Distributed monitoring method for triacetin industrial production line

Abstract

The application is particularly a distributed monitoring method for a triacetin industrial production line, relates to the technical field of chemical production process control, and comprises the following steps: synthesizing a physical hard threshold based on system measurement uncertainty, combining residual super-threshold duration to determine flowmeter soft failure, and classifying failure causes according to the size relationship between measured values and theoretical values. In the application, a heat-mass coupling mechanism verification model is constructed based on the law of thermodynamic conservation, thereby breaking through the technical limitation of traditional DCS / PLC systems that can only identify dominant faults such as sensor disconnection and over-range, and accurately capturing the in-range hidden drift faults of flowmeters caused by acetic acid corrosion and byproduct scaling; compared with a machine learning diagnosis scheme that relies on historical data, the method does not require training samples, the control logic is completely interpretable, and is suitable for the working condition fluctuation characteristics of triacetin production.

Description

Technical Field

[0001] This invention relates to the field of chemical production process control technology, and in particular to a distributed monitoring method for an industrial production line of triacetylglycerol. Background Technology

[0002] As an important fine chemical product, triacetin is widely used in food additives, pharmaceutical excipients, plasticizers and other fields. Its product quality directly depends on the accuracy of the feed ratio of glycerol and acetic acid and the stability of the reaction temperature in the synthesis reaction.

[0003] Currently, industrial production lines generally use distributed control systems (DCS / PLC) that rely on measurement data from field instruments such as flow meters and temperature sensors for regulation.

[0004] However, the synthesis reaction of triacetylglycerol is a highly corrosive and scale-prone chemical process. During long-term operation, the inner wall of the flowmeter is susceptible to corrosion by acetic acid, leading to parameter drift. Furthermore, the adhesion of reaction byproducts can cause changes in the flow cross-section, resulting in implicit measurement deviations within the flow range. Traditional DCS / PLC systems can only identify explicit faults such as signal disconnection and exceeding the flow range, but cannot detect drift faults within the sensor's range. When a flowmeter experiences implicit drift, the control system adjusts the mixing ratio based on erroneous signals, leading to material imbalance, excessive product acid value, and decreased yield. Moreover, the fault is highly concealed, requiring shutdown and investigation only after a large number of defective products have been produced, resulting in significant economic losses.

[0005] The existing improvement approach has obvious flaws: although using high-precision imported instruments can delay the failure, it significantly increases the equipment cost and cannot fundamentally eliminate drift.

[0006] Therefore, a distributed monitoring method for industrial production lines of triacetylglycerol is proposed to address the aforementioned problems. Summary of the Invention

[0007] The purpose of this invention is to provide a distributed monitoring method for an industrial production line of triacetylglycerol in order to solve the above-mentioned problems.

[0008] To achieve the above objectives, the present invention adopts the following technical solution:

[0009] Distributed monitoring methods for industrial triacetylglycerol production lines include:

[0010] Based on the energy conservation law of the triacetylglycerol synthesis reaction, key variables, physical constants and discretized energy balance equations related to the reaction are hard-coded, and a heat-mass coupling mechanism verification model is constructed.

[0011] The material transport time delay was determined by step response experiments, and a dynamic queue was constructed to match historical flow data with the current temperature change rate, while time synchronization and data preprocessing were completed.

[0012] Obtain the measured value of the flow sensor and the theoretical value of the flow calculated in reverse, and calculate the absolute residual between the two;

[0013] Based on the system measurement uncertainty, a physical hard threshold is synthesized, and the duration of residual exceeding the threshold is combined to determine the flow meter soft fault. Based on the relationship between the measured value and the theoretical value, the cause of the fault is specifically classified.

[0014] After a fault is triggered, the PID controller feedback signal source is seamlessly switched to the virtual calculated value, a correction coefficient is generated and a notification is sent, and online quality monitoring is linked to maintain production.

[0015] Preferably, the construction of a heat-mass coupling mechanism verification model based on the energy conservation law of the triacetylglycerol synthesis reaction, hard-coding key variables, physical constants, and discretized energy balance equations related to the reaction, specifically includes:

[0016] Define production process variables:

[0017] Glycerin flow Acetic acid flow rate Heating steam flow rate Feed temperature Real-time temperature of the reactor ;

[0018] And physical constants:

[0019] Specific heat capacity at constant pressure of glycerol and acetic acid , Enthalpy of reaction for the formation of triacetylglycerol Heat dissipation coefficient of the reactor ;

[0020] Construct the fundamental equation of energy conservation: ;

[0021] in, The heat input for steam; The reaction is exothermic; This refers to the amount of heat dissipated from the reactor to the environment. This refers to the change in heat storage of materials within the reaction system.

[0022] Preferably, the method further includes the derivation of discrete formulas that can be calculated for the control system:

[0023] Each energy term is decomposed into an expression directly related to the input variables, and based on the sampling period of the DCS. After discretization, we get:

[0024] ;

[0025] in, The latent heat of vaporization for heating steam; The effective heat transfer area of ​​the reactor; The ambient temperature; The theoretical reaction rate is calculated based on the Arrhenius equation; Total mass of materials; For the specific heat capacity of the mixture; The discretized rate of temperature change .

[0026] Preferably, the step-response experiment is used to determine the material transport time delay and construct a dynamic queue to match historical flow data with the current temperature change rate, while simultaneously completing time synchronization and data preprocessing. Specifically, this includes:

[0027] Under stable production conditions, the glycerol flow rate is increased stepwise while simultaneously collecting reactor temperature data. A response curve for the rate of temperature change is plotted. The time difference between the start of the rising edge of the curve and the moment of the flow rate step is the [temperature change rate]. ;

[0028] Similarly, the time delays corresponding to the acetic acid flow rate and the steam flow rate were measured respectively. , , ;

[0029] Establish a first-in-first-out (FIFO) data queue for each flow variable;

[0030] Real-time traffic data is collected and written to the tail of the queue, while simultaneously reading from the head of the queue. Historical traffic data at any given time ;

[0031] Read historical traffic data Rate of temperature change calculated at the current time Pair them one by one to form flow rate-temperature change rate data pairs.

[0032] Preferably, the step of obtaining the measured value of the flow sensor and the theoretical flow value calculated in reverse, and calculating the absolute residual between the two, specifically includes:

[0033] Real-time acquisition of flow meter readings, acquisition frequency and sampling period. Maintain consistency;

[0034] The collected measurements are checked for range. If the value exceeds the flow meter's range, it is considered an invalid signal, triggering a visible fault alarm. If the value is within the range, it is marked as a valid measurement. .

[0035] Preferably, the method further includes:

[0036] Select the real-time temperature of the reactor Temperature change rate Heating steam flow rate Feed temperature ;

[0037] The discrete energy balance equation is rearranged into a linear equation in one variable concerning the target flow rate. Solving this equation using algebraic operations yields a unique solution for the target flow rate, which is the theoretically calculated value of the target flow rate under the current operating conditions. ;

[0038] Perform process constraint verification on the calculated values, if If the flow rate exceeds the allowable range of the process, the current operating condition is determined to be abnormal, and it will not be included in the residual calculation. At the same time, an abnormal operating condition prompt will be triggered. If it is within the process range, it will be marked as a valid estimated value.

[0039] Calculate the absolute deviation between the physical measurement value and the mechanism estimation value, i.e. The unit is consistent with the unit of flow rate.

[0040] Preferably, the process of synthesizing a physical hard threshold based on system measurement uncertainty, determining a flowmeter soft fault based on the duration of residual exceeding the threshold, and specifically classifying the fault cause based on the relationship between the measured value and the theoretical value, specifically includes:

[0041] We analyzed all the physical factors affecting residual calculation, including the basic error of the feed flow meter, the measurement error of the temperature sensor, the basic error of the steam flow meter, and the heat dissipation coefficient. The calibration error;

[0042] The root sum-of-squares method is used to synthesize the uncertainties of each individual term to obtain the threshold. ;

[0043] General fault diagnosis logic:

[0044] The judgment logic needs to combine the magnitude and duration of the residual to ensure the accuracy of fault diagnosis. Specific rules:

[0045] Normal state determination: If the residual values ​​of consecutive sampling periods all meet the following conditions... < If so, the flow sensor is considered to be working normally;

[0046] Soft fault determination: If the residual value satisfies > And the duration exceeds the safety threshold. If so, the flow meter is determined to have experienced non-lethal drift;

[0047] Instantaneous deviation handling: If > But the duration is less If the deviation is detected, it is determined to be a momentary fluctuation in operating conditions, and no fault alarm is triggered; only the deviation event is recorded.

[0048] Preferably, the method further includes specific diagnostics for triacetylglycerol production lines:

[0049] False high reading fault: If > If the soft fault judgment condition is met, then the flow meter reading is determined to be falsely high;

[0050] Low reading fault: If < If the soft fault determination criteria are met, the flow meter reading is determined to be falsely low.

[0051] Preferably, the step of seamlessly switching the PID controller feedback signal source to the virtual calculated value after triggering a fault, generating a correction coefficient and pushing a notification, and linking with online quality monitoring to maintain production, specifically includes:

[0052] When the system determines that the flow meter has a soft fault, it automatically triggers the control weight transfer process.

[0053] After the switch, the system forms a soft measurement control loop based on a thermodynamic model with the reactor temperature as the core.

[0054] The system calculates the ratio of the measured value to the estimated value in real time, i.e., the correction factor. Correction factor Characterizes the degree of drift of the flow meter.

[0055] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0056] 1. This invention, by constructing a thermo-mass coupling mechanism verification model based on the thermodynamic conservation law, overcomes the technical limitations of traditional DCS / PLC systems that can only identify explicit faults such as sensor disconnection and over-range. It can accurately capture implicit drift faults within the flow meter caused by acetic acid corrosion and by-product scaling. Compared with machine learning diagnostic schemes that rely on historical data, this method does not require training samples, the control logic is fully interpretable, and it is adapted to the operating condition fluctuation characteristics of triacetylglycerol production.

[0057] 2. This invention, through physical hard threshold discrimination and residual trend analysis, can accurately locate faults in their early stages, avoiding the control system from adjusting material ratios based on erroneous flow signals. It addresses the root causes of problems such as excessive acid value and yield fluctuations, significantly reducing the rate of defective products and raw material waste costs, and ensuring the stability and consistency of triacetylglycerol product quality. Attached Figure Description

[0058] Further details, features, and advantages of this application are disclosed in the following description of exemplary embodiments in conjunction with the accompanying drawings, in which:

[0059] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0060] Several embodiments of this application will now be described in more detail with reference to the accompanying drawings to enable those skilled in the art to implement this application. This application may be embodied in many different forms and for various purposes and should not be limited to the embodiments set forth herein. These embodiments are provided to make this application thorough and complete, and to fully convey the scope of this application to those skilled in the art. The embodiments described do not limit this application.

[0061] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It will be further understood that terms such as those defined in commonly used dictionaries shall be interpreted as having a meaning consistent with their meaning in the relevant field and / or the context of this specification, and shall not be interpreted in an idealized or overly formal sense unless expressly defined herein.

[0062] Example 1

[0063] Its specific implementation method is combined with the appendix Figure 1 Please provide a detailed explanation.

[0064] Appendix Figure 1 The flowchart of the distributed monitoring method for an industrial production line of triacetylglycerol provided in this embodiment of the invention shows the complete steps from constructing a thermal-mass coupling mechanism verification model to specifically classifying the causes of failures based on the relationship between measured and theoretical values.

[0065] In this embodiment, it includes:

[0066] Based on the energy conservation law of the triacetylglycerol synthesis reaction, key variables, physical constants and discretized energy balance equations related to the reaction are hard encoded, and a thermal-mass coupling mechanism verification model that does not rely on historical data is constructed.

[0067] Specifically, it includes:

[0068] The core objective of this step is to establish a deterministic verification model that does not rely on historical data and is based solely on physical conservation laws, providing a theoretical benchmark for subsequent sensor fault diagnosis. The model construction process must incorporate the process characteristics of the triacetylglycerol esterification reaction to ensure a high degree of fit between the equations and actual production conditions.

[0069] Input variable definition:

[0070] The selection of variables must meet the principle of covering the entire reaction energy input-transfer-output chain. The following key production process variables are set, and all variables correspond to the data acquisition points of the field instruments on the production line:

[0071] Glycerin flow This corresponds to the measurement value of the Coriolis mass flow meter on the glycerol feed line, in units of kg / h. This variable directly determines the molar ratio of the reactants.

[0072] Acetic acid flow rate The corresponding electromagnetic flowmeter reading on the acetic acid feed line is in kg / h. Excess acetic acid is a key condition for high yield of triacetylglycerol and needs to be precisely controlled.

[0073] Heating steam flow rate This corresponds to the vortex flow meter measurement value on the heating steam pipeline of the reactor jacket, with the unit being kg / h, and is the core energy input source of the reaction system.

[0074] Feed temperature , corresponding to the temperature sensor reading at the outlet of the feed mixer, in °C, characterizing the initial energy state of the reactants when they enter the reactor;

[0075] Real-time temperature of the reactor The value corresponds to the measurement of the armored thermocouple inside the reactor vessel, in °C, and is a direct characterization parameter of the energy balance state of the reaction system.

[0076] And physical constants:

[0077] The physical constants entered must be inherent properties of the triacetylglycerol synthesis reaction, and their values ​​must be derived from authoritative chemical property handbooks or precise laboratory measurement data. Furthermore, a constant calibration interface must be set in the system to facilitate fine-tuning based on batch differences in raw materials.

[0078] Specific heat capacity at constant pressure of glycerol and acetic acid , The unit is kJ / (kg·℃), and the value must match the temperature range of the production conditions (e.g., 25-150℃) to avoid calculation deviations in the high-temperature reaction stage caused by constant values ​​at room temperature;

[0079] enthalpy of reaction for the formation of triacetylglycerol The unit is kJ / mol, and this reaction is exothermic. The value is negative and is determined by differential scanning calorimetry (DSC), characterizing the heat change per unit amount of reactant when it is converted into product.

[0080] Heat dissipation coefficient of the reactor The unit is The coefficient needs to be calibrated through a reactor heat loss experiment. The specific method is as follows: under the condition of no feed and no reaction, a constant flow of steam is introduced into the jacket, the temperature change curve of the reactor body over time is recorded, and the coefficient is calculated by combining the heat transfer formula. The coefficient decay caused by the aging of the reactor body insulation layer should be taken into account, and an annual calibration cycle should be set. This process is a direct reference to the existing technology and will not be elaborated here.

[0081] Core verification equation:

[0082] A real-time equilibrium equation for the reaction system must be established based on the law of conservation of energy. The equation needs to cover four major energy terms: heat input from steam, heat released from the reaction, heat dissipation from the system, and heat storage by materials. At the same time, it needs to be converted into a discretized formula to adapt to the periodic operation logic of DCS / PLC.

[0083] Construct the fundamental equation of energy conservation: ;

[0084] in,

[0085] The heat input for steam;

[0086] The reaction is exothermic;

[0087] This refers to the amount of heat dissipated from the reactor to the environment.

[0088] This refers to the change in heat storage of materials within the reaction system.

[0089] It also includes the derivation of discrete formulas that can be computed for the control system:

[0090] Each energy term is decomposed into an expression directly related to the input variables, and based on the sampling period of the DCS. Discretize (e.g., 1s) to obtain:

[0091] ;

[0092] in,

[0093] The latent heat of vaporization of heating steam, expressed in kJ / kg, is determined based on the steam pressure.

[0094] The effective heat transfer area of ​​the reactor is expressed in units of... The parameters are determined by the structural parameters of the vessel body;

[0095] The ambient temperature is measured in °C and is collected in real time by temperature sensors within the workshop.

[0096] The theoretical reaction rate is calculated based on the Arrhenius equation, in mol / h. The calculation formula is as follows: , Pre-exponential factor, The activation energy of the reaction. The gas constant is , These represent the molar concentrations of glycerol and acetic acid in the reaction vessel, respectively.

[0097] The total mass of materials is the total mass of all materials (reactants, products, and solvents) in the reactor. It represents the total amount of heat storage material in the system, expressed in kg, and is calculated from the cumulative feed flow rate and the product discharge flow rate. It reflects the basic heat storage capacity of the system and is a parameter relating the rate of temperature change to the amount of heat stored.

[0098] The specific heat capacity of the mixture is the average specific heat capacity of the mixture in the reactor, which reflects the heat required to raise the temperature of a unit mass of material by 1°C. The unit is kJ / (kg・°C). It is obtained by mass-weighted average of the specific heat capacities of glycerol, acetic acid, and triacetic acid glyceride. It quantifies the heat storage capacity of the mixture and is a key coefficient for converting the rate of temperature change into heat storage.

[0099] The discretized rate of temperature change This represents the rate of change in heat storage of the reaction system per unit time. This discrete formula needs to be permanently written into the DCS's computation module, with the computation cycle matching the sampling cycle to ensure real-time performance.

[0100] The material transport time delay was determined by step response experiment and a dynamic queue was constructed to achieve accurate matching between historical flow data and current temperature change rate. At the same time, median filtering was used to remove glitch from the flow signal and retain the drift trend, thus completing time synchronization and data preprocessing.

[0101] Specifically, it includes:

[0102] Because there is a physical transmission delay in the material from the flow meter detection point to the reaction zone of the reactor, and the original signal collected by the field instrument is subject to noise interference, this step requires timing compensation and signal preprocessing to eliminate the impact of data distortion on the subsequent verification results.

[0103] Time delay compensation:

[0104] The core of time delay compensation is to determine the material transfer time delay. This allows for precise matching of historical flow data with the current temperature change rate. The specific operation process is as follows:

[0105] A step response experiment was conducted to determine the reaction rate. Under stable production conditions, the glycerol flow rate was increased by 10% in a step, while the reactor temperature data was collected at a high frequency of 10Hz. The response curve of the temperature change rate was plotted, and the time difference between the start of the rising edge of the curve and the moment of the flow rate step was determined. ;

[0106] Similarly, the time delays corresponding to the acetic acid flow rate and the steam flow rate were measured respectively. , , And store them separately in the system;

[0107] In the DCS, a first-in-first-out (FIFO) data queue is established for each flow variable, with a queue length of... (Round up), for example when =10s When the time is 1 second, the queue length is 10.

[0108] The system collects traffic data in real time and writes it to the tail of the queue, while simultaneously reading it from the head of the queue. Historical traffic data at any given time ;

[0109] Read historical traffic data Rate of temperature change calculated at the current time One-to-one pairing is performed to form flow rate-temperature change rate data pairs, ensuring that the two parameters involved in the verification correspond to the same reaction condition in the time dimension;

[0110] The goal of signal preprocessing is to remove transient noise and retain effective trends. The process requires lossless filtering algorithms to avoid filtering out key features of sensor drift. Specific operations include:

[0111] Deburring: The original flow signal is processed using a median filtering algorithm, with the filter window size set to... ( (Select 3~5), the specific choice depends on the typical duration of the on-site pressure fluctuation; the core operation of median filtering is: select the value within the window. The algorithm sorts data points by size and takes the median value as the output value of the window. This algorithm can effectively remove spike signals caused by instantaneous fluctuations in pipeline pressure (such as abnormally high / low values ​​that last for 1 to 2 sampling cycles) and does not smooth out the trend changes of the signal.

[0112] Trend Preservation Verification: After filtering, the system needs to calculate the linear fitting slope of the signal and compare the rate of change of the slope before and after filtering. If the rate of change is less than 5%, the filtering is deemed effective; if it is greater than 5%, the filtering window size is automatically reduced to ensure that the slow drift trend of the sensor (such as a daily drift of 0.1%) is completely preserved. This trend is the core detection object for subsequent soft fault diagnosis.

[0113] The system acquires the measured values ​​from the flow sensor and the theoretical flow values ​​calculated in reverse based on stable parameters such as temperature in parallel, calculates the absolute residual between the two, realizes cross-verification of physical measurement and mechanism calculation, and reconstructs the virtual instrument.

[0114] Specifically, it includes:

[0115] Logic A (Acquisition of physical measurement values):

[0116] The core of this logic is to ensure the authenticity and integrity of the measured values. Specific operations include:

[0117] Signal Acquisition and Transmission: The DCS acquires the flow meter's measured values ​​in real time via a 4~20mA current signal or the PROFIBUS bus protocol, with varying acquisition frequency and sampling period. To maintain consistency, shielded cables are used in signal transmission lines to improve anti-interference capabilities, and the shielding layer is grounded at one end.

[0118] Data validity assessment: The system performs range verification on the collected measurements. If the value exceeds the flow meter's range (e.g., 0~1000 kg / h), it is determined to be an invalid signal, directly triggering a visible fault alarm; if it is within the range, it is marked as a valid measurement value. .

[0119] Also includes:

[0120] Logic B (mechanism estimation calculation):

[0121] The core of this logic is to use more stable temperature sensor data to inversely calculate the theoretical value of the flow rate. The specific operation is as follows:

[0122] Select the real-time temperature of the reactor Temperature change rate Heating steam flow rate Feed temperature With parameters such as thermocouples, RTDs, and steam flow meters as inputs, the corresponding sensors are less affected by corrosion and scaling, and their operational stability is higher than that of feed flow meters.

[0123] The discrete energy balance equation is rearranged to be related to the target flow rate (e.g.) The linear equation in one variable, taking glycerol flow rate as an example, can be simplified to the following form: In the formula, , The coefficients related to other known parameters are solved through algebraic operations to obtain the target flow rate (e.g., ...). The unique solution to ) is the target flow rate under the current operating conditions (e.g., Theoretical estimated value ;

[0124] In the formula, , The specific methods for obtaining it are as follows:

[0125] Primitive energy balance equation Substitute the reaction rate The expression, because The reaction rate is based on the Arrhenius equation, where the molar concentration of glycerol is... With glycerol flow Directly related ( ∝ ),therefore It can be represented as: , It is a composite coefficient composed of known parameters such as pre-exponential factor, activation energy, temperature, and acetic acid concentration;

[0126] Summarized as follows: Compare with the standard form of a linear equation in one variable. We can obtain:

[0127] ;

[0128] ;

[0129] The physical meaning of 'a': It reflects the degree to which changes in glycerol flow rate affect the heat released by the reaction. It is determined by both the kinetic and thermodynamic characteristics of the reaction and is a dynamic coefficient related to the operating conditions.

[0130] The physical meaning of 'b': It reflects the comprehensive energy difference between steam input, heat loss, and heat storage changes under the current operating conditions. At that time, the energy of the entire reaction system reaches equilibrium.

[0131] Perform process constraint verification on the calculated values, if If the flow rate exceeds the allowable range of the process (e.g., the normal flow rate range of glycerin is 200~300 kg / h), the current operating condition is determined to be abnormal and will not be included in the residual calculation, while triggering an abnormal operating condition prompt; if it is within the process range, it is marked as a valid estimated value.

[0132] Example explanation: When the heating steam flow rate is stable at 80 kg / h, the reaction vessel temperature change rate is 0.3℃ / min, and the ambient temperature is 25℃, substituting these values ​​into the energy balance equation and solving in reverse, the theoretical value of the glycerol flow rate should be 250 kg / h. =250kg / h; if the physical flow meter measures 270kg / h at this time, there is a significant deviation between the two.

[0133] Residual generation:

[0134] Residual error is a core indicator for measuring the deviation between sensor measurements and theoretical values. Specific calculation and storage rules are as follows:

[0135] Calculate the absolute deviation between the physical measurement value and the mechanism estimation value, i.e. The unit is consistent with the flow rate unit;

[0136] The system establishes a residual data sequence and stores the residual values ​​of 100 consecutive sampling periods in chronological order, forming a residual change trend curve. This facilitates the analysis of the continuous characteristics of the deviation during subsequent fault diagnosis and avoids misjudgment caused by a single instantaneous deviation.

[0137] Based on the system measurement uncertainty, a physical hard threshold is synthesized. Combined with the duration of residual exceeding the threshold, a soft fault of the flow meter is determined. Based on the relationship between the measured value and the theoretical value, the fault is specifically classified as either acetic acid corrosion or by-product scaling.

[0138] Specifically, it includes:

[0139] This step abandons the traditional statistical probability discrimination method and adopts a hard threshold discrimination logic based on the inherent physical characteristics of the system to achieve accurate diagnosis and specific classification of sensor faults.

[0140] Physical hard threshold setting:

[0141] threshold The value of must be strictly based on the combined result of the system measurement uncertainty to ensure the objectivity and rationality of the threshold. Specific calculation method:

[0142] We analyzed all the physical factors affecting residual calculation, including the basic error of the feed flow meter, the measurement error of the temperature sensor, the basic error of the steam flow meter, and the heat dissipation coefficient. The calibration error;

[0143] The root summation method is used to synthesize the uncertainties of each individual term. ,in , , , These are the basic error of the feed flow meter, the measurement error of the temperature sensor, the basic error of the steam flow meter, and the heat dissipation coefficient. The uncertainty corresponding to the calibration error is used to obtain the threshold. ;

[0144] The random error range reflecting the measured flow rates of glycerol and acetic acid is the fundamental measurement error of the flow sensor itself.

[0145] Basis for value determination: Derived from the accuracy class of the flow meter. For example, if the basic error of the flow meter is ±0.5% of full scale (FS), under the assumption of uniform distribution (errors occur with equal probability within the range), the standard uncertainty is: ;

[0146] The random error range reflecting the measured values ​​of reactor temperature and feed temperature is the measurement error of the temperature sensor.

[0147] Basis for value selection: Derived from the accuracy class of the temperature sensor. For example, if the measurement error of the thermocouple is ±0.2°C, under the assumption of uniform distribution, the standard uncertainty is: ;

[0148] The random error range reflecting the measured value of heating steam flow rate is the basic measurement error of the steam flow meter.

[0149] Basis for value selection: Derived from the accuracy class of the steam flow meter. For example, if the basic error of the vortex flow meter is ±1.0%FS, under the assumption of uniform distribution, the standard uncertainty is: ;

[0150] The random error range reflecting the calibration process of the heat dissipation coefficient of the reactor is the measurement and fitting error of the heat loss experiment.

[0151] The value is derived from the calibration error of the heat dissipation coefficient. For example, if... The calibration error is ±3%. Assuming a uniform distribution, the standard uncertainty is: .

[0152] The system dynamically adjusts the threshold according to changes in production conditions. For example, when the production load increases to 120% of the rated load, the basic error of the flow meter will increase slightly. The system automatically amplifies the threshold proportionally to avoid misjudgment under high load conditions.

[0153] General fault diagnosis logic:

[0154] The judgment logic needs to combine the magnitude and duration of the residual to ensure the accuracy of fault diagnosis. Specific rules:

[0155] Normal state determination: If the residual values ​​of 10 consecutive sampling periods all meet the requirements... < If the flow sensor is found to be in normal working condition, the system maintains the original control logic and the residual sequence continues to be updated.

[0156] Soft fault determination: If the residual value satisfies > Furthermore, the duration of this deviation exceeds the safety threshold. If the flow meter experiences a non-lethal drift (i.e., a soft fault), then it is determined that the flow meter has experienced a non-lethal drift. The setting is based on the maximum allowable duration of process deviation. For example, if the deviation lasts for more than 30 seconds, it will cause the reactant ratio to deviate from the process window. Set to 30 seconds;

[0157] Instantaneous deviation handling: If > But the duration is less If the deviation is detected, it is determined to be a momentary fluctuation in operating conditions. No fault alarm is triggered, but the deviation event is recorded for subsequent system optimization analysis.

[0158] It also includes specific diagnostics for triacetylglycerol production lines:

[0159] Based on the corrosiveness of materials and the scaling characteristics of byproducts in the triacetylglycerol synthesis reaction, the types and causes of drift failures are accurately classified, providing clear guidance for subsequent maintenance. Specific classification rules are as follows:

[0160] False high reading fault: If > If the soft fault judgment conditions are met, the flow meter reading is determined to be falsely high. The cause of the fault is located as long-term corrosion of the flow meter by acetic acid causing the inner wall of the flow meter to become thinner, or the electrode polarization parameters of the electromagnetic flow meter drifting, resulting in the measured value being higher than the actual flow rate. The typical characteristic of this fault is that the residual increases linearly with the increase of operating time.

[0161] Low reading fault: If < If the soft fault judgment conditions are met, the flow meter reading is determined to be falsely low. The cause of the fault is located as follows: reaction byproducts (such as glycerol diacetate, polymer) adhere to the inner wall of the flow meter, resulting in a narrowing of the flow cross section, or scale buildup in the measuring tube of the Coriolis flow meter causes a shift in the vibration frequency, which in turn makes the measured value lower than the actual flow rate. The typical characteristic of this fault is that the residual error increases exponentially with the increase of operating time.

[0162] After a fault is triggered, the PID controller feedback signal source is switched to the virtual calculated value without disturbance, a correction coefficient is generated and a human-machine prompt is pushed, and online quality monitoring is linked to maintain production. The original control logic is restored after the flow meter is repaired and calibrated.

[0163] Specifically, it includes:

[0164] The control weight transfer employs a disturbance-free switching algorithm to achieve a smooth transition of the control signal source, avoiding production fluctuations caused by sudden changes in flow. Specific operation:

[0165] Switching trigger conditions: When the system determines that the flow meter has a soft fault, it automatically triggers the control weight transfer process without manual intervention;

[0166] Bumpy switching is achieved by recording the current output value of the PID controller at the moment of switching and transferring the controller's feedback signal source (PV) from the physical flow meter measurement value. Switch to mechanism estimation value At the same time, the controller's manual / automatic switching module is locked in automatic mode to avoid sudden changes in output values;

[0167] Control loop reconfiguration: After switching, the system forms a soft measurement control loop based on a thermodynamic model with the reactor temperature as the core. The control logic is adjusted to maintain the reactor temperature at the process set value by adjusting the theoretical values ​​of steam flow and feed flow, thereby ensuring the accuracy of material proportioning.

[0168] Switching stability verification: The system monitors the rate of change of the controller output value in real time. If the rate of change is less than 5% / s, the switching is determined to be disturbance-free; if it is greater than 5% / s, the output limiting is automatically triggered to ensure the stability of the production process.

[0169] Correction coefficient generation and human-computer interaction prompts:

[0170] Generate correction coefficients and push notification messages to provide data support for subsequent sensor repair and calibration. Specific steps are as follows:

[0171] Correction factor calculation: The system calculates the ratio of the measured value to the estimated value in real time, which is the correction factor. Correction factor Characterizing the degree of drift of the flow meter, for example when When the value is 1.08, it indicates that the flow meter's measured value is 8% lower than the actual value.

[0172] Human-Machine Interface (HMI) prompts: A fault prompt window pops up on the HMI, displaying the following information: faulty sensor location (e.g., glycerol feed flow meter), fault type (e.g., false low reading), correction factor k, and fault duration; at the same time, the faulty sensor is marked with a flashing red mark on the flowchart interface to facilitate quick location by the operator;

[0173] Correction coefficient storage: The system stores the correction coefficient in association with the fault occurrence time and operating parameters to form a fault file, which is used for subsequent analysis of sensor drift patterns and optimization of maintenance cycles.

[0174] Online quality monitoring linkage: The self-healing control loop is linked with an online quality analyzer (such as a near-infrared spectrometer) to monitor the acid value and ester content of the product in real time. If the index deviates from the process range, the system automatically adjusts the correction coefficient of the calculated value, forming a control-detection-correction process.

[0175] Maintenance timing recommendations: The system pushes the best maintenance timing prompts on the HMI based on the current production batch progress. For example, if the current batch has 2 hours of remaining production time, it is recommended to stop the machine for maintenance after the batch is completed to avoid product waste caused by mid-process shutdown.

[0176] Switching back after fault recovery: After the flowmeter maintenance and calibration are completed, the operator confirms the recovery signal on the HMI, and the system automatically executes the reverse switching process: first, the control weights are gradually transferred from the calculated values ​​to the measured values, and then the residual value is stabilized. < After the status is restored, the original control logic is restored, the fault markers are cleared, and the fault file is updated.

[0177] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0178] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0179] It should be noted that, in this document, the use of relational terms such as "first" and "second" is merely for distinguishing one entity or operation from another, and does not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "include," "contain," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the statement "includes a…" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0180] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0181] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0182] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0183] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0184] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0185] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0186] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A distributed monitoring method for an industrial production line of triacetylglycerol, characterized in that, include: Based on the energy conservation law of the triacetylglycerol synthesis reaction, key variables, physical constants and discretized energy balance equations related to the reaction are hard-coded, and a heat-mass coupling mechanism verification model is constructed. The material transport time delay was determined by step response experiments, and a dynamic queue was constructed to match historical flow data with the current temperature change rate, while time synchronization and data preprocessing were completed. Obtain the measured value of the flow sensor and the theoretical value of the flow calculated in reverse, and calculate the absolute residual between the two; Based on the system measurement uncertainty, a physical hard threshold is synthesized, and the duration of residual exceeding the threshold is combined to determine the flow meter soft fault. Based on the relationship between the measured value and the theoretical value, the cause of the fault is specifically classified. After a fault is triggered, the PID controller feedback signal source is seamlessly switched to the virtual calculated value, a correction coefficient is generated and a notification is sent, and online quality monitoring is linked to maintain production.

2. The distributed monitoring method for an industrial production line of triacetylglycerol according to claim 1, characterized in that, Based on the energy conservation law of the triacetylglycerol synthesis reaction, key variables, physical constants, and discretized energy balance equations related to the reaction are hard-coded, and a heat-mass coupling mechanism verification model is constructed, specifically including: Define production process variables: Glycerin flow Acetic acid flow rate Heating steam flow rate Feed temperature Real-time temperature of the reactor ; And physical constants: Specific heat capacity at constant pressure of glycerol and acetic acid , Enthalpy of reaction for the formation of triacetylglycerol Heat dissipation coefficient of the reactor ; Construct the fundamental equation of energy conservation: ; in, The heat input for steam; The reaction is exothermic; This refers to the amount of heat dissipated from the reactor to the environment. This refers to the change in heat storage of materials within the reaction system.

3. The distributed monitoring method for an industrial production line of triacetylglycerol according to claim 2, characterized in that, It also includes the derivation of discrete formulas that can be computed for the control system: Each energy term is decomposed into an expression directly related to the input variables, and based on the sampling period of the DCS. After discretization, we get: ； in, The latent heat of vaporization for heating steam; The effective heat transfer area of ​​the reactor; The ambient temperature; The theoretical reaction rate is calculated based on the Arrhenius equation; Total mass of materials; For the specific heat capacity of the mixture; The discretized rate of temperature change .

4. The distributed monitoring method for an industrial production line of triacetylglycerol according to claim 1, characterized in that, Material transport time delay was determined through step response experiments, and a dynamic queue was constructed to pair historical flow data with the current temperature change rate. Simultaneously, time synchronization and data preprocessing were performed, specifically including: Under stable production conditions, the glycerol flow rate is increased stepwise while simultaneously collecting reactor temperature data. A response curve for the rate of temperature change is plotted. The time difference between the start of the rising edge of the curve and the moment of the flow rate step is the [temperature change rate]. ; Similarly, the time delays corresponding to the acetic acid flow rate and the steam flow rate were measured respectively. , , ; Establish a first-in-first-out (FIFO) data queue for each flow variable; Real-time traffic data is collected and written to the tail of the queue, while simultaneously reading from the head of the queue. Historical traffic data at any given time ; Read historical traffic data Rate of temperature change calculated at the current time Pair them one by one to form flow rate-temperature change rate data pairs.

5. The distributed monitoring method for an industrial production line of triacetylglycerol according to claim 1, characterized in that, Obtain the measured value from the flow sensor and the theoretical flow value calculated from the reverse calculation, and calculate the absolute residual between the two, specifically including: Real-time acquisition of flow meter readings, acquisition frequency and sampling period. Maintain consistency; The collected measurements are checked for range. If the value exceeds the flow meter's range, it is considered an invalid signal, triggering a visible fault alarm. If the value is within the range, it is marked as a valid measurement. .

6. The distributed monitoring method for an industrial production line of triacetylglycerol according to claim 5, characterized in that, Also includes: Select the real-time temperature of the reactor Temperature change rate Heating steam flow rate Feed temperature ; The discrete energy balance equation is rearranged into a linear equation in one variable concerning the target flow rate. Solving this equation using algebraic operations yields a unique solution for the target flow rate, which is the theoretically calculated value of the target flow rate under the current operating conditions. ; Perform process constraint verification on the calculated values, if If the flow rate exceeds the allowable range of the process, the current operating condition is determined to be abnormal, and it will not be included in the residual calculation. At the same time, an abnormal operating condition prompt will be triggered. If it is within the process range, it will be marked as a valid estimated value. Calculate the absolute deviation between the physical measurement value and the mechanism estimation value, i.e. The unit is consistent with the unit of flow rate.

7. The distributed monitoring method for an industrial production line of triacetylglycerol according to claim 1, characterized in that, Based on the system measurement uncertainty, a physical hard threshold is synthesized. Combined with the duration of residual exceeding the threshold, soft faults in the flowmeter are determined. Furthermore, based on the relationship between measured and theoretical values, specific fault causes are classified, including: We analyzed all the physical factors affecting residual calculation, including the basic error of the feed flow meter, the measurement error of the temperature sensor, the basic error of the steam flow meter, and the heat dissipation coefficient. The calibration error; The root sum-of-squares method is used to synthesize the uncertainties of each individual term to obtain the threshold. ; General fault diagnosis logic: The judgment logic needs to combine the magnitude and duration of the residual to ensure the accuracy of fault diagnosis. Specific rules: Normal state determination: If the residual values ​​of consecutive sampling periods all meet the following conditions... < If so, the flow sensor is considered to be working normally; Soft fault determination: If the residual value satisfies > And the duration exceeds the safety threshold. If so, the flow meter is determined to have experienced non-lethal drift; Instantaneous deviation handling: If > But the duration is less If the deviation is detected, it is determined to be a momentary fluctuation in operating conditions, and no fault alarm is triggered; only the deviation event is recorded.

8. The distributed monitoring method for an industrial production line of triacetylglycerol according to claim 7, characterized in that, It also includes specific diagnostics for triacetylglycerol production lines: False high reading fault: If > If the soft fault judgment condition is met, then the flow meter reading is determined to be falsely high; Low reading fault: If < If the soft fault determination criteria are met, the flow meter reading is determined to be falsely low.

9. The distributed monitoring method for an industrial production line of triacetylglycerol according to claim 1, characterized in that, Upon triggering a fault, the PID controller feedback signal source is seamlessly switched to the virtual calculated value, a correction coefficient is generated and a notification is pushed, and online quality monitoring is linked to maintain production. Specifically, this includes: When the system determines that the flow meter has a soft fault, it automatically triggers the control weight transfer process. After the switch, the system forms a soft measurement control loop based on a thermodynamic model with the reactor temperature as the core. The system calculates the ratio of the measured value to the estimated value in real time, i.e., the correction factor. Correction factor Characterizes the degree of drift of the flow meter.

Images