A reaction kettle liquid level gauge fault control method and system based on an industrial internet

Abstract

The present application relates to the technical field of industrial internet, and provides a reaction kettle liquid level meter fault control method and system based on industrial internet, the method comprising: determining a fault root cause according to reaction kettle liquid level meter operation data, reaction kettle replenishment measurement data and a distinguishing result of abnormal state of the liquid level meter, and generating operation guidance information according to the fault root cause; sending the operation guidance information to an operation and maintenance terminal to guide the operation and maintenance terminal to check and process a communication link, a sensor body or a reaction kettle working condition; when a data credibility evaluation result meets a preset recovery condition, and a deviation between a measured liquid level output by the reaction kettle liquid level meter and reaction kettle liquid level information is within a preset allowable range, restoring a feedback source of liquid level control from the reaction kettle liquid level information output by a digital twin to the measured liquid level output by the reaction kettle liquid level meter according to a preset recovery observation period. The present application has the effect of improving the continuity, stability and intelligent level of industrial production.

Description

Technical Field

[0001] This invention relates to the technical field of industrial internet, specifically to a method and system for fault control of reaction vessel level gauges based on industrial internet. Background Technology

[0002] In modern industrial production, especially in fields like chemical and pharmaceutical manufacturing where continuous production is crucial, the reactor, as a core piece of equipment, is vital for ensuring production efficiency, product quality, and operational safety through accurate monitoring and stable control of its internal liquid level. With the widespread application of Industrial Internet technology, equipment on the production floor is interconnected, and large amounts of operational data are uploaded to cloud platforms in real time, aiming to optimize the entire production process through data analysis. However, in actual industrial environments, reactor level gauges often encounter various complex operating conditions, such as media corrosion, fluctuations under high temperature and pressure, and the gradual aging of the sensors themselves. These factors can easily lead to level gauge malfunctions, such as sensor network interruptions or data transmission link disruptions.

[0003] In traditional control methods, once the reactor level gauge malfunctions, the control system often stalls due to the lack of accurate feedback signals, necessitating emergency shutdowns or manual intervention. This not only causes unplanned production line interruptions but also severely impacts the stability and economic efficiency of continuous production. Current technologies, when facing level gauge failures, primarily rely on hard-wired feedback from a single sensor or simple local redundancy switching, lacking the global collaborative processing capabilities required in an industrial internet environment. Specifically, traditional control methods struggle to distinguish between communication interruptions and sensor soft faults within milliseconds, and the limited computing power of local controllers prevents the rapid generation of reliable alternative control strategies when data is missing. Furthermore, when a fault occurs, maintenance personnel often struggle to quickly pinpoint the root cause (sensor damage, wiring issues, or network congestion), resulting in long repair cycles, compromising production continuity, and failing to meet the high availability and adaptive control requirements of current smart manufacturing.

[0004] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0005] This application discloses a fault control method and system for reactor level gauges based on the Industrial Internet, aiming to solve the problems of production interruption caused by reactor level gauge failures, difficulty in quickly locating the root cause of the failure, and lack of global collaborative processing capabilities under traditional control methods.

[0006] The technical solution of this application is as follows: In a first aspect, this application discloses a fault control method for a reactor level gauge based on the Industrial Internet, comprising the following steps: The system acquires the operating data of the reactor level gauge, the supplementary measurement data of the reactor, and the physical constraints of the reactor, and conducts a reliability assessment of the operating data of the reactor level gauge to obtain the data reliability assessment results. A digital twin of the reactor and its corresponding level gauge is constructed, and the reactor level gauge operation data, data reliability assessment results, reactor supplementary measurement data, and reactor physical constraints are synchronized to the digital twin; the digital twin is a digital model used to simulate reactor level changes and level gauge response states. Based on the operating data and data reliability assessment results of the reactor level gauge in the digital twin, abnormal states of the level gauge are identified, and the abnormal states of the level gauge are classified into communication interruption or sensor failure, thus obtaining the classification results of abnormal states of the level gauge. In response to abnormal conditions of the level gauge, based on the supplementary measurement data and physical constraints of the reactor in the digital twin, the reactor level information is calculated, and the feedback source for level control is switched from the measured level output by the reactor level gauge to the reactor level information output by the digital twin. Based on the liquid level information of the reactor, control commands are generated to adjust the reactor actuators, and the control commands are transmitted to the reactor actuators. Based on the operating data of the reactor level gauge, the supplementary measurement data of the reactor, and the differentiation results of the abnormal state of the level gauge, the root cause of the fault is determined, and operation guidance information is generated based on the root cause of the fault. The operation guidance information is sent to the operation and maintenance terminal to guide the operation and maintenance terminal to check and handle the communication link, sensor body, or reactor operating condition. When the data reliability assessment results meet the preset recovery conditions, and the deviation between the measured liquid level output by the reactor level gauge and the reactor liquid level information is within the preset allowable range, the feedback source of liquid level control will be restored from the reactor liquid level information output by the digital twin to the measured liquid level output by the reactor level gauge according to the preset recovery observation period.

[0007] This technical solution enables intelligent identification of reactor level gauge malfunctions, rapid switching of control feedback sources, precise location of the root cause of the malfunction, and provision of operational guidance, thereby effectively ensuring the continuity and stability of industrial production and avoiding downtime losses caused by level gauge malfunctions.

[0008] Secondly, this application also discloses a fault control system for a reactor level gauge based on the Industrial Internet, used to perform fault control of the reactor level gauge based on the Industrial Internet, including: The credibility assessment module is used to acquire the operating data of the reactor level gauge, the supplementary measurement data of the reactor, and the physical constraints of the reactor, and to assess the credibility of the operating data of the reactor level gauge to obtain the data credibility assessment result. The digital twin construction module is used to build a digital twin of the reactor and its corresponding level gauge, and to synchronize the reactor level gauge operation data, data reliability assessment results, reactor supplementary measurement data and reactor physical constraints to the digital twin; The abnormal state identification module is used to identify abnormal states of the level gauge based on the operating data and data reliability assessment results of the reactor level gauge in the digital twin, and to distinguish between communication interruption or sensor failure, thus obtaining the distinction result of the abnormal state of the level gauge. The liquid level measurement switching module is used to respond to abnormal conditions of the liquid level gauge, supplement the measurement data of the reactor in the digital twin and the physical constraints of the reactor, calculate the liquid level information of the reactor, and switch the feedback source of liquid level control from the measured liquid level output by the reactor liquid level gauge to the reactor liquid level information output by the digital twin. The control command generation module is used to generate control commands for adjusting the reactor actuator based on the reactor liquid level information, and to transmit the control commands to the reactor actuator. The fault root cause determination module is used to determine the fault root cause based on the operating data of the reactor level gauge, the supplementary measurement data of the reactor, and the differentiation results of the abnormal state of the level gauge. Based on the fault root cause, it generates operation guidance information and sends the operation guidance information to the operation and maintenance terminal to guide the operation and maintenance terminal to check and handle the communication link, sensor body, or reactor operating condition. The liquid level measurement recovery module is used to restore the liquid level control feedback source from the digital twin output to the measured liquid level output by the reactor liquid level gauge when the data reliability assessment result meets the preset recovery conditions and the deviation between the measured liquid level output by the reactor liquid level gauge and the reactor liquid level information is within the preset allowable range, according to the preset recovery observation period.

[0009] This technical solution provides a system for implementing the aforementioned fault control method. Through modular design, it ensures efficient collaboration among various functions, thereby achieving comprehensive and intelligent control of reactor level gauge malfunctions and ensuring the continuity and stability of production.

[0010] Beneficial Effects: The industrial internet-based fault control method for reactor level gauges disclosed in this application, by introducing digital twin technology, can acquire and evaluate the reliability of level gauge operating data in real time. Combined with supplementary measurement data from the reactor and physical constraints, a high-precision digital twin model is constructed. When the level gauge malfunctions, this method can quickly identify and distinguish between communication interruption and sensor failure, avoiding misjudgments and processing delays caused by unclear fault types in traditional methods. After a fault occurs, the system can calculate the reactor level information based on the digital twin and promptly switch the feedback source for level control from the faulty level gauge to the digital twin, thereby ensuring that the control system can continuously obtain accurate level data and avoid unplanned production line interruptions. Furthermore, this method can accurately determine the root cause of the fault based on the fault differentiation results and generate operational guidance information to be sent to the maintenance terminal, greatly shortening the fault diagnosis and repair time. Through the above technical solutions, this application effectively solves the problems of production stagnation caused by level gauge failure, difficulty in fault location, and lack of global collaborative processing capability in the prior art, significantly improving the continuity, stability and intelligence level of industrial production, and providing strong technical support for intelligent manufacturing. Attached Figure Description

[0011] Figure 1 This is a flowchart of a method for fault control of a reactor level gauge based on the Industrial Internet, as described in one embodiment of the present invention. Figure 2 This is a flowchart of a method for fault control of a reactor level gauge based on the Industrial Internet, as described in another embodiment of the present invention. Figure 3 This is a system block diagram of a reaction vessel level gauge fault control system based on the Industrial Internet, according to another embodiment of the present invention. Explanation of reference numerals in the attached figures: 1. Fault control system for reactor level gauge based on industrial internet; 11. Reliability assessment module; 12. Digital twin construction module; 13. Abnormal state identification module; 14. Measurement level switching module; 15. Control command generation module; 16. Fault root cause determination module; 17. Measurement level recovery module. Detailed Implementation

[0012] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0013] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0014] This application proposes a fault control method for reactor level gauges based on the Industrial Internet, combining... Figure 1 As shown, it includes: S1. Acquire the operating data of the reactor level gauge, the supplementary measurement data of the reactor, and the physical constraints of the reactor, and conduct a reliability assessment of the operating data of the reactor level gauge to obtain the data reliability assessment result. S2, construct a digital twin of the reactor and its corresponding level gauge, and synchronize the reactor level gauge operation data, data reliability assessment results, reactor supplementary measurement data, and reactor physical constraints to the digital twin; the digital twin refers to a digital model used to simulate reactor level changes and level gauge response states; S3, based on the operating data and data reliability assessment results of the reactor level gauge in the digital twin, identify the abnormal state of the level gauge and classify the abnormal state of the level gauge into communication interruption or sensor body failure, and obtain the classification result of the abnormal state of the level gauge. S4, in response to an abnormal state of the level gauge, calculates the liquid level information of the reactor based on the supplementary measurement data and physical constraints of the reactor in the digital twin, and switches the feedback source of the liquid level control from the measured liquid level output by the reactor level gauge to the reactor liquid level information output by the digital twin; S5, Generate control commands for adjusting the reactor actuator based on the reactor liquid level information, and transmit the control commands to the reactor actuator; S6. Based on the operating data of the reactor level gauge, the supplementary measurement data of the reactor, and the differentiation results of the abnormal state of the level gauge, determine the root cause of the fault, generate operation guidance information based on the root cause of the fault, and send the operation guidance information to the operation and maintenance terminal to guide the operation and maintenance terminal to check and handle the communication link, sensor body or reactor operating condition. S7. When the data credibility assessment result meets the preset recovery conditions, and the deviation between the measured liquid level output by the reactor level gauge and the reactor level information is within the preset allowable range, the feedback source of the liquid level control is restored from the reactor level information output by the digital twin to the measured liquid level output by the reactor level gauge according to the preset recovery observation period.

[0015] To facilitate a clearer understanding of the technical solutions in this application, some key terms and implementation environments are explained first. "Reaction vessel level gauge operating data" refers to data collected and output in real time by the reaction vessel level gauge. This data may include measured liquid level, echo signal strength, timestamp, communication status code, etc., reflecting the current operating status of the reaction vessel level gauge and the liquid level measurement results. "Reaction vessel supplementary measurement data" refers to auxiliary measurement data obtained from other sensors or control systems of the reaction vessel, in addition to the reaction vessel level gauge operating data. This data may include mass flow rate data, temperature data, pressure data, stirring motor operating parameters, etc., used to assist in judging the internal working conditions of the reaction vessel and the trend of liquid level changes. "Reaction vessel physical constraints" refers to the inherent physical and process constraints of the reaction vessel during its design and operation, such as the reaction vessel's geometric dimensions, volumetric structure, material physicochemical properties, inlet and outlet relationships, and process parameters. These provide a constraint basis for reaction vessel level estimation and digital twin modeling.

[0016] A "digital twin" refers to a digital model used to simulate changes in the liquid level in a reactor and the response state of the level gauge. Through data synchronization and model calculation, it can reflect the real-time operating behavior of the reactor and its corresponding level gauge. "Data reliability assessment results" refer to the evaluation results obtained after analyzing the reliability, completeness, and validity of the reactor level gauge's operating data. "Abnormal level gauge status" refers to abnormal operating states that occur during the operation of the reactor level gauge, which may include communication interruptions and sensor malfunctions. "Communication interruption" refers to an abnormal data transmission link between the reactor level gauge and the control system, resulting in data failure to upload normally, interrupted timestamps, or abnormal data transmission delays. "Sensor malfunction" refers to hardware damage, performance degradation, abnormal echo signals, or measurement drift of the reactor level gauge sensor itself, leading to inaccurate measurement data or unstable output. "Reactor level information" refers to the internal liquid level data of the reactor calculated based on the digital twin under abnormal level gauge status. "Reactor actuator" refers to the physical actuator used to regulate the reactor liquid level, such as a feed pump, discharge valve, or discharge pump.

[0017] The implementation environment of this application can be built on an industrial internet platform. The industrial internet platform can include a data acquisition layer, an edge computing layer, and a cloud platform layer. The data acquisition layer is used to collect operational data from the reactor level gauge, supplementary measurement data from the reactor, and status data from the reactor actuators. The edge computing layer is used for reliability assessment, anomaly identification, rapid level calculation, and control switching of real-time data. The cloud platform layer can be used to store reactor physical constraints, historical operational data, fault records, and digital twin model parameters, and output operational guidance information to the maintenance terminal. Data interaction and command transmission between the layers can be achieved through industrial networks, OPC UA, MQTT, or other industrial communication protocols.

[0018] The fault control method for reactor level gauges based on the Industrial Internet proposed in this application focuses on simulating reactor level changes and level gauge response states using a digital twin. When a communication interruption or sensor malfunction occurs in the reactor level gauge, the system uses supplementary measurement data and physical constraints from the reactor to calculate the reactor level information. The feedback source for level control is switched from the measured level output by the reactor level gauge to the reactor level information output by the digital twin. Subsequently, the system generates control commands based on the calculated reactor level information and, combined with the differentiation results of the abnormal level gauge state, determines the root cause of the fault and provides operational guidance information. After the level gauge returns to normal, the system, based on data reliability assessment results, the deviation between the measured level and the reactor level information, and a preset recovery observation period, restores the feedback source for level control to the measured level output by the reactor level gauge, thus forming a closed-loop control process of level gauge fault identification, alternative control, fault guidance, and safe recovery.

[0019] First, the operating data of the reactor level gauge, supplementary measurement data of the reactor, and physical constraints of the reactor are acquired. The reliability of the reactor level gauge operating data is then assessed to obtain the data reliability assessment results. In practice, the reactor level gauge operating data can be acquired in real time through the reactor level gauge's own communication interface. For example, it can be read from the reactor level gauge via communication methods such as Modbus, Profibus, Ethernet / IP, and OPC UA, providing information such as measured liquid level, echo signal strength, timestamp, diagnostic status code, and communication status. Supplementary measurement data of the reactor can be obtained from other process instruments or control systems within the reactor. For example, it can be obtained from mass flow meters for feed and discharge flow, from temperature sensors for internal temperature, from pressure sensors for internal pressure, and from agitator motor controllers for agitation speed, current, and load status. Relevant operating data can also be read from DCS or SCADA systems. Reactor physical constraints can be extracted from reactor design documents, process specifications, equipment ledgers, or historical operating data and stored in the form of parameters, rules, or models.

[0020] After acquiring the operational data from the reactor level gauge, the system performs a reliability assessment on this data. Specifically, this assessment can be conducted through communication verification, data range checks, data change rate checks, timestamp continuity checks, and echo signal quality checks. For example, when the measured liquid level output by the reactor level gauge is within a reasonable range, the measured liquid level change rate matches the inlet and outlet status reflected in the supplementary measurement data of the reactor, the timestamps are continuous, and the echo signal strength meets preset requirements, a high-reliability data reliability assessment result can be obtained. Conversely, when the measured liquid level changes abruptly, remains unchanged for a long time, the timestamp is interrupted, the echo signal attenuates significantly, or the communication status code is abnormal, a low-reliability data reliability assessment result can be obtained. This data reliability assessment result is used to subsequently identify abnormal states of the level gauge and determine whether the feedback source for level control needs to be switched.

[0021] Secondly, a digital twin of the reactor and its corresponding level gauge is constructed, and the reactor level gauge operating data, data reliability assessment results, supplementary measurement data of the reactor, and physical constraints of the reactor are synchronized to the digital twin. In specific implementation, the digital twin can include a reactor level change model and a level gauge response model. The reactor level change model can be established based on the reactor's geometry, volumetric structure, material density, feed flow rate, discharge flow rate, temperature, pressure, and process reaction, used to simulate the process of level change within the reactor over time. The level gauge response model can be established based on the level gauge's response time, measurement accuracy, echo signal characteristics, drift characteristics, and abnormal response features, used to simulate the output state of the reactor level gauge under different operating conditions. The digital twin can be implemented using a physical mechanism-based modeling approach, a data-driven model trained with historical operating data, or a combination of physical and data-driven models.

[0022] After the digital twin is built, the system synchronizes the reactor level gauge's operating data, data reliability assessment results, supplementary reactor measurement data, and reactor physical constraints to the digital twin. Data synchronization can be achieved through real-time data interfaces, industrial internet gateways, or edge computing nodes. For example, the measured liquid level, echo signal strength, and timestamp output by the reactor level gauge, the mass flow rate data, temperature data, pressure data, and stirring motor operating parameters from the supplementary reactor measurement data, and the volume parameters and process constraints from the reactor's physical constraints can be synchronized to the digital twin via MQTT, OPC UA, or industrial database interfaces. Through data synchronization, the digital twin can continuously reflect the actual operating status of the reactor under normal or abnormal conditions of the level gauge, providing a model basis for level estimation and fault control.

[0023] Next, based on the operational data and data reliability assessment results of the reactor level gauge in the digital twin, the system identifies abnormal states of the level gauge and classifies them into communication interruption or sensor malfunction. Specifically, the system can determine whether the reactor level gauge is in an abnormal state based on the measured liquid level, echo signal strength, timestamp, and communication status information in the reactor level gauge's operational data, combined with the data reliability assessment results. When the timestamp continuity is interrupted, the data is not refreshed for a long time, the communication status code is abnormal, or the data transmission delay exceeds a preset range, the abnormal state of the level gauge can be classified as a communication interruption. When the reactor level gauge can still continuously output data, but the measured liquid level shows abnormal fluctuations, the measured value is significantly inconsistent with the material change trend reflected by the supplementary measurement data of the reactor, the echo signal strength is severely attenuated, the signal smoothness is abnormal, or the measurement drift continues to increase, the abnormal state of the level gauge can be classified as a sensor malfunction. By distinguishing between communication interruption and sensor malfunction, a basis can be provided for subsequent liquid level feedback switching, fault root cause location, and generation of operation guidance information.

[0024] Then, in response to the abnormal state of the level gauge, based on the supplementary measurement data and physical constraints of the reactor in the digital twin, the reactor level information is calculated, and the feedback source for level control is switched from the measured level output by the reactor level gauge to the reactor level information output by the digital twin. Specifically, when the abnormal state of the level gauge indicates that the measured level output by the reactor level gauge is no longer reliable, the digital twin uses synchronized supplementary measurement data and physical constraints of the reactor to calculate the level. For example, based on the feed mass flow rate, discharge mass flow rate, material density, reactor geometry, and volume curve, the change in material volume inside the reactor can be calculated through material balance, and then converted into a change in level. For situations where temperature and pressure affect material density or the gas-liquid state inside the reactor, corrections can be made by combining temperature data, pressure data, and process property parameters. For situations where stirring causes liquid level fluctuations or measurement disturbances, the level calculation results can be smoothed or dynamically compensated by combining the operating parameters of the stirring motor. The resulting reactor level information serves as an alternative level feedback source in the case of abnormal level gauge conditions.

[0025] When switching the feedback source for level control, the system can automatically perform the switch based on the abnormal state of the level gauge, data reliability assessment results, and operational safety policies. Alternatively, it can issue a confirmation prompt to the operator before performing the switch in situations requiring manual confirmation. After the switch is completed, the level control loop no longer uses the measured level output from the reactor level gauge as the primary feedback, but instead uses the reactor level information output by the digital twin as the feedback source. This process maintains the continuity of the level control process even if the reactor level gauge experiences a communication interruption or sensor failure, preventing malfunctions or shutdowns of the reactor actuator due to lost level feedback.

[0026] Subsequently, control commands for adjusting the reactor actuator are generated based on the reactor level information and transmitted to the reactor actuator. Specifically, after the feedback source for level control is switched to the reactor level information output by the digital twin, the control system generates control commands based on the deviation between the reactor level information and the preset level setpoint, combined with a preset level control strategy. The preset level control strategy may include PID control, feedforward control, rule-based control, or model predictive control. For example, when the reactor level information is lower than the level setpoint, control commands can be generated to increase the feed pump speed, increase the feed valve opening, or decrease the discharge valve opening; when the reactor level information is higher than the level setpoint, control commands can be generated to decrease the feed pump speed, decrease the feed valve opening, increase the discharge valve opening, or increase the discharge pump speed. These control commands are transmitted to the reactor actuator via the industrial network, enabling the reactor actuator to adjust the level based on the reactor level information calculated by the digital twin.

[0027] Simultaneously, based on the operational data of the reactor level gauge, supplementary measurement data of the reactor, and the differentiation results of the abnormal state of the level gauge, the root cause of the fault is determined, and operation guidance information is generated based on the root cause and sent to the operation and maintenance terminal. In specific implementation, when the abnormal state of the level gauge is classified as a communication interruption, the system can combine network logs, communication status codes, data refresh intervals, equipment online status, and control system interface status to determine whether the root cause is a communication cable malfunction, interface module failure, network equipment malfunction, or protocol communication malfunction. When the abnormal state of the level gauge is classified as a sensor body failure, the system can combine abnormal changes in the measured liquid level, echo signal strength attenuation, signal noise changes, temperature and pressure conditions, and supplementary measurement data of the reactor to determine whether the root cause is damage to the level gauge sensor hardware, probe contamination, media adhesion, measurement drift, sensor aging, or abnormal state of the media inside the reactor. Based on the determined root cause, the system generates operation guidance information, such as prompts to check the communication line between the level gauge and the control system, check the interface module, restart the communication equipment, clean the level gauge probe, replace the level gauge sensor, and check whether the media inside the reactor is blocked or adhered. Operation guidance information can be sent to mobile applications, tablets, SCADA system interfaces, maintenance work order systems, or other maintenance terminals to guide maintenance personnel in inspection and handling.

[0028] Finally, when the data reliability assessment results meet the preset recovery conditions, and the deviation between the measured liquid level output by the reactor level gauge and the reactor liquid level information is within the preset allowable range, the feedback source for liquid level control is restored from the reactor liquid level information output by the digital twin to the measured liquid level output by the reactor level gauge according to the preset recovery observation period. Specifically, after the maintenance is completed or the level gauge automatically recovers, the system continues to collect the reactor level gauge's operating data and re-evaluates the reliability. The preset recovery conditions may include continuous output of measured liquid level, continuous timestamps, normal communication status, echo signal strength restored to the preset range, reasonable measured liquid level change rate, and the data reliability assessment result reaching a high reliability level. At the same time, the system compares the measured liquid level output by the reactor level gauge with the reactor liquid level information output by the digital twin and calculates the deviation between the two. If the deviation is within the preset allowable range, it indicates that the measured liquid level output by the reactor level gauge is consistent with the result calculated by the digital twin and meets the conditions for restoration as the control feedback source.

[0029] To prevent the level gauge from malfunctioning again after a brief recovery, a preset recovery observation period is set in the system. During this period, the reliability assessment results of the data, the deviation between the measured level and the reactor level information, the echo signal strength, and the communication status are continuously monitored. Only when the reactor level gauge's operating data meets the recovery conditions throughout the entire preset recovery observation period, and the deviation between the measured level output by the reactor level gauge and the reactor level information remains within a preset allowable range, is the feedback source for level control restored from the reactor level information output by the digital twin to the measured level output by the reactor level gauge. Through this recovery process, frequent switching between abnormal and recovery states of the control feedback source can be avoided, ensuring the stability and safety of the reactor level control process. Thus, this application forms a complete closed loop from reactor level gauge operating data reliability assessment, level gauge abnormal state identification, digital twin substitute level estimation, reactor actuator control, fault root cause location, to feedback source restoration.

[0030] Optional, combined Figure 2 As shown, based on the operating data and data reliability assessment results of the reactor level gauge in the digital twin, the abnormal state of the level gauge is identified, and the abnormal state of the level gauge is classified into communication interruption or sensor failure. The steps to obtain the classification result of the abnormal state of the level gauge include: A1, Read the measured liquid level, echo signal strength and timestamp information output by the liquid level gauge from the operating data of the reactor liquid level gauge; A2, reads supplementary measurement data and physical constraints of the reactor from the digital twin. The supplementary measurement data of the reactor includes mass flow data, temperature data, pressure data and stirring motor operating parameters; A3, obtain real-time operating status of the reactor; A4. Based on the supplementary measurement data of the reactor, the physical constraints of the reactor, and the real-time operating conditions of the reactor, a physical behavior baseline of the reactor is constructed. The physical behavior baseline is used to calculate the liquid level change trajectory and echo signal intensity range of the level gauge under healthy conditions. A5. Based on the data reliability assessment results, the measured liquid level, echo signal strength, and timestamp information output by the liquid level gauge are compared with the physical behavior baseline to obtain the liquid level gauge output deviation. The liquid level gauge output deviation includes communication delay, signal smoothness index, and echo attenuation slope. A6. When the timestamp information is interrupted or the communication delay exceeds the preset communication threshold, the abnormal state of the level gauge is classified as a communication interruption, and a communication interruption classification result is generated. A7. When the timestamp information is normal and the signal smoothness index or echo attenuation slope exceeds the preset signal threshold, the abnormal state of the level gauge is classified as a sensor body failure, and a sensor body failure classification result is generated.

[0031] Specifically, when performing the above steps, it is first necessary to obtain key information such as the measured liquid level, echo signal strength, and timestamp from the reactor level gauge's operating data. The measured liquid level is the liquid level reading directly output by the level gauge; the echo signal strength reflects the attenuation of the level gauge's detection signal; and the timestamp information records the time of data generation, which is crucial for determining the timeliness and completeness of data transmission. Simultaneously, supplementary reactor measurement data read from the digital twin, such as mass flow rate data, temperature data, pressure data, and stirring motor operating parameters, as well as the reactor's physical constraints, provide comprehensive information on the reactor's internal operating conditions and physical characteristics. This data forms the basis for constructing an accurate baseline of physical behavior. Obtaining real-time reactor operating conditions, such as stable operating conditions, feeding conditions, or discharging conditions, helps to more accurately assess the level gauge's operating status.

[0032] Based on this, a physical behavior baseline for the reactor is constructed using supplementary measurement data, physical constraints, and real-time operating conditions. This baseline is a dynamic model used to calculate the expected liquid level change trajectory and echo signal intensity range under ideal health conditions. For example, under specific feed rates, temperatures, and pressures, the theoretically expected liquid level rise and the expected echo signal intensity level can be predicted using this baseline.

[0033] Subsequently, based on the data reliability assessment results, the measured liquid level, echo signal strength, and timestamp information output by the level gauge were compared with the aforementioned physical behavior baseline to quantify the level gauge's output deviation. This deviation is specifically manifested in communication delay, signal smoothness index, and echo attenuation slope. Communication delay refers to the time delay between data transmission from the level gauge and reception by the system; the signal smoothness index reflects the fluctuation or noise level of the measured liquid level; and the echo attenuation slope indicates the trend of echo signal strength change over time or distance. These deviation indicators can reveal potential anomalies in the level gauge from different dimensions.

[0034] Furthermore, based on these deviation indicators, the abnormal states of the level gauge are differentiated. When the timestamp information is interrupted, indicating a potential problem with the data transmission link or a communication delay exceeding a preset communication threshold, the abnormal state of the level gauge is clearly classified as a communication interruption. This usually signifies a fault in the data transmission link, such as an unstable network connection or a damaged communication module. Conversely, when the timestamp information is normal, but the signal smoothness index or echo attenuation slope exceeds a preset signal threshold, the abnormal state of the level gauge is classified as a sensor body fault. This typically indicates aging or damage to internal sensor components or interference from the external environment, leading to decreased measurement accuracy or deteriorated signal quality.

[0035] Optionally, the steps for constructing a baseline of the physical behavior of the reactor based on supplementary measurement data, physical constraints, and real-time operating conditions include: Based on the data reliability assessment results, echo signal strength, and timestamp information, determine whether the operating data of the reactor level gauge meets the selection criteria for the reference true value. When the real-time operating status of the reactor indicates that the reactor is in a stable operating condition, and the operating data of the reactor level gauge meets the reference true value selection conditions, the measured liquid level output by the level gauge is taken as the reference true value. By using the reference true value, supplementary measurement data of the reactor, and physical constraints of the reactor, the actual effective geometric parameters or actual effective material physical property parameters of the reactor can be calculated in reverse. The actual effective geometric parameters or actual effective material physical property parameters obtained by reverse calculation are updated to the digital twin to obtain the updated digital twin; Based on the updated digital twin, a baseline of the physical behavior of the reactor is constructed.

[0036] Specifically, the "reference truth value selection criteria" refer to the standards used to screen level gauge measurement data for those with high reliability and suitable as a true reference for liquid level. These criteria comprehensively consider the data reliability assessment results, echo signal strength, and the integrity and stability of the timestamp information. For example, when the data reliability assessment result is highly reliable, the echo signal strength is stable and within the normal range, and the timestamp information is continuous and uninterrupted, the data can be considered to meet the reference truth value selection criteria. The purpose is to ensure that the selected reference truth value has a high degree of accuracy and representativeness.

[0037] The "reference true value" can be understood as the measured value that is considered closest to the actual liquid level in the reactor under specific stable operating conditions. When the reactor is in stable operating conditions and the liquid level gauge's operating data meets the above-mentioned reference true value selection criteria, the measured liquid level output by the liquid level gauge is considered a reliable true liquid level and is thus selected as the reference true value.

[0038] In practical applications, "reverse engineering" refers to using known reference values ​​(i.e., actual liquid level), supplementary measurement data from the reactor (such as mass flow rate data, temperature data, pressure data, etc.), and physical constraints of the reactor (such as geometric dimensions, material density models, etc.) to derive the "actual effective geometric parameters" or "actual effective material physical property parameters" of the reactor originally set or estimated in the digital twin through reverse engineering or optimization algorithms. For example, if there are slight deviations between the preset reactor inner diameter or material density model in the digital twin and the actual situation, these parameters can be calibrated through reverse engineering to better match the current actual operating conditions. The purpose is to dynamically calibrate the internal parameters of the digital twin, enabling it to more accurately simulate the real physical behavior of the reactor.

[0039] Therefore, updating the digital twin with the actual effective geometric parameters or actual effective material physical property parameters obtained through reverse calculation means correcting and optimizing the physical model parameters within the digital twin. For example, the cross-sectional area parameter of the reactor used for liquid level calculation in the digital twin can be updated, or the functional model of the relationship between material density and temperature and pressure can be updated. The purpose is to enable the digital twin to more accurately reflect the current physical state and material properties of the reactor, thereby providing a more accurate basis for subsequent liquid level calculation and baseline construction.

[0040] Ultimately, "constructing a physical behavior baseline for the reactor based on the updated digital twin" refers to using a calibrated and optimized digital twin model to regenerate or adjust the liquid level change trajectory and echo signal intensity range of the level gauge under healthy conditions. Because the accuracy of the digital twin parameters is improved, the constructed physical behavior baseline will more closely resemble the actual operating behavior of the reactor, thereby improving the accuracy of abnormal state identification.

[0041] Optionally, the steps to compare the measured liquid level, echo signal strength, and timestamp information output by the level gauge with the physical behavior baseline, based on the data reliability assessment results, include: Acquire supplementary state data inside the reactor, including bubble entrainment degree, solid particle suspension concentration, and agitator operation status; The echo signal feature extraction strategy is adjusted based on the degree of bubble entrainment and the concentration of suspended solid particles. Based on the adjusted echo signal feature extraction strategy, echo signal features are extracted by combining the echo signal intensity with the echo signal's spectral distribution, waveform distortion, and signal-to-noise ratio. Based on the operating status of the agitator, the physical behavior baseline is corrected to obtain the corrected physical behavior baseline. Based on the data reliability assessment results, the measured liquid level, timestamp information and echo signal characteristics output by the liquid level gauge are compared with the corrected physical behavior baseline to obtain the liquid level change deviation, time response deviation and echo characteristic deviation. The output deviation of the level gauge is quantified based on the deviation of liquid level change, time response, and echo characteristic.

[0042] Specifically, acquiring supplementary data on the internal state of the reactor refers to obtaining real-time dynamic operating information within the reactor through additional sensors or model calculations. This includes: bubble entrainment level (the content or distribution of bubbles in the liquid within the reactor, affecting the propagation speed and attenuation of sound or radar waves in the medium); solid particle suspension concentration (the density or distribution of solid particles in the liquid, potentially causing scattering and attenuation of echo signals); and agitator operating status (the speed, direction, or presence of abnormalities in the agitator, causing liquid surface fluctuations, bubble entrainment, or changes in the local flow field, thus affecting the level gauge's measurement). This supplementary data is crucial for understanding the performance of the level gauge under complex operating conditions.

[0043] The strategy of adjusting the echo signal feature extraction based on the degree of bubble entrainment and the concentration of suspended solid particles can be understood as dynamically selecting or optimizing the algorithms and parameters used to extract effective features from the original echo signal for different levels of bubbles and solid particles. For example, when the degree of bubble entrainment is high, a more robust signal processing algorithm can be used to filter out noise caused by bubbles, or the frequency range of feature extraction can be adjusted to avoid the bubble resonance frequency; when the concentration of suspended solid particles is high, the signal gain can be adjusted or Doppler effect analysis can be used to distinguish between liquid level echoes and particle scattering echoes. The aim is to ensure that the true characteristics of the liquid level gauge echo signal can still be accurately captured in complex media.

[0044] In practical applications, echo signal features are extracted by combining the echo signal intensity with the echo signal's spectral distribution, waveform distortion, and signal-to-noise ratio (SNR). The spectral distribution reflects the signal's frequency components, the waveform distortion indicates interference or attenuation during propagation, and the SNR directly measures signal quality. These features provide a more comprehensive description of the echo signal's health, going beyond simple intensity values.

[0045] Furthermore, correcting the physical behavior baseline based on the agitator's operating status refers to adjusting the pre-established physical behavior baseline in real time, taking into account the impact of the agitator's operation on the dynamics of the reactor's liquid level and the characteristics of the echo signal. For example, when the agitator is running at high speed, eddies or fluctuations may occur on the liquid surface, causing periodic deviations in the level gauge readings; the agitator's operation may also alter the homogeneity of the medium, affecting the propagation of the echo signal. By incorporating the agitator's operating status, the physical behavior baseline can be made closer to the actual operating conditions of the reactor, thereby improving the accuracy of the comparison.

[0046] Therefore, combining the data reliability assessment results, the measured liquid level, timestamp information, and echo signal characteristics output by the level gauge are compared with the corrected physical behavior baseline to obtain the liquid level change deviation, time response deviation, and echo characteristic deviation. The liquid level change deviation reflects the difference between the measured liquid level and the liquid level estimated by the corrected baseline; the time response deviation reflects the difference between the liquid level gauge's response speed and the theoretical response speed, which may be related to communication delay or internal processing delay of the sensor; the echo characteristic deviation reflects the difference between the actual echo signal characteristics and the echo characteristics predicted by the corrected baseline, directly indicating the health status of the sensor itself.

[0047] Finally, the level gauge output deviation is quantified based on the level change deviation, time response deviation, and echo characteristic deviation. This is typically achieved by combining the deviations of these three dimensions into a unified level gauge output deviation index using methods such as weighted averaging, fuzzy logic, or machine learning models, in order to more comprehensively and accurately assess the overall operating status of the level gauge.

[0048] Optionally, when the real-time operating status of the reactor indicates that the reactor is in a stable operating condition, and the operating data of the reactor level gauge meets the reference true value selection criteria, the step of using the measured liquid level output by the level gauge as the reference true value includes: Based on the real-time operating conditions of the reactor, supplementary measurement data of the reactor, and physical constraints of the reactor, physical constraint verification conditions for selecting the reference true value are determined. The physical constraint verification conditions include the reasonable range of liquid level change, the reasonable range of time response, and the reasonable range of echo signal. The measured liquid level, echo signal strength, and timestamp information output by the liquid level gauge are compared with the physical constraint verification conditions to obtain multiple deviation indicators, including time delay, smoothness index, and echo attenuation slope. Based on the data credibility assessment results, determine whether the operating data of the reactor level gauge meets the data credibility conditions; Determine whether the deviation indicator exceeds the preset decrease threshold; When the operating data of the reactor level gauge meets the data reliability conditions and all deviation indicators do not exceed the preset drop threshold, the operating data of the reactor level gauge is determined to meet the reference true value selection conditions, and the measured liquid level output by the level gauge is taken as the reference true value.

[0049] Specifically, when determining the physical constraint verification conditions used for selecting the reference true value, a series of verification standards can be dynamically set based on the real-time operating conditions of the reactor, such as feed and discharge rates, stirring speed, temperature, and pressure, combined with supplementary measurement data of the reactor (such as mass flow rate data, temperature data, pressure data, and stirring motor operating parameters) and the physical constraints of the reactor (such as the reactor's geometric dimensions, material density, viscosity, etc.). These standards include a reasonable range for liquid level changes, used to limit the fluctuation amplitude of liquid level under stable operating conditions; a reasonable range for time response, used to evaluate the timeliness of the liquid level gauge's output response; and a reasonable range for echo signals, used to check the quality and stability of the echo signals. These verification conditions aim to ensure the physical rationality of the liquid level gauge data from multiple dimensions.

[0050] The comparison of the measured liquid level, echo signal strength, and timestamp information output by the level gauge with the aforementioned physical constraint verification conditions aims to quantify various deviations in the level gauge data. For example, by comparing the timestamp information with the reasonable range of time response, the communication delay deviation index can be obtained, reflecting the timeliness of data transmission; by analyzing the fluctuation of the measured liquid level with the reasonable range of liquid level changes, the signal smoothness index can be obtained, reflecting the stability of the liquid level data; and by comparing the echo signal strength with the reasonable range of echo signals, the echo attenuation slope can be obtained, reflecting the signal quality. These deviation indices provide a more detailed assessment of the operating status of the level gauge.

[0051] In practical applications, data reliability assessment results are a crucial basis for determining whether the operating data of the reactor level gauge meets the data reliability criteria. Data reliability criteria can be understood as data meeting preset standards in terms of completeness, accuracy, consistency, and timeliness. For example, if the data reliability assessment results show that the level gauge data has significant missing data, abnormal fluctuations, or serious discrepancies with historical data, it is considered that it does not meet the data reliability criteria.

[0052] Furthermore, determining whether the deviation indicators exceed the preset drop threshold aims to establish a tolerance level to distinguish between normal fluctuations and potential anomalies. The preset drop threshold can be determined based on historical operating data, expert experience, or statistical analysis methods, and is used to define the normal fluctuation range of each deviation indicator. Only when all deviation indicators do not exceed these thresholds is it considered that the level gauge data is physically reliable.

[0053] Therefore, only when the operating data of the reactor level gauge not only meets the data reliability conditions, but also when all its deviation indicators (such as time delay, smoothness index, and echo attenuation slope) do not exceed the preset descent threshold, can the operating data be finally determined to meet the reference true value selection conditions, and the measured liquid level output by the level gauge be used as the reference true value. This multi-verification mechanism ensures the high accuracy and reliability of the selected reference true value.

[0054] Optionally, the steps for determining whether the deviation index exceeds a preset decrease threshold include: Obtain the material batch information and reactor identification information for the current reactor; Based on the material batch information and reactor identification information, retrieve the threshold adjustment parameters that match the real-time operating conditions of the reactor from the preset parameter set; The preset decreasing threshold is corrected based on the threshold adjustment parameters to obtain the corrected preset decreasing threshold. The deviation index is compared with the corrected preset decline threshold to determine whether the deviation index exceeds the corrected preset decline threshold.

[0055] Specifically, obtaining the material batch information and reactor identification information of the current reactor is intended to provide necessary contextual information for subsequent threshold adjustments. Material batch information may include the type, formula, and production stage of the material being processed in the reactor, while reactor identification information uniquely identifies the specific reactor equipment, such as its model, serial number, or historical operating characteristics. This information forms the basis for refined threshold adjustments. Based on the material batch information and reactor identification information, threshold adjustment parameters matching the real-time operating conditions of the reactor are retrieved from a preset parameter set. The preset parameter set can be a knowledge base stored in a database or configuration table, containing threshold adjustment parameters pre-calibrated or learned from historical data for different material batches, reactor models, and real-time operating conditions (e.g., temperature, pressure, stirring speed). These parameters can be multiplicative factors, additive offsets, or other complex functional relationships used to correct the original preset threshold decrease. In practical applications, the preset threshold decrease is corrected based on the threshold adjustment parameters to obtain the corrected preset threshold decrease. The correction process can adjust parameters based on the retrieved threshold, dynamically adjusting the original preset drop threshold using a pre-defined algorithm (e.g., linear correction, non-linear correction, or rule-based correction). For example, for certain batches of materials prone to bubble formation or with high viscosity, the echo signal from the level gauge may be naturally weak or highly volatile. In such cases, the drop threshold for the echo attenuation slope can be appropriately widened. Conversely, for reaction vessels with extremely high level accuracy requirements, it may be necessary to tighten the threshold within a reasonable range of level changes. Therefore, the deviation index is compared with the corrected preset drop threshold to determine whether the deviation index exceeds the corrected preset drop threshold. By using the corrected, more adaptive threshold, it is possible to more accurately determine whether the level gauge operating data meets the reference true value selection criteria, thereby improving the accuracy of data reliability assessment.

[0056] Optionally, when the real-time operating status of the reactor indicates that the reactor is in a stable operating condition, and the operating data of the reactor level gauge meets the reference true value selection criteria, the step of using the measured liquid level output by the level gauge as the reference true value includes: When the real-time operating status of the reactor indicates that the reactor is in a stable operating condition, and the operating data of the reactor level gauge meets the reference true value selection conditions, the measured liquid level output by the level gauge is compared with the theoretical liquid level value calculated in the digital twin. When the deviation between the measured liquid level output by the level gauge and the theoretical liquid level value remains stable within a preset time window and is within a preset deviation range, the measured liquid level output by the level gauge is confirmed as the initial reference true value. Based on the initial reference true value, combined with historical deviation records and liquid level gauge model parameters, the initial reference true value is systematically corrected to obtain the corrected initial reference true value; the historical deviation record is used to represent the deviation record formed between the measured liquid level output by the liquid level gauge and the liquid level estimated by the digital twin during the historical operating cycle; Use the corrected initial reference truth value as the reference truth value.

[0057] Specifically, comparing the measured liquid level output by the level gauge with the theoretical liquid level value calculated in the digital twin involves using the physical model within the digital twin and real-time operating data (such as material composition information, temperature data, pressure data, and inlet / outlet mass flow rate data) to accurately calculate the current theoretical liquid level of the reactor when the reactor is under stable operating conditions and the level gauge data meets the reference true value selection criteria. Then, the actual liquid level measured by the level gauge is compared with this theoretical liquid level value to preliminarily assess the accuracy of the level gauge measurement.

[0058] Specifically, when the deviation between the measured liquid level output by the level gauge and the theoretical liquid level remains stable and within a preset deviation range within a preset time window, the measured liquid level output by the level gauge is confirmed as the initial reference true value. The preset time window aims to ensure that the measured value of the level gauge remains stable over a period of time, avoiding the impact of instantaneous fluctuations on the determination of the reference true value; the preset deviation range is used to judge the consistency between the measured value and the theoretical value, ensuring that the initial reference true value has high reliability.

[0059] In practical applications, based on the initial reference true value, and combining historical deviation records and level gauge model parameters, a systematic deviation correction is performed on the initial reference true value to obtain the corrected initial reference true value. Historical deviation records can be understood as the deviation trends and patterns between the level gauge's measured level and the level calculated by its digital twin during long-term operation, such as long-term drift and periodic errors. Level gauge model parameters include information such as the inherent characteristics, calibration curves, and aging models of that model of level gauge. By comprehensively considering these historical data and equipment characteristics, individualized systematic deviation correction parameters for a specific level gauge can be generated, thereby more accurately correcting the initial reference true value and eliminating or reducing systematic errors caused by factors such as equipment aging and environmental changes.

[0060] Optionally, the step of comparing the measured liquid level output by the level gauge with the theoretical liquid level value calculated in the digital twin may include the following process: Based on the supplementary measurement data of the reactor, real-time material composition information, real-time temperature data, real-time pressure data, and inlet and outlet mass flow rate data are obtained; Based on real-time material composition information, real-time temperature data, and real-time pressure data, determine the real-time density and real-time viscosity of the current material mixture; Based on real-time density, real-time viscosity, physical constraints of the reactor, and inlet and outlet mass flow rate data, the material balance estimation model in the digital twin is corrected. The theoretical liquid level of the reactor is calculated using the modified material balance estimation model. The measured liquid level output by the level gauge is compared with the theoretical liquid level value.

[0061] The supplementary measurement data for the reactor can be understood as auxiliary measurement data, in addition to the output data of the level gauge itself, used to more comprehensively reflect the internal state and operating conditions of the reactor. Specifically, this data includes real-time material composition information, real-time temperature data, real-time pressure data, and inlet and outlet mass flow rate data. Real-time material composition information is used to characterize the chemical composition of the materials inside the reactor, while real-time temperature and pressure data reflect the thermodynamic state inside the reactor, and inlet and outlet mass flow rate data are directly related to the dynamic changes of the materials inside the reactor. The acquisition of this data aims to provide accurate and comprehensive input for subsequent theoretical level estimation.

[0062] Specifically, based on the acquired real-time material composition information, real-time temperature data, and real-time pressure data, the real-time density and viscosity of the material mixture within the reactor can be accurately determined. Material density and viscosity are key physical parameters affecting liquid level estimation, and they dynamically change with variations in material composition, temperature, and pressure. Determining these parameters in real time ensures the accuracy of subsequent calculations.

[0063] Furthermore, based on the determined real-time density, real-time viscosity, physical constraints of the reactor, and inlet / outlet mass flow rate data, the material balance prediction model in the digital twin can be modified. The material balance prediction model is the core model used by the digital twin to simulate changes in reactor liquid level; it predicts the liquid level by considering the inflow and outflow of materials, reactions, and changes in physical properties. By modifying this model in real time, it can dynamically adapt to changes in actual operating conditions, thereby improving the accuracy of theoretical liquid level prediction.

[0064] Therefore, using the modified material balance estimation model, the theoretical liquid level of the reactor can be calculated. This theoretical liquid level represents the ideal liquid level predicted by the digital twin under the current precise operating conditions.

[0065] Finally, the measured liquid level output by the level gauge is compared with the calculated theoretical liquid level value. The purpose of this comparison is to evaluate the accuracy and reliability of the level gauge measurement, providing a basis for subsequent confirmation of the initial reference true value.

[0066] Optionally, the steps to systematically correct the initial reference true value by combining historical deviation records and level gauge model parameters to obtain the corrected initial reference true value include: Obtain the identification information of the reaction vessel and the level gauge; Based on the identification information of the reactor and the level gauge, retrieve the historical deviation records, operating history data, aging characteristic data and level gauge model parameters that match the current reactor and level gauge from the digital twin; Based on historical deviation records, operational history data, aging characteristic data, and level gauge model parameters, corresponding individualized systematic deviation correction parameters are generated. The initial reference true value is corrected by using individualized systematic bias correction parameters to obtain the corrected initial reference true value.

[0067] Specifically, obtaining the identification information of the reactor and level gauge means that, in an industrial internet environment, each reactor and level gauge typically has a unique identifier, such as a device ID, serial number, or asset tag. This identification information is used to accurately locate and access historical data and configuration parameters associated with specific equipment.

[0068] The process of retrieving historical deviation records, operational history data, aging characteristic data, and level gauge model parameters from the digital twin to match the current reactor and level gauge can be understood as follows: the digital twin, as a digital mapping of the physical device, stores all relevant information about the device from deployment to the present moment. Historical deviation records represent the deviation between the measured level output by the level gauge and the level calculated by the digital twin over a historical operating cycle. These records reflect the long-term drift and systematic errors of the level gauge under different operating conditions. Operational history data includes the level gauge's measured values, environmental parameters, maintenance records, etc., over different time periods, used to analyze its performance evolution. Aging characteristic data describes the performance degradation pattern of the sensor over time, such as decreased sensitivity and prolonged response time. Level gauge model parameters include the factory calibration data, technical specifications, and accuracy class of that level gauge model.

[0069] In practical applications, generating corresponding individualized systematic deviation correction parameters based on historical deviation records, operational history data, aging characteristic data, and level gauge model parameters refers to using data analysis and machine learning algorithms to comprehensively consider the above multi-dimensional information and generate a set of customized correction parameters for a specific reactor and level gauge. For example, a regression model can be constructed, taking historical deviations, operational data, and aging characteristics as inputs, and outputting a dynamic correction factor.

[0070] Among them, the initial reference true value is corrected by using individualized systematic deviation correction parameters to obtain the corrected initial reference true value. The purpose is to ensure that the obtained reference true value is not only based on the comparison under real-time stable operating conditions, but also fully considers the long-term characteristics and individual differences of the level gauge itself, thereby improving the accuracy and reliability of the reference true value.

[0071] This application also discloses a fault control system for a reactor level gauge based on the Industrial Internet, used to perform fault control of the reactor level gauge based on the Industrial Internet, combined with... Figure 3 As shown, the industrial internet-based reactor level gauge fault control system 1 includes: The credibility assessment module 11 is used to acquire the operating data of the reactor level gauge, the supplementary measurement data of the reactor, and the physical constraints of the reactor, and to assess the credibility of the operating data of the reactor level gauge to obtain the data credibility assessment result. The digital twin construction module 12 is used to construct a digital twin of the reactor and its corresponding level gauge, and to synchronize the reactor level gauge operation data, data reliability assessment results, reactor supplementary measurement data and reactor physical constraints to the digital twin. The abnormal state identification module 13 is used to identify the abnormal state of the liquid level gauge based on the operating data and data reliability assessment results of the liquid level gauge in the digital twin, and to distinguish the abnormal state of the liquid level gauge as either communication interruption or sensor body failure, thereby obtaining the distinction result of the abnormal state of the liquid level gauge. The liquid level measurement switching module 14 is used to respond to the abnormal state of the liquid level gauge, calculate the liquid level information of the reactor based on the supplementary measurement data and physical constraints of the reactor in the digital twin, and switch the feedback source of liquid level control from the measured liquid level output by the reactor liquid level gauge to the reactor liquid level information output by the digital twin. The control command generation module 15 is used to generate control commands for adjusting the reactor actuator based on the reactor liquid level information, and to transmit the control commands to the reactor actuator. The fault root cause determination module 16 is used to determine the fault root cause based on the operating data of the reactor level gauge, the supplementary measurement data of the reactor, and the differentiation results of the abnormal state of the level gauge. It also generates operation guidance information based on the fault root cause and sends the operation guidance information to the operation and maintenance terminal to guide the operation and maintenance terminal to check and handle the communication link, sensor body, or reactor operating condition. The liquid level recovery module 17 is used to restore the liquid level control feedback source from the digital twin to the liquid level output by the reactor liquid level gauge when the data credibility assessment result meets the preset recovery conditions and the deviation between the measured liquid level output by the reactor liquid level gauge and the reactor liquid level information is within the preset allowable range, according to the preset recovery observation period.

[0072] The reliability assessment module is used to acquire the operating data of the reactor level gauge, supplementary measurement data of the reactor, and physical constraints of the reactor, and to assess the reliability of the reactor level gauge operating data to obtain the data reliability assessment result. The specific methods for data acquisition and reliability assessment have been described in the above embodiments and will not be repeated here. It is important to emphasize that the reliability assessment module can be implemented as an independent software service or an embedded processing unit, capable of receiving data streams from field devices in real time and executing preset data verification and analysis algorithms to generate the data reliability assessment result.

[0073] The digital twin construction module is used to build digital twins of the reactor and its corresponding level gauge, and to synchronize the reactor level gauge's operating data, data reliability assessment results, supplementary measurement data of the reactor, and the reactor's physical constraints to the digital twin. The construction method and data synchronization mechanism of the digital twin have already been described in the above embodiments and will not be repeated here. It is important to emphasize that the digital twin construction module can be implemented as a cloud-based or edge-based modeling and data management platform, responsible for the initialization of the digital twin model, parameter configuration, and bidirectional data synchronization with the physical entity, ensuring the consistency between the virtual model and the physical entity's state.

[0074] The abnormal state identification module is used to identify abnormal states of the level gauge based on the operating data and data reliability assessment results of the reactor level gauge in the digital twin. It classifies these abnormal states into communication interruptions or sensor malfunctions, thus obtaining a classification result for the abnormal state of the level gauge. The specific methods for abnormal state identification and classification have already been described in the above embodiments and will not be repeated here. It is important to emphasize that the abnormal state identification module can be implemented as an intelligent analysis engine, capable of continuously monitoring the data in the digital twin and using pattern recognition, anomaly detection algorithms, and other technologies to perform real-time judgment and classification of the operating status of the level gauge.

[0075] The liquid level measurement switching module is used to respond to abnormal conditions of the liquid level gauge. Based on the supplementary measurement data and physical constraints of the reactor in the digital twin, it calculates the liquid level information of the reactor and switches the feedback source of the liquid level control from the measured liquid level output by the reactor level gauge to the reactor level information output by the digital twin. The specific methods for liquid level information calculation and feedback source switching have been described in the above embodiments and will not be repeated here. It should be emphasized that the liquid level measurement switching module can be implemented as a control logic unit. When an abnormality is detected in the liquid level gauge, it can automatically, or after operator confirmation, switch the liquid level feedback signal source of the control system from the physical liquid level gauge to the liquid level information calculated by the digital twin to maintain the normal operation of the control loop.

[0076] The control command generation module generates control commands for adjusting the reactor actuator based on the reactor liquid level information and transmits these commands to the reactor actuator. The specific method for generating control commands has been described in the above embodiments and will not be repeated here. It is important to emphasize that the control command generation module can be implemented as a software function block integrated into a DCS or PLC. Based on the liquid level information provided by the digital twin and combined with a preset control algorithm, it calculates and outputs adjustment commands for the reactor actuator in real time.

[0077] The fault root cause determination module is used to determine the root cause of the fault based on the operating data of the reactor level gauge, the supplementary measurement data of the reactor, and the differentiation results of the abnormal state of the level gauge. It then generates operation guidance information based on the root cause and sends this information to the maintenance terminal. The specific methods for determining the fault root cause and generating operation guidance information have already been described in the above embodiments and will not be repeated here. It is important to emphasize that the fault root cause determination module can be implemented as a diagnostic expert system or a rule-based inference engine, capable of comprehensively analyzing multi-source data and anomaly differentiation results to automatically locate the root cause of the fault and generate targeted maintenance suggestions.

[0078] The liquid level measurement recovery module is used to restore the liquid level control feedback source from the digital twin's output reaction vessel liquid level information to the measured liquid level output by the reaction vessel liquid level gauge when the data reliability assessment result meets the preset recovery conditions and the deviation between the measured liquid level output by the reaction vessel liquid level gauge and the reaction vessel liquid level information is within the preset allowable range, according to a preset recovery observation period. The specific methods for judging the recovery conditions and restoring the feedback source have been described in the above embodiments and will not be repeated here. It should be emphasized that the liquid level measurement recovery module can be implemented as a status monitoring and switching management unit, which continuously monitors the recovery status and data consistency of the liquid level gauge, and smoothly switches the liquid level control feedback source back to the physical liquid level gauge after meeting the safe recovery conditions and after a preset observation period.

[0079] The above are merely embodiments of this application and are not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A fault control method for a reactor level gauge based on the Industrial Internet, characterized in that, include: Acquire the operating data of the reactor level gauge, the supplementary measurement data of the reactor, and the physical constraints of the reactor, and conduct a reliability assessment on the operating data of the reactor level gauge to obtain the data reliability assessment result; A digital twin of the reactor and its corresponding level gauge is constructed, and the reactor level gauge operating data, the data reliability assessment results, the reactor supplementary measurement data, and the reactor physical constraints are synchronized to the digital twin; the digital twin refers to a digital model used to simulate the reactor level change and the level gauge response state. Based on the operating data of the reactor level gauge in the digital twin and the data reliability assessment results, the abnormal state of the level gauge is identified, and the abnormal state of the level gauge is classified into communication interruption or sensor body failure, thus obtaining the classification result of the abnormal state of the level gauge. In response to the abnormal state of the level gauge, based on the supplementary measurement data of the reactor in the digital twin and the physical constraints of the reactor, the reactor level information is calculated, and the feedback source of the level control is switched from the measured level output by the reactor level gauge to the reactor level information output by the digital twin; Based on the reactor liquid level information, a control command is generated for adjusting the reactor actuator, and the control command is transmitted to the reactor actuator; Based on the operating data of the reactor level gauge, the supplementary measurement data of the reactor, and the differentiation results of the abnormal state of the level gauge, the root cause of the fault is determined, and operation guidance information is generated based on the root cause of the fault. The operation guidance information is sent to the operation and maintenance terminal to guide the operation and maintenance terminal to check and handle the communication link, sensor body, or reactor operating condition. When the data reliability assessment result meets the preset recovery conditions, and the deviation between the measured liquid level output by the reactor level gauge and the reactor level information is within the preset allowable range, the feedback source of the liquid level control is restored from the reactor level information output by the digital twin to the measured liquid level output by the reactor level gauge according to the preset recovery observation period.

2. The method for fault control of a reactor level gauge based on the Industrial Internet according to claim 1, characterized in that, The step of identifying abnormal states of the level gauge based on the operating data of the reactor level gauge in the digital twin and the data reliability assessment results, and classifying the abnormal states of the level gauge into communication interruption or sensor failure, and obtaining the classification result of the abnormal state of the level gauge includes: Read the measured liquid level, echo signal strength, and timestamp information output by the liquid level gauge from the operating data of the reactor liquid level gauge; The supplementary measurement data and physical constraints of the reactor are read from the digital twin. The supplementary measurement data includes mass flow rate data, temperature data, pressure data, and stirring motor operating parameters. Obtain real-time operating status of the reactor; Based on the supplementary measurement data of the reactor, the physical constraints of the reactor, and the real-time operating conditions of the reactor, a physical behavior baseline of the reactor is constructed. The physical behavior baseline is used to calculate the liquid level change trajectory and echo signal intensity range of the level gauge under healthy conditions. Based on the data reliability assessment results, the measured liquid level output by the level gauge, the echo signal strength, the timestamp information, and the physical behavior baseline are compared to obtain the level gauge output deviation, which includes communication delay, signal smoothness index, and echo attenuation slope. When the timestamp information is interrupted or the communication delay exceeds a preset communication threshold, the abnormal state of the level gauge is classified as a communication interruption, and a communication interruption classification result is generated. When the timestamp information is normal, and the signal smoothness index or the echo attenuation slope exceeds the preset signal threshold, the abnormal state of the level gauge is classified as a sensor body failure, and a sensor body failure classification result is generated.

3. The method for fault control of a reactor level gauge based on the Industrial Internet according to claim 2, characterized in that, The step of constructing a baseline of the physical behavior of the reactor based on supplementary measurement data, physical constraints of the reactor, and real-time operating conditions of the reactor includes: Based on the data reliability assessment results, the echo signal strength, and the timestamp information, determine whether the operating data of the reactor level gauge meets the reference true value selection conditions; When the real-time operating condition of the reactor indicates that the reactor is in a stable operating condition, and the operating data of the reactor level gauge meets the reference true value selection conditions, the measured liquid level output by the level gauge is taken as the reference true value. Using the reference true value, the supplementary measurement data of the reactor, and the physical constraints of the reactor, the actual effective geometric parameters or actual effective material physical property parameters of the reactor are calculated in reverse. The actual effective geometric parameters or actual effective material physical property parameters obtained by reverse calculation are updated to the digital twin to obtain the updated digital twin; Based on the updated digital twin, a baseline of the physical behavior of the reactor is constructed.

4. The method for fault control of a reactor level gauge based on the Industrial Internet according to claim 2, characterized in that, The step of comparing the measured liquid level output by the level gauge, the echo signal strength, the timestamp information, and the physical behavior baseline, based on the data reliability assessment results, to obtain the liquid level gauge output deviation includes: Acquire supplementary state data inside the reactor, including bubble entrainment degree, solid particle suspension concentration, and agitator operation status; The echo signal feature extraction strategy is adjusted based on the degree of bubble entrainment and the concentration of suspended solid particles. Based on the adjusted echo signal feature extraction strategy, and combined with the echo signal intensity corresponding to the echo signal's spectral distribution, waveform distortion degree, and signal-to-noise ratio, echo signal features are extracted. Based on the operating state of the agitator, the physical behavior baseline is corrected to obtain the corrected physical behavior baseline; Based on the data reliability assessment results, the measured liquid level output by the level gauge, the timestamp information, and the echo signal characteristics are compared with the corrected physical behavior baseline to obtain the liquid level change deviation, time response deviation, and echo characteristic deviation. The liquid level gauge output deviation is quantified based on the liquid level change deviation, the time response deviation, and the echo characteristic deviation.

5. The method for fault control of a reactor level gauge based on the Industrial Internet according to claim 3, characterized in that, The step of using the measured liquid level output by the liquid level gauge as the reference true value when the real-time operating condition of the reactor indicates that the reactor is in a stable operating condition and the operating data of the reactor level gauge meets the reference true value selection conditions includes: Based on the real-time operating conditions of the reactor, the supplementary measurement data of the reactor, and the physical constraints of the reactor, physical constraint verification conditions for selecting the reference true value are determined. The physical constraint verification conditions include the reasonable range of liquid level change, the reasonable range of time response, and the reasonable range of echo signal. The measured liquid level output by the level gauge, the echo signal strength, the timestamp information, and the physical constraint verification conditions are compared to obtain multiple deviation indicators, including time delay, smoothness index, and echo attenuation slope. Based on the data reliability assessment results, it is determined whether the operating data of the reactor level gauge meets the data reliability conditions; Determine whether the deviation index exceeds a preset decrease threshold; When the operating data of the reactor level gauge meets the data reliability condition and all the deviation indicators do not exceed the preset drop threshold, it is determined that the operating data of the reactor level gauge meets the reference true value selection condition, and the measured liquid level output by the level gauge is used as the reference true value.

6. The method for fault control of a reactor level gauge based on the Industrial Internet according to claim 5, characterized in that, The step of determining whether the deviation index exceeds a preset decrease threshold includes: Obtain the material batch information and reactor identification information for the current reactor; Based on the material batch information and the reactor identification information, a threshold adjustment parameter matching the real-time operating condition of the reactor is retrieved from a preset parameter set; The preset decreasing threshold is corrected according to the threshold adjustment parameter to obtain the corrected preset decreasing threshold; The deviation index is compared with the corrected preset decline threshold to determine whether the deviation index exceeds the corrected preset decline threshold.

7. The method for fault control of a reactor level gauge based on the Industrial Internet according to claim 3, characterized in that, The step of using the measured liquid level output by the liquid level gauge as the reference true value when the real-time operating condition of the reactor indicates that the reactor is in a stable operating condition and the operating data of the reactor level gauge meets the reference true value selection conditions includes: When the real-time operating condition of the reactor indicates that the reactor is in a stable operating condition, and the operating data of the reactor level gauge meets the reference true value selection conditions, the measured liquid level output by the level gauge is compared with the theoretical liquid level value calculated in the digital twin. When the deviation between the measured liquid level output by the level gauge and the theoretical liquid level value remains stable within a preset time window and is within a preset deviation range, the measured liquid level output by the level gauge is confirmed as the initial reference true value. Based on the initial reference true value, and in conjunction with historical deviation records and liquid level gauge model parameters, a systematic deviation correction is performed on the initial reference true value to obtain the corrected initial reference true value; the historical deviation records are used to represent the deviation records formed between the measured liquid level output by the liquid level gauge and the liquid level calculated by the digital twin during the historical operating cycle; Use the corrected initial reference truth value as the reference truth value.

8. The method for fault control of a reactor level gauge based on the Industrial Internet according to claim 7, characterized in that, The step of comparing the measured liquid level output by the level gauge with the theoretical liquid level value calculated in the digital twin includes: Based on the supplementary measurement data of the reactor, real-time material composition information, real-time temperature data, real-time pressure data, and inlet and outlet mass flow rate data are obtained; Based on the real-time material composition information, the real-time temperature data, and the real-time pressure data, determine the real-time density and real-time viscosity of the current material mixture; Based on the real-time density, the real-time viscosity, the physical constraints of the reactor, and the inlet and outlet mass flow rate data, the material balance estimation model in the digital twin is corrected; The theoretical liquid level of the reactor is calculated using the modified material balance estimation model. The measured liquid level output by the level gauge is compared with the theoretical liquid level value.

9. A method for fault control of a reactor level gauge based on the Industrial Internet, as described in claim 7, is characterized in that, The step of systematically correcting the initial reference true value by combining historical deviation records and level gauge model parameters to obtain the corrected initial reference true value includes: Obtain the identification information of the reaction vessel and the level gauge; Based on the identification information of the reactor and the level gauge, retrieve historical deviation records, operating history data, aging characteristic data and level gauge model parameters that match the current reactor and level gauge from the digital twin; Based on the historical deviation records, the operational history data, the aging characteristic data, and the level gauge model parameters, corresponding individualized systematic deviation correction parameters are generated. The initial reference true value is corrected using the individualized systematic bias correction parameter to obtain the corrected initial reference true value.

10. A fault control system for a reactor level gauge based on the Industrial Internet, used to perform fault control of the reactor level gauge based on the Industrial Internet, characterized in that, include: The credibility assessment module is used to acquire the operating data of the reactor level gauge, the supplementary measurement data of the reactor, and the physical constraints of the reactor, and to assess the credibility of the operating data of the reactor level gauge to obtain the data credibility assessment result. A digital twin construction module is used to construct a digital twin of a reactor and its corresponding level gauge, and to synchronize the reactor level gauge operation data, the data reliability assessment results, the reactor supplementary measurement data, and the reactor physical constraints to the digital twin. An abnormal state identification module is used to identify abnormal states of the level gauge based on the operating data of the reactor level gauge in the digital twin and the data reliability assessment results, and to classify the abnormal state of the level gauge into communication interruption or sensor body failure, thereby obtaining the classification result of the abnormal state of the level gauge. The liquid level measurement switching module is used to respond to the abnormal state of the liquid level gauge, calculate the liquid level information of the reactor based on the supplementary measurement data of the reactor in the digital twin and the physical constraints of the reactor, and switch the feedback source of liquid level control from the measured liquid level output by the reactor liquid level gauge to the reactor liquid level information output by the digital twin; The control command generation module is used to generate control commands for adjusting the reactor actuator based on the reactor liquid level information, and to transmit the control commands to the reactor actuator. The fault root cause determination module is used to determine the fault root cause based on the operating data of the reactor level gauge, the supplementary measurement data of the reactor, and the differentiation results of the abnormal state of the level gauge, and generate operation guidance information based on the fault root cause, and send the operation guidance information to the operation and maintenance terminal to guide the operation and maintenance terminal to check and handle the communication link, sensor body or reactor operating condition. The liquid level measurement recovery module is used to restore the liquid level control feedback source from the reactor liquid level information output by the digital twin to the measured liquid level output by the reactor liquid level gauge according to a preset recovery observation period when the data reliability assessment result meets the preset recovery conditions and the deviation between the measured liquid level output by the reactor liquid level gauge and the reactor liquid level information is within a preset allowable range.

Images