assist logical judgment (i.e., if the parent indicator is abnormal, when the subordinate real-time indicators and test indicators it contains are normal, if there is a qualitative analysis sub-indicator, the state of the qualitative analysis sub-indicator is abnormal); the data source includes the alarm items bound to other system modules, quality inspection results, or DCS data collection points associated with custom indicators, and sets the normal range of the collection point values ​​or the alarm judgment formula; before the alarm status triggers the judgment, the start and stop status of the subordinate indicators is judged: when the indicator stop and start status is enabled, the indicator data source is used to collect the indicator status to determine whether an alarm is triggered, or the collected real-time data is used to determine whether the indicator is abnormal; when the indicator stop and start status is disabled, the background real-time calculation judgment will automatically skip this indicator; the prerequisite judgment formula refers to a formula used to preliminarily determine whether this indicator needs to be judged or calculated before the judgment or calculation of a certain indicator. The prerequisite judgment formula takes precedence over the data source judgment when performing real-time calculations in the background, thereby realizing a locking alarm: for example, there are two parallel feed pipelines in the production process, and it is necessary to monitor in real time whether the feed flow rate is within the limit range. The judgment of the valve switch state is set to the prerequisite judgment formula IF (switch==0, 0, 1). When the valve state is closed, the judgment of the feed flow rate limit range can be skipped.

[0008] The product model is defined and managed by maintaining the basic data of the product model and the model's subordinate indicator attributes.

[0009] Furthermore, the subordinate indicator attributes also include an indicator tracing address. By setting a tracing or analysis page, the abnormal indicator will be jumped to the tracing or analysis page.

[0010] Furthermore, the subordinate indicator attributes also include responsible persons, handling measures and scope of influence. The responsible person is associated with the department, position or personnel in the organization. When an alarm is triggered, a message will be pushed to the relevant personnel. Multiple push forms are supported, including APP, DingTalk, enterprise WeChat, SMS, etc.

[0011] Furthermore, the graphical process early warning method also includes an indicator alarm record, which is generated by summarizing the indicator alarm records according to the product model and statistics, and is used to record the historical alarm information of the indicator. The historical alarm information is displayed according to the model classification to facilitate the later analysis and statistics.

[0012] Compared with the prior art, the present invention has the following advantages: The present invention deeply integrates production condition parameters and material BOM information, abstracts key product indicators and connects them in series in the form of a mind map. In terms of monitoring, it can present the current value of each indicator and the production status in real time, and intuitively display the overall picture of production in a graphical way. In terms of the alarm function, based on preset rules and dynamic data analysis, through real-time calculation and judgment in the background, alarms are issued in time for abnormal situations, and treatment measures and early warning prompts of the impact range are given. At the same time, when a problem occurs, you can quickly backtrack along the mind map, set the backtracking start time, backtracking direction, and backtracking cycle frequency, etc., accurately locate the source of the abnormality, thereby effectively improving production efficiency, ensuring product quality, and providing strong support for the management of complex production processes.

[0013] The present invention is applied to the manufacturing execution system of the chemical industry to analyze the relationship between key indicators of chemical products. On the premise of effectively maintaining basic production data, it can also connect with other modules of the manufacturing execution system to realize visual monitoring, traceability and early warning of product production, and quickly respond to product production anomalies in combination with message push. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a flow chart of a graphical process early warning method based on chemical production monitoring and tracing of the present invention; Figure 2 This is a schematic diagram of the product model constructed in Example 2 of the present invention. DETAILED DESCRIPTION

[0015] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0016] The present invention discloses a graphical process early warning method based on chemical production monitoring and tracing, comprising: building a product model, real-time monitoring of the model, historical backtracking of the model, pushing of alarm messages, and recording of indicator alarms.

[0017] Among them, the product model is constructed, and the definition and management of the model is realized by maintaining the basic data of the product model and the model indicator attributes; the basic data of the product model includes the key indicators determined according to the chemical production process and their subordinate sub-indicators, as well as the related production condition parameters and material list.

[0018] And define the product model attributes. The core attributes include the basic attributes of the model and the definition of subordinate indicators. The basic attributes include the model code, model name, and start and stop status. Subordinate indicator definition, subordinate indicators are key indicators abstracted from the product model. These indicators are divided into indicator levels through product material BOM and production conditions. Subordinate indicator attributes include: indicator code, indicator name, model, parent indicator, indicator type, data source, start and stop status, prerequisite judgment formula, indicator traceability address, responsible person, handling measures and scope of impact.

[0019] The specific meaning of the parent indicator refers to the indicator that the current indicator may affect or the operating condition indicator and quality inspection indicator of the downstream equipment in the process production process.

[0020] There are three types of indicators, namely real-time indicators, laboratory indicators, and qualitative analysis indicators. Among them, real-time indicators are data collection points on the DCS side of key production conditions, laboratory indicators are laboratory data of key material indicators, and qualitative analysis indicators have no relevant data sources. Their main function is to assist in logical judgment, and they are sub-indicators that are directly related to the abnormal state of the parent indicator. For example, the parent indicator, the fluctuation of the brine high-level tank liquid level, is a real-time indicator. When its current real-time value is judged to be out of limit, the status is alarm. There are two possibilities for causing the fluctuation of the parent indicator, the brine high-level tank liquid level, which are the corresponding pump pressure fluctuation and the abnormal display of the liquid level gauge. However, these two sub-indicators have no data source to access the system, so they are set as qualitative analysis indicators that are directly related to the parent indicator, the fluctuation of the brine high-level tank liquid level. Once the parent indicator is abnormal, the sub-indicator will also be marked as abnormal to prompt the user to check.

[0021] The data source can be bound to the alarm items of other system modules, quality inspection results, or DCS data collection points associated with custom indicators, and the normal range of the collection point value or the alarm judgment formula can be set to facilitate subsequent calculations.

[0022] Enable or disable status. When the status is enabled, the indicator status is collected through the indicator data source (the data source can come from a third-party system) to determine whether an alarm is triggered, or the collected real-time data is used to determine whether the indicator is abnormal based on the set rules. If not, the background real-time calculation will automatically skip this indicator.

[0023] The prerequisite judgment formula will take precedence over the data source judgment when it is calculated and judged in real time in the background. It is mainly for realizing the locking alarm. That is, assuming that there are two parallel feeding pipelines, the feeding flow is monitored in real time to see if it is within the limit range. The judgment of the valve switch state is set as the prerequisite judgment formula. When the valve state is closed, the judgment of the limit range of the indicator feeding flow can be skipped.

[0024] The indicator tracing address is used to jump to the set tracing or analysis page by double-clicking the screen indicator when an abnormality occurs in the indicator of the product model.

[0025] The responsible person is associated with the department, position or person in the organization, and when an alarm is triggered, a message will be pushed to the relevant personnel.

[0026] Model enable / disable. When the status is enabled, this model will participate in the background real-time calculation and judgment. If not, the background real-time judgment will automatically skip this model.

[0027] Graphical real-time monitoring: After the product model is defined, the monitoring mind map is automatically generated through the defined model and subordinate indicators to achieve real-time monitoring of the model: real-time display of indicator values, current indicator abnormal status, and indicators that may be affected or the consequences caused by the warning; the model is automatically expanded to the guide Figure 2 The level node supports one-click expansion and one-click zooming of all indicators of the model, supports zooming of the mind map screen and one-click adaptive screen. The screen displays the current value of the indicator and the alarm status in real time. The indicator background is calculated and determined in real time. Normal indicators are displayed in blue, alarm indicators are highlighted in red and flashing, and qualitative analysis indicators are abnormal and flashing in yellow. The parent indicator displays the number of all alarm indicators of its subordinates with a red superscript, so that users can view the current alarm number without expanding the mind map.

[0028] Model historical backtracking: Automatically generate a monitoring mind map through the defined models and subordinate indicators, switch the monitoring mode to historical backtracking, set the time step frequency and direction, trace the historical data of model indicators forward or backward from historical moments, and view the abnormal evolution of the model.

[0029] Alarm message push is to push the model indicators that generate alarms to the relevant persons in charge in real time, so as to facilitate timely handling of alarms. For indicators that trigger alarms, multiple push forms are supported, including APP, DingTalk, enterprise WeChat, SMS, etc.

[0030] The indicator alarm record is a record of indicator alarms generated and summarized according to the product model.

[0031] The indicator alarm record is used to record the historical alarm information of the indicator and display the historical alarm information according to the model classification to facilitate the subsequent analysis and statistics.

[0032] Example 1

[0033] like Figure 1 The figure shows a schematic flow chart of the graphical process warning method of the present invention. Product model construction is the basis for system implementation. After the construction is completed, real-time monitoring of the model and historical backtracking of the model are achieved through data acquisition and real-time judgment, and the alarm indicators are pushed to the relevant responsible persons and archived.

[0034] First, the product model is established, and the product production process is comprehensively analyzed. In combination with industry standards and internal quality requirements of the enterprise, key indicators that affect product quality, performance, etc. are determined. These indicators cover various dimensions from raw material characteristics to physical and chemical parameters in the production process, as well as functional parameters of the final product. For example, in the production of chemical products, key indicators include raw material purity, reaction temperature, pressure, reaction time, product purity, density, etc. For each key indicator, its subordinate sub-indicators are further analyzed. Taking reaction temperature as an example, sub-indicators such as temperature setting values ​​and temperature fluctuation range at different reaction stages can be included; for product purity, sub-indicators such as main component content and impurity content limit can be included. These sub-indicators together constitute a complete product model structure. The determined key indicators and their sub-indicators are associated with production condition parameters and material BOM (bill of materials). Production condition parameters such as equipment operating parameters (speed, power, etc.) and environmental parameters (humidity, ventilation, etc.) are linked to corresponding indicators through data interfaces or manual entry.

[0035] The second step is to realize data calculation and monitoring screen display. Through the connection of established DCS and other data sources, real-time data collection of production conditions, including temperature, pressure, flow, etc., is used to obtain real-time parameters in the production process. At the same time, material test information such as concentration and quality inspection level are obtained from the quality management system. According to the pre-established product model and data association relationship, the real-time value of each key indicator and sub-indicator is calculated. The collected raw data is processed using the normal range of the set indicator data value or the alarm judgment formula to determine whether to trigger an alarm, and the real-time value of the indicator is displayed on the monitoring interface.

[0036] Monitoring is achieved through graphical interaction. Users can zoom, pan, view indicators one by one, expand or shrink all indicator elements with one click, and other operations on the interface. The indicator status is displayed at the same time. Normal indicators are displayed in blue, alarm indicators are highlighted in red and flashing, and qualitative analysis indicators are abnormal and flashing in yellow. The parent indicator displays the number of all alarm indicators of its subordinates through a red superscript, which is convenient for operators to view the indicator status intuitively. It supports double-clicking or right-clicking to view the detailed historical data trend or historical test details list of the indicator. For example, it shows the fluctuation of product purity in the past month. The operator can clearly see the value change trend of the purity indicator at different time points, as well as the relationship between it and the production condition parameters (such as temperature, material batch changes, etc.).

[0037] On the monitoring interface, the historical backtracking function is supported, and the screen monitoring mode can be switched with one click. The historical backtracking function provides a time selection function, and the user can specify any historical time as the start time. The system interacts with the database that stores historical production data to obtain all relevant production condition data and test information starting from the start time. The data stored in the database should have sufficient data to ensure data integrity to support detailed historical analysis. According to the time step frequency set by the user (such as 5 seconds, 30 seconds, 5 minutes) and the screen refresh frequency (such as 2 seconds, 5 seconds), the system gradually extracts the data at the corresponding time points, displays it on the screen, and simultaneously judges the alarm status. At the same time, the backtracking toolbar displays the current backtracking time in real time. It also supports viewing the historical details of indicators.

[0038] During the product model building phase, a responsible person is assigned for each key indicator and sub-indicator. The responsible person can be an operator, technical engineer, quality management personnel, etc. in the production department, and is determined based on the company's organizational structure and division of responsibilities. It supports setting up departments and positions, and notification methods include but are not limited to APP, SMS, WeChat for Enterprise, DingTalk, etc., so that users can respond to production anomalies in a timely manner.

[0039] The system automatically records each alarm event and forms an alarm record. The alarm record is stored in a special database table. Each record contains detailed information about the alarm event, such as alarm indicator, alarm time, release time (if the alarm has been released), alarm value, alarm threshold, responsible person, processing status, etc. This can comprehensively record the entire process of the alarm. The queried alarm records are counted and analyzed to generate reports and visual charts. For example, the frequency of occurrence of different types of alarms, the distribution of the number of alarms for each indicator, etc. can be counted to help enterprise managers understand the weak links and high-incidence areas of problems in the production process, so as to take targeted improvement measures and optimize the production process and quality control system.

[0040] The present invention is not limited to the above-mentioned embodiments. Any person who, after learning of the present invention, makes a technical solution having the same or similar technical solution as the present invention, shall be included in the protection scope of the present invention.

[0041] The techniques, shapes, and structural parts not described in detail in the present invention are all well-known techniques.

[0042] Example 2

[0043] This embodiment takes the high-boiling tower production process in the distillation process of chlor-alkali chemical resin production as an example. The distillation mainly involves heating and evaporating crude vinyl chloride containing higher impurities in a low-boiling tower to remove acetylene impurities with lower boiling points, and then heating and removing ethylene dichloride impurities with higher boiling points in a high-boiling tower.

[0044] The key production condition indicators in the high-low boiling tower distillation process are the high tower temperature, high tower pressure and high tower liquid level in the tower. The key and only material is the raw material crude vinyl chloride. Taking the high tower pressure as an example, the key indicators affecting the high tower pressure are disassembled and sorted out according to the process. There are four key indicators affecting the high tower pressure: 1. Reboiler power: The reboiler is used to provide heat source for the tower. The higher the reboiler power, the higher the temperature in the tower, which in turn affects the pressure. However, the reboiler power is also affected by the temperature of the heat source (circulating water). If the temperature of the heat source is high, the power used by the reboiler to heat the heat source can be reduced to achieve the same heating effect. In addition, the larger the opening of the heat source valve, the more circulating water enters the reboiler per unit time (the circulating water temperature is lower than the water temperature in the reboiler), and the reboiler needs to increase the power for heating.

[0045] 2. Condensate valve opening: The heating of the reboiler causes the vinyl chloride to be converted into gas phase and rise to the top of the tower. The more gas phase vinyl chloride is accumulated, the greater the pressure in the tower. The larger the valve opening of the top condenser, the more gas phase vinyl chloride will be liquefied into liquid and enter the reflux tank, thereby reducing the pressure in the tower.

[0046] 3. Condensate temperature: At the same condensate valve opening, the lower the condensate temperature, the faster the vinyl chloride in the gas phase at the top of the tower will liquefy, thereby reducing the pressure inside the high tower.

[0047] 4. Ethylene dichloride content in crude vinyl chloride feed: Different from the above three key indicators of operating conditions, the ethylene dichloride content in crude vinyl chloride feed is the raw material test data. The higher the ethylene dichloride content in the raw material, the less vinyl chloride can be evaporated in the gas phase, and the pressure in the gas phase vinyl chloride reduction tower will also decrease.

[0048] Similarly, for other key production condition indicators in the high boiling tower production process: high tower liquid level, high tower temperature, respectively confirm the corresponding key indicators and their subordinate sub-indicators, such as Figure 2 As shown, establish a product model.

[0049] Through the corresponding collection and measurement points on site, the corresponding production conditions and index data, such as temperature, pressure, flow, valve opening, etc., are obtained. The data source is configured as "real-time index" in the system and the associated points are collected in real time. For other subordinate index attributes with "laboratory index", such as the content of dichloroethane in vinyl chloride feed, the data source is configured as "laboratory index", and the latest laboratory data is obtained by associating quality inspection items (which can be third-party systems).

[0050] Set corresponding upper and lower limits for each subordinate indicator in the product model. Once an indicator is out of the limit, it is considered abnormal and an alarm is triggered. In addition, built-in formulas can be set for more complex alarms. An example is as follows: The reboiler temperature is affected by the heat source temperature and the heat source opening. Assume that the reboiler temperature and the heat source temperature are both within the normal range. Now the heat source temperature suddenly rises and lasts for one minute, but the increased temperature is still within the temperature limit. Assume that the mass of the original water in the reboiler is m1, the specific heat capacity is c, and the current temperature in the reboiler is T 1a , the original water inlet temperature is T 2a , rising to T 2b , the water flow per unit time is l2, assuming that after time t, the heat entering the reboiler is Q in = l2c(T 2b -T 2a )t, according to the heat balance , the temperature after mixing is, , now assume that the upper limit of the reboiler temperature is T 1b , then l2c(T 2b -T 2a )t=(m1+l2t)(T 1b -T 1a )c, and the solution is time t=(m1T 1b -m1T 1a ) / ( l2T 2b +l2T 1a -l2T 2a -l2T 1b ), through the above formula, we can deal with the situation where the factor index changes within the normal range but continues to run and causes the parent index to alarm. This formula is configured in the parent index warning formula. If the heat source temperature rises for one minute, the mind map monitoring screen will warn you how long the reboiler will exceed the limit. The system has built-in similar calculation formulas as alarm judgment formulas for users to refer to.

Claims

1. A graphical process early warning method based on chemical production monitoring and tracing, characterized in that: The steps include: Build product model: According to the chemical production process, determine the key indicators and their subordinate sub-indicators, and associate the production condition parameters and bill of materials to build the product model; Data storage and processing: real-time collection of production data, acquisition of material testing information, real-time values ​​of key indicators and sub-indicators, storage of corresponding time points in the historical database, and processing of each data to determine whether to trigger an alarm; Graphical monitoring: Generate a monitoring mind map based on the product model to display the current values ​​and alarm status of key indicators and sub-indicators; Historical backtracking: Taking any designated historical time as the starting time, according to the set time step frequency and screen refresh frequency, the data of the corresponding time point is extracted from the historical database, and graphically displayed and the alarm status of each time point is synchronously displayed.

2. The graphical process early warning method according to claim 1, characterized in that: The graphical process early warning method also includes pushing alarm messages: designating responsible persons for each key indicator and sub-indicator in the product model, and pushing corresponding responsible persons for indicators that trigger alarms in real time.

3. The graphical process early warning method according to claim 2, characterized in that: The graphical process early warning method also includes product model management: defining the attributes of the product model, including model basic attributes and subordinate indicator attributes; the model basic attributes include model number, model name and start and stop status; The subordinate indicators are key indicators and sub-indicators confirmed in the product model construction; the subordinate indicator attributes include indicator code, indicator name, model, parent indicator, indicator type, data source, start and stop status, and prerequisite judgment formula; the parent indicator refers to the indicator that the current indicator will affect or the operating condition indicator and quality inspection indicator of the downstream equipment in the process of process production; the indicator types include three types, namely real-time indicators, test indicators, and qualitative analysis indicators. The real-time indicator is the production condition data collected in real time, the test indicator is the obtained material test information data, and the qualitative analysis indicator is a sub-indicator directly related to the abnormal state of the parent indicator; the data source includes the alarm items bound to other system modules, quality inspection results, or DCS data collection points associated with custom indicators, and the normal range of the collection point value or the alarm judgment formula is set; before the alarm state is triggered, the start and stop status of the subordinate indicators is judged: when the indicator stop and start status is enabled, the indicator state is collected through the indicator data source to determine whether an alarm is triggered, or the collected real-time data is used to determine whether the indicator is abnormal; When the indicator is disabled, the background real-time calculation will automatically skip this indicator; the prerequisite judgment formula refers to a formula used to preliminarily determine whether this indicator needs to be judged or calculated before the judgment or calculation of a certain indicator; the prerequisite judgment formula takes precedence over the data source judgment in the background real-time calculation, thereby realizing the locking alarm.

4. The graphical process early warning method according to claim 3, characterized in that: The subordinate indicator attributes also include an indicator tracing address. By setting a tracing or analysis page, the abnormal indicator will be jumped to the tracing or analysis page accordingly.

5. The graphical process early warning method according to claim 4, characterized in that: The subordinate indicator attributes also include responsible persons, handling measures and scope of influence.

6. The graphical process early warning method according to claim 5, characterized in that: The graphical process early warning method also includes indicator alarm records, which are generated by summarizing the indicator alarm records according to the product model and statistics, so as to record the historical alarm information of the indicators.

Images