A gas-liquid separation and condensation recovery device for antimony-free polyester esterification process

Abstract

The present application relates to the technical field of polyester esterification process, and discloses a gas-liquid separation and condensation recovery device for antimony-free polyester esterification process, which realizes closed-loop optimization through four core modules: a flow field characteristic prediction module predicts the average diameter of entrained liquid droplets based on the high-frequency fluctuation characteristics of the import instantaneous differential pressure; a separation efficiency solving module dynamically corrects the critical liquid flooding gas velocity in combination with the above characteristics, calculates the actual separation efficiency and the mass flow rate of escaped droplets; a heat transfer attenuation prediction module predicts the dynamic coking dirt resistance of the condenser based on the escape flow rate and catalyst activity, and updates the real-time total heat transfer coefficient; and a comprehensive energy efficiency optimization module constructs an energy efficiency objective function covering product revenue and heat transfer attenuation penalty, and iteratively outputs the optimal cooling water control flow rate. The present application realizes the deep coupling of flow field sensing, coking prediction and energy efficiency optimization, and significantly improves the operation stability and comprehensive economic benefits of the device under dynamic working conditions.

Description

Technical Field

[0001] This invention belongs to the field of polyester esterification process technology, and particularly relates to a gas-liquid separation and condensation recovery device for antimony-free polyester esterification process. Background Technology

[0002] With the continuous improvement of global environmental standards, antimony-free polyester production processes, especially novel processes based on titanium-based catalysts, are rapidly replacing traditional antimony-containing technologies. While these catalysts effectively avoid the risk of heavy metal pollution, their inherent high catalytic activity and strong adhesion characteristics pose significant challenges in the esterification reaction stage. At the reactor outlet, the intense vaporization process of water and ethylene glycol generates a complex gas-liquid two-phase flow, which carries a large number of unreacted terephthalic acid particles and high-viscosity oligomers. These components are highly susceptible to the dynamic characteristics of the flow field, but the design concept of traditional gas-liquid separation and condensation devices is still limited to the steady-state conditions of antimony-based processes, generally employing separation internals with fixed geometric structures. Such devices fail to recognize the inherent pressure pulsation phenomenon in the esterification boiling system. This pulsation, originating from the nonlinear characteristics of reaction kinetics, continuously interferes with the gas flow shear force distribution and disrupts the stability of the liquid film. In actual operation, high-frequency pressure fluctuations induce secondary droplet breakage effects and exacerbate the risk of flooding in the demister, causing a large number of micron-sized oligomer droplets to escape from the separation boundary. After escaping droplets enter the condensation system, they undergo irreversible deposition on the high-temperature heat exchanger tube walls. Their high reactivity promotes rapid secondary polycondensation of oligomers, forming a dense and stubborn coking fouling layer. This fouling layer accumulation not only causes a continuous deterioration in heat transfer resistance and a significant decrease in condensation efficiency, but also disrupts the system's vacuum stability, destroys thermodynamic phase equilibrium conditions, and ultimately leads to color degradation and uneven molecular weight distribution in the polyester product. Furthermore, existing equipment lacks the capability for deep coupling analysis of the dynamic parameters of the gas-liquid two-phase fluid, the evolution trend of heat transfer attenuation, and the energy consumption of power equipment. This prevents the construction of a real-time adaptive control mechanism, resulting in frequent system instability and additional energy consumption during fluctuations in operating conditions. Therefore, the existing technology system is significantly inadequate in addressing the unique dynamic flow field and coking problems inherent in antimony-free polyester processes. Summary of the Invention

[0003] The purpose of this invention is to provide a gas-liquid separation and condensation recovery device for antimony-free polyester esterification process, aiming to solve the above-mentioned problems.

[0004] This invention is implemented as follows: a gas-liquid separation and condensation recovery device for antimony-free polyester esterification process, comprising: a flow field characteristic prediction module: acquiring the instantaneous differential pressure signal at the inlet of the gas-liquid separator and extracting its high-frequency fluctuation characteristics, and predicting the Souter mean diameter of entrained droplets by combining fluid physical property parameters; and a separation efficiency calculation module: calculating the total liquid phase flow rate entering the separator based on the ratio of gas flow pressure to surface tension, dynamically correcting the critical flooding velocity according to the high-frequency fluctuation characteristics, and calculating the actual gas-liquid separation efficiency and the mass flow rate of escaped droplets escaping to the condenser by combining the Souter mean diameter. The module includes: a heat transfer attenuation prediction module, a comprehensive energy efficiency optimization module, and a comprehensive energy efficiency optimization module. The module predicts the dynamic coking fouling thermal resistance of the condenser based on the escape droplet mass flow rate, characteristic activity parameters of the antimony-free catalyst, and intrinsic properties of the fouling, and updates its real-time total heat transfer coefficient. The module also includes a comprehensive net energy efficiency benefit objective function that includes product recovery benefits and equipment power consumption with a heat transfer coefficient attenuation penalty. By calculating the gradient of the total derivative of this objective function with respect to cooling water flow rate, a gradient-based numerical iterative optimization algorithm is used to continuously update the independent variables along the gradient's upward direction to solve the problem and output the optimal cooling water control flow rate command.

[0005] A further technical solution involves the flow field feature prediction module, where the prediction model for the Sauter mean diameter of the entrained droplets includes a static hydrodynamic baseline term and a dynamic pulsation correction term. The static hydrodynamic baseline term characterizes the initial aerodynamic breaking effect of the flow field's basic shear force on the droplets, and is positively correlated with the surface tension at the gas-liquid interface, and negatively correlated with the gas phase dynamic pressure and Reynolds number. The dynamic pulsation correction term characterizes the secondary aerodynamic breaking effect of pressure pulsations on the droplets, and is constructed as a decreasing function of the ratio of the root mean square of the high-frequency fluctuations of the instantaneous differential pressure signal to its arithmetic mean, used to reduce and correct the static hydrodynamic baseline term.

[0006] A further technical solution includes a soft-sensor derivation sub-model, a critical flooding gas velocity correction model, and an actual separation efficiency calculation model in the separation efficiency solution module. The specific configuration is as follows: The soft-sensor derivation sub-model is used to calculate the total liquid phase mass flow rate entering the gas-liquid separator. Its established gas-liquid entrainment mapping relationship is positively correlated with the total hot-side gas phase mass flow rate and the local gas phase dynamic pressure, and negatively correlated with the gas-liquid interface surface tension. The critical flooding gas velocity correction model is established based on the standard flooding constant of the separator internals and introduces a characterizing liquid film. The dynamic pulsation characteristic variable with secondary entrainment sensitivity is reduced and corrected. The dynamic pulsation characteristic variable is determined by the ratio of the root mean square of high-frequency fluctuations to the arithmetic mean. The actual separation efficiency calculation model is constructed as a truncated probability distribution function. When the apparent gas velocity inside the separator is lower than the corrected critical flooding velocity, the separation efficiency has an exponential asymptotic growth relationship with the droplet size ratio and velocity margin. When it is equal to or higher than the critical flooding velocity, the separation efficiency is zero. The total uncaptured liquid phase mass flow rate is the escape droplet mass flow rate to the condenser.

[0007] A further technical solution involves constructing an evolution model for the dynamic coking fouling thermal resistance of the condenser as an asymptotically saturated function over time within the heat transfer attenuation prediction module. This evolution model is derived from the dynamic equilibrium relationship between the fouling deposition rate term and the airflow shear stripping rate term. The fouling deposition rate term is positively coupled with the antimony-free catalyst activity temperature correction coefficient, the escaped droplet mass flow rate, and the oligomer mass fraction. The airflow shear stripping rate term is positively correlated with the square of the apparent gas velocity inside the separator and the gas phase density. The real-time total heat transfer coefficient is calculated based on the thermal resistance of the initial design total heat transfer coefficient and the dynamic coking fouling thermal resistance using the principle of series thermal resistance superposition.

[0008] A further technical solution involves the comprehensive energy efficiency optimization module, where the comprehensive net energy efficiency benefit objective function is constructed as a difference model between the economic benefit item of target product recovery and the total system operating cost item. The economic benefit item is linearly positively correlated with the actual gas-liquid separation efficiency. The total system operating cost item includes the circulating water pump power consumption item and the induced draft fan power consumption item, wherein the circulating water pump power consumption item is positively correlated with the cube of the cooling water control flow rate, and the induced draft fan power consumption item incorporates the attenuation degree of the real-time total heat transfer coefficient as a nonlinear back pressure penalty factor. In each control cycle, the comprehensive energy efficiency optimization module uses the cooling water control flow rate as the independent variable and employs a numerical iterative optimization algorithm to calculate the gradient of the total derivative of the objective function, continuously updating the independent variable along the gradient ascent direction until the objective function converges to the extreme value, thereby outputting the optimal cooling water control flow rate.

[0009] Further technical solutions also include a parameter adaptive calibration module, which is used to acquire the instantaneous differential pressure signal characteristics and actual separated and captured data under offline operating conditions, and obtain dimensionless correction coefficients, high-frequency pulsation interaction calibration coefficients and secondary entrainment sensitivity calibration coefficients by nonlinear regression fitting, and correct them online by gradient descent method according to the deviation of the actual heat transfer coefficient during device operation.

[0010] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0011] 1. This invention overcomes the shortcomings of traditional fixed-structure separators in dealing with nonlinear pressure pulsations. By extracting the high-frequency fluctuation characteristics of instantaneous differential pressure and incorporating them into the prediction of the Sauter mean diameter and the dynamic correction of the critical flooding velocity, this invention effectively quantifies the destructive effect of severe flow field fluctuations on liquid film stability, and achieves accurate solutions for gas-liquid separation efficiency and escape material quantity in antimony-free esterified complex boiling systems.

[0012] 2. An innovative model for the evolution of thermal resistance due to dynamic coking in condensers was constructed, and the decrease in heat transfer coefficient caused by coking was directly introduced into the total system operating cost as a nonlinear back pressure penalty factor. By constructing a comprehensive net energy efficiency objective function that includes both recovery benefits and penalty energy consumption and performing numerical iteration optimization, the system can adaptively output the optimal cooling water flow rate, completely solving the energy efficiency imbalance problem caused by dynamic heat transfer decay.

[0013] 3. To address the engineering challenge of empirical parameters becoming invalid due to variations in flow field characteristics and coking patterns under different operating conditions, this device introduces a dual-layer calibration strategy combining offline nonlinear regression calibration and online gradient descent correction. This allows the core fluid dynamics and heat transfer empirical coefficients to match the actual operating conditions in real time, avoiding model distortion and thus ensuring the device's prediction accuracy and optimal energy efficiency under long-term complex operating conditions. Attached Figure Description

[0014] Figure 1 This invention provides a schematic diagram of the connection between the physical device and the measurement and control unit.

[0015] Figure 2 This is a block diagram of the internal virtual function module architecture of the controller provided by the present invention;

[0016] Figure 3 The flowchart illustrates the execution of the energy efficiency optimization control algorithm provided by this invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0018] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0019] like Figure 1 , Figure 2 and Figure 3 As shown, an embodiment of the present invention provides a gas-liquid separation and condensation recovery device for an antimony-free polyester esterification process, comprising:

[0020] The flow field characteristic prediction module acquires the instantaneous differential pressure signal at the gas-liquid separator inlet and extracts its high-frequency fluctuation characteristics. Combined with fluid property parameters, it predicts the Sauter mean diameter of entrained droplets. The gas-liquid separator is the core component of the device, used to separate droplets from the gas phase flow to improve gas phase purity and recover liquid phase products. The instantaneous differential pressure signal is the pressure difference data measured in real time at the gas-liquid separator inlet; its changes reflect the dynamic behavior and fluctuations within the flow field. High-frequency fluctuation characteristics are the rapid, periodic, or non-periodic pressure change patterns contained in the instantaneous differential pressure signal. These fluctuations are usually related to phenomena such as fluid turbulence, interphase interactions, or equipment vibration. Fluid property parameters are quantities describing the physical properties of fluids, such as density, viscosity, and surface tension. These parameters have a significant impact on fluid flow and phase separation processes. Entrained droplets are tiny liquid particles carried in the gas phase flow, typically formed by mechanisms such as gas flow shearing, boiling, or breakup. The Sotter mean diameter is a statistical measure representing the average size of a droplet swarm. It is defined as the ratio of the total volume of all droplets in the swarm to their total surface area, and it reflects the mass and heat transfer characteristics of the droplets.

[0021] The separation efficiency calculation module calculates the total liquid flow rate entering the separator based on the ratio of gas kinetic pressure to surface tension, dynamically corrects the critical flooding velocity according to the high-frequency fluctuation characteristics, and calculates the actual gas-liquid separation efficiency and the mass flow rate of escaped droplets escaping to the condenser by combining the Sotter mean diameter. Gas kinetic pressure is the kinetic pressure generated by airflow motion; its magnitude is related to airflow velocity and density, and is an important parameter describing the force of airflow on droplets. Surface tension is the contractile force generated by intermolecular forces on the liquid surface, which affects droplet formation, stability, and breakup behavior. Total liquid flow rate is the total mass or volume of all liquid phase substances entering the gas-liquid separator within a specific unit of time. The critical flooding velocity is the phenomenon where, under specific operating conditions, the gas phase velocity reaches a certain critical value, at which point a large amount of liquid phase begins to be entrained by the gas phase, leading to a sharp decline in separation performance or even failure. The actual gas-liquid separation efficiency is the ratio of the mass of liquid phase successfully separated from the gas phase to the total mass of liquid phase entering the separator under actual operating conditions. The escape droplet mass flow rate is the total mass of droplets that fail to be captured by the gas-liquid separator and enter the subsequent condensation system with the gas phase flow within a specific time unit.

[0022] The heat transfer attenuation prediction module predicts the dynamic coking fouling thermal resistance of the condenser based on the escape droplet mass flow rate, characteristic activity parameters of the antimony-free catalyst, and intrinsic fouling properties, updating its real-time overall heat transfer coefficient. The characteristic activity parameters of the antimony-free catalyst are quantitative indicators describing its catalytic activity under specific reaction conditions; its high activity may lead to secondary reactions of oligomers on the condenser surface. Intrinsic fouling properties describe the physicochemical properties of the fouling layer itself, such as density, thermal conductivity, and adhesion; these properties affect the fouling formation rate and thermal resistance. The dynamic coking fouling thermal resistance of the condenser is the additional thermal resistance formed on the condenser heat exchange surface due to oligomer coking; its value changes dynamically over time, reflecting the degree to which coking hinders heat transfer efficiency. The real-time overall heat transfer coefficient is the actual heat transfer performance index of the condenser at a certain moment, comprehensively considering the heat transfer resistance of the heat exchange wall, fluid boundary layer, and fouling layer.

[0023] The comprehensive energy efficiency optimization module constructs a comprehensive net energy efficiency benefit objective function that encompasses product recovery benefits and equipment power consumption including a heat transfer coefficient decay penalty. By calculating the gradient of the total derivative of this objective function with respect to cooling water flow rate, a gradient-based numerical iterative optimization algorithm is used to continuously update the independent variables along the gradient's ascending direction to solve the problem and output the optimal cooling water control flow rate command. Product recovery benefits are the economic benefits derived from recovering valuable liquid products (such as ethylene glycol) through the gas-liquid separation and condensation processes. The heat transfer coefficient decay penalty term in the energy efficiency objective function quantifies the additional energy consumption or economic loss caused by the decrease in condenser heat transfer efficiency (due to coking). Equipment power consumption refers to the electrical energy consumed by power equipment such as fans and pumps during the operation of the gas-liquid separation and condensation recovery unit. The comprehensive net energy efficiency benefit objective function is a mathematical expression used to quantify the overall economic benefits of the unit during operation, comprehensively considering product recovery benefits and various operating costs (including additional energy consumption caused by heat transfer decay). Cooling water flow rate is the mass or volume of cooling water flowing through the condenser to cool the gas phase per unit time. The total derivative gradient is the rate of change of the objective function with respect to the cooling water flow rate, and its direction indicates the direction in which the objective function value increases the most rapidly. Numerical iterative algorithms are methods that gradually find a solution to a problem through repeated calculations and approximations; here, they are used to solve for the optimal cooling water flow rate. The optimal cooling water control flow rate command is calculated by the optimization algorithm and represents the cooling water flow rate setpoint that maximizes the overall net energy efficiency benefit objective function.

[0024] In response, this application proposes a gas-liquid separation and condensation recovery device for antimony-free polyester esterification processes. The device includes a flow field characteristic prediction module, a separation efficiency solution module, a heat transfer attenuation prediction module, and a comprehensive energy efficiency optimization module.

[0025] The flow field characteristic prediction module acquires the instantaneous differential pressure signal at the gas-liquid separator inlet and extracts its high-frequency fluctuation characteristics. Combined with fluid property parameters, it predicts the Sauter mean diameter of entrained droplets. Specifically, the instantaneous differential pressure signal can be acquired in real time by a differential pressure sensor installed on the gas-liquid separator inlet pipe. High-frequency fluctuation characteristics can be extracted from the raw differential pressure signal using signal processing techniques, such as Fourier transform or wavelet analysis. Fluid property parameters, such as gas phase density, viscosity, and surface tension, can be measured by online sensors or obtained by looking up tables based on process conditions. The prediction of the Sauter mean diameter of entrained droplets can be estimated based on existing empirical models or simplified physical models. For example, it can be estimated by simply correlating gas flow pressure with surface tension, or by assuming that the droplet size is inversely proportional to the square of the gas flow velocity.

[0026] The separation efficiency calculation module calculates the total liquid flow rate entering the separator based on the ratio of gas flow pressure to surface tension, dynamically corrects the critical flooding velocity according to the aforementioned high-frequency fluctuation characteristics, and calculates the actual gas-liquid separation efficiency and the escape droplet mass flow rate to the condenser by combining the Sauter mean diameter. Specifically, the total liquid flow rate entering the separator can be initially calculated by multiplying the total gas mass flow rate by an empirical entrainment coefficient. The correction of the critical flooding velocity can be based on an empirical formula, adjusting the traditional fixed critical flooding velocity by introducing a correction factor related to the high-frequency fluctuation characteristics. The calculation of the actual gas-liquid separation efficiency can adopt a standard separation efficiency model based on droplet size distribution and separator geometry. For example, it can assume that all droplets larger than a certain critical value are captured, while droplets smaller than that critical value partially or completely escape. Thus, the escape droplet mass flow rate can be obtained by multiplying the total liquid flow rate by (1 minus the actual separation efficiency).

[0027] The heat transfer attenuation prediction module is used to predict the dynamic coking fouling thermal resistance of the condenser based on the escape droplet mass flow rate, characteristic activity parameters of the antimony-free catalyst, and intrinsic fouling properties, and to update its real-time overall heat transfer coefficient. Specifically, the prediction of dynamic coking fouling thermal resistance can be based on a simplified fouling formation model that treats fouling thermal resistance as a linear or nonlinear function of escape droplet mass flow rate, catalyst activity, and fouling properties, taking into account the cumulative effect of fouling over time. For example, it can be assumed that the fouling thermal resistance is positively correlated with escape droplet mass flow rate and catalyst activity, and increases monotonically with operating time. The real-time update of the overall heat transfer coefficient can be obtained by adding the reciprocal of the initial design overall heat transfer coefficient to the predicted fouling thermal resistance and then taking the reciprocal.

[0028] The comprehensive energy efficiency optimization module is used to construct a comprehensive net energy efficiency benefit objective function that includes product recovery benefits and equipment power consumption including heat transfer coefficient decay penalty. By calculating the total derivative gradient of the objective function with respect to cooling water flow rate, a gradient-based numerical iterative optimization algorithm is used to continuously update the independent variables along the gradient ascent direction to solve the problem and output the optimal cooling water control flow rate command.

[0029] The following example will provide a more detailed explanation of the above technical solution:

[0030] Assume a gas-liquid separation and condensation recovery unit is operating at a production site for an antimony-free polyester esterification process. During the esterification reaction, a large amount of gas-liquid two-phase flow carrying oligomer droplets is generated at the reactor outlet.

[0031] First, the flow field feature prediction module continuously acquires the instantaneous differential pressure signal at the inlet of the gas-liquid separator. For example, hundreds of differential pressure data points are collected per second using a high-precision differential pressure sensor installed on the inlet pipe. These raw signals are sent to the signal processing unit to extract their high-frequency fluctuation characteristics, such as calculating the ratio of the root mean square value to the arithmetic mean of the instantaneous differential pressure signal sequence to quantify the turbulence intensity and pulsation degree of the flow field. Simultaneously, the system acquires fluid properties such as gas density, viscosity, and surface tension under the current operating conditions. Combining these high-frequency fluctuation characteristics and fluid properties, the flow field feature prediction module predicts the Sauter mean diameter of the entrained droplets. For example, when an increase in high-frequency fluctuations is detected, the predicted Sauter mean diameter of the droplets will decrease accordingly, reflecting the phenomenon of secondary breakup of the droplets under intense turbulence.

[0032] Next, the separation efficiency calculation module receives the predicted Sauter mean diameter. This module first calculates the total liquid mass flow rate entering the gas-liquid separator based on the ratio of gas flow pressure to surface tension, combined with the total gas mass flow rate. Subsequently, this module dynamically corrects the critical flooding velocity using the aforementioned high-frequency fluctuation characteristics. For example, when the high-frequency fluctuation characteristics indicate severe flow field pulsation, the critical flooding velocity is corrected downwards to reflect the actual situation where flooding is more likely to occur. Based on the corrected critical flooding velocity and the predicted Sauter mean diameter, the separation efficiency calculation module calculates the actual gas-liquid separation efficiency. This allows for accurate quantification of the escape droplet mass flow rate that fails to be captured by the separator and escapes to the condenser. For example, if the actual separation efficiency is 90% and the total liquid flow rate is 100 kg / h, then the escape droplet mass flow rate is 10 kg / h.

[0033] Subsequently, the heat transfer decay prediction module receives the escape droplet mass flow rate. This module combines characteristic activity parameters of the antimony-free catalyst (e.g., the rate constant of oligomer polycondensation at a specific temperature) and intrinsic fouling properties (e.g., the density and thermal conductivity of the oligomer fouling) to predict the thermal resistance of dynamically coking fouling on the condenser heat exchange surface. For example, the predicted fouling thermal resistance increases more rapidly over time as the escape droplet mass flow rate increases. Based on the predicted fouling thermal resistance, the module updates the real-time overall heat transfer coefficient of the condenser. For example, if the initial design overall heat transfer coefficient is... The predicted fouling thermal resistance is The real-time total heat transfer coefficient It will be updated to .

[0034] Finally, the comprehensive energy efficiency optimization module utilizes the aforementioned real-time total heat transfer coefficient to construct a global comprehensive net energy efficiency benefit objective function. This objective function comprehensively considers the benefits from recovering products such as ethylene glycol, as well as the power consumption of equipment such as cooling water pumps and induced draft fans. Specifically, the induced draft fan power consumption component incorporates a penalty term related to the decay of the real-time total heat transfer coefficient, quantifying the additional energy consumption increased due to the decrease in heat transfer efficiency caused by coking. For example, when the real-time total heat transfer coefficient... As the flow rate decreases, the penalty term increases, reflecting the higher induced draft fan power required to maintain the same condensation effect. This module calculates the gradient of the total derivative of the objective function with respect to the cooling water flow rate and employs a numerical iterative algorithm—for example, starting from an initial cooling water flow rate and gradually adjusting along the gradient direction—until the objective function reaches its maximum value, thereby solving for and outputting the optimal cooling water control flow command.

[0035] Through the coordinated operation of the above modules, the device can sense the dynamics of the flow field in real time, predict the separation effect and coking trend, and adaptively optimize the cooling water flow rate based on this, thereby ensuring product quality while achieving economical operation of the device.

[0036] This application further proposes that, in the flow field characteristic prediction module, the prediction model for the Sauter mean diameter of entrained droplets includes a static hydrodynamic baseline term and a dynamic pulsation correction term. The static hydrodynamic baseline term characterizes the initial breakup effect of the basic shear force of the flow field on the droplets, and is positively correlated with the surface tension of the gas-liquid interface, and negatively correlated with the gas phase dynamic pressure and Reynolds number. The dynamic pulsation correction term characterizes the secondary aerodynamic breakup effect of pressure pulsation on the droplets, and is constructed as a decreasing function of the ratio of the root mean square of the high-frequency fluctuations of the instantaneous differential pressure signal to its arithmetic mean, used to reduce and correct the static hydrodynamic baseline term. The prediction model formula for the Sauter mean diameter of entrained droplets is as follows:

[0037]

[0038] in: This represents the Sauter mean diameter of the entrained droplets; Indicates the dimensionless correction coefficient; Indicates surface tension; Indicates the density of the gas phase; Indicates the apparent velocity in the gas phase; Indicates the hydraulic diameter of the separator inlet; Indicates the dynamic viscosity of the gas phase; Indicates the calibration coefficient for high-frequency pulsation interaction; The root mean square of the high-frequency fluctuations in the instantaneous differential pressure signal sequence; This represents the arithmetic mean of the instantaneous differential pressure signal sequence.

[0039] The core of this predictive model lies in accurately characterizing the Sauter mean diameter of the entrained droplets. , defined as the average ratio of droplet volume to surface area, is a key parameter characterizing the size distribution of droplet swarms and directly affects gas-liquid separation efficiency. Dimensionless correction coefficient. As an empirical parameter in the model, it is used to calibrate the prediction results overall to make them more consistent with actual working conditions. Its value can be determined through experimental data fitting or computational fluid dynamics (CFD) simulation. Surface tension This is an inherent physical property of the liquid phase, reflecting the droplet's ability to resist deformation and breakage. It can be obtained through online measurement or calculation based on a physical property database. (Gas phase density) and apparent velocity of the gas phase These parameters collectively determine the kinetic energy of the gas phase and significantly influence the force state and breakup behavior of the droplets. These parameters can be acquired in real-time by process sensors or calculated through material balance. (Separator inlet hydraulic diameter) This is a characteristic length that characterizes the geometry of the separator, affecting the flow characteristics of the airflow within the separator, and is typically a design parameter for the equipment. Gas phase dynamic viscosity This reflects the resistance characteristics of gas-phase flow and affects the effect of gas flow shear force on droplets. It can be queried from a physical property database or calculated based on temperature and pressure. High-frequency pulsating interaction calibration coefficient. This is an empirical coefficient used to quantify the impact of high-frequency flow field pulsations on droplet breakup or coalescence. Its value can be obtained through offline experiments or regression analysis of historical operating data. The root mean square of high-frequency fluctuations in the instantaneous differential pressure signal sequence... It is a statistical measure of the intensity of high-frequency component fluctuations in an instantaneous differential pressure signal, and can be calculated by filtering the original differential pressure signal. It is the arithmetic mean of the instantaneous differential pressure signal sequence. This represents the average level of the differential pressure signal. It is used as a benchmark value and combined with the root mean square of high-frequency fluctuations to form a dimensionless fluctuation intensity index.

[0040] This scheme constructs a Sauter mean diameter prediction model that incorporates the high-frequency fluctuation characteristics of the actual flow field. It quantifies the dynamically generated flow field information and introduces it into the droplet size prediction process, obtaining droplet diameter prediction results that closely match the actual conditions of antimony-free polyester esterification. This provides accurate basic parameters for subsequent process calculations throughout the entire process. The prediction model first uses surface tension... With gas phase dynamic pressure ( The ratio of ) is used as the basic calculation term, combined with the gas phase Reynolds number characteristics ( The correction term, based on conventional fluid properties and separator structural parameters, yields a basic droplet diameter prediction that conforms to the fundamental laws of two-phase flow. Among these, the gas phase density... Apparent velocity in the gas phase Separator inlet hydraulic diameter Gas phase dynamic viscosity These parameters respectively reflect the fundamental influence of gas phase flow state, separator inlet structure on droplet size, and surface tension. This reflects the influence of the liquid phase's inherent properties on droplet aggregation ability, consistent with the fundamental physical laws governing droplet size variation in two-phase flow. Building upon this, a correction term incorporating high-frequency fluctuation characteristics is introduced, employing the root mean square of high-frequency fluctuations from the instantaneous differential pressure signal sequence. Arithmetic mean of instantaneous differential pressure signal sequence The dimensionless ratio characterizes the relative severity of flow field fluctuations, avoiding interference from differences in the absolute value of differential pressure under different operating conditions, and can accurately reflect the intensity of flow field pulsations during actual operation. Combined with the calibration coefficients for high-frequency pulsation interaction obtained from calibration... This allows for accurate quantification of the impact of flow field fluctuations on droplet diameter, consistent with the actual pattern that the more intense the flow field fluctuations, the more likely the droplets are to undergo secondary breakup and the smaller their average diameter. Working in conjunction with the flow field characteristic prediction module, this model acquires the instantaneous differential pressure signal at the gas-liquid separator inlet and extracts its high-frequency fluctuation characteristics. Combined with fluid property parameters, it can predict the Sauter mean diameter of entrained droplets in real time and dynamically. This provides more accurate input for subsequent separation efficiency calculations, heat transfer attenuation predictions, and comprehensive energy efficiency optimization, ensuring that the entire device can achieve more precise operation control and optimization under the complex and variable conditions of the antimony-free polyester esterification process.

[0041] As one specific implementation, this predictive model can be implemented in the device's control system. First, a high-precision differential pressure sensor is installed at the inlet of the gas-liquid separator to acquire instantaneous differential pressure signals in real time. These signals can be transmitted to an embedded controller or industrial PC at a high sampling frequency (e.g., hundreds or even thousands of times per second). Within this controller, a digital signal processing algorithm can be run to perform high-pass filtering on the acquired instantaneous differential pressure signal sequence to extract high-frequency fluctuation components and calculate their root mean square (RMS). Simultaneously, the original signal is either low-pass filtered or its arithmetic mean is directly calculated. Gas phase density Apparent velocity in the gas phase Gas phase dynamic viscosity Fluid properties such as surface tension can be obtained through online sensors (e.g., flow meters, thermometers, pressure gauges) and calculated in real time using a property database. The inlet hydraulic diameter of the separator can be obtained using empirical formulas or databases based on the liquid phase composition and temperature. Dimensionless correction coefficients are pre-input into the system as fixed design parameters. Calibration coefficients for interaction with high-frequency pulsations The system can be calibrated through offline experiments or computational fluid dynamics simulations, combined with actual operating data, and stored in the controller. All these parameters are input into a preset predictive model formula, which is then calculated in real time by the controller. For example, the apparent velocity of the gas phase is measured. = 2 m / s, gas phase density = 1.0 kg / m 3 gas dynamic viscosity = Pa·s, hydraulic diameter of separator = 1.0 m, surface tension = 0.04 N / m. Arithmetic mean of differential pressure signal. = 5000 Pa, root mean square of high-frequency fluctuations = 500 Pa. Calibration constant. = 500, = 2.0. Calculate the effect of gas phase Reynolds number: Correction item Calculate the pulsation correction term: ratio ,but .calculate : This outputs the Sauter mean diameter of the entrained droplets under the current operating conditions. The calculation result can then be used as an input parameter to the separation efficiency solution module for subsequent separation efficiency calculations and system optimization.

[0042] The above technical solution solves the problem of insufficient accuracy caused by the failure to consider dynamic flow field fluctuations in the original prediction. It provides accurate input parameters for subsequent separation efficiency calculation, coking prediction, and energy efficiency optimization, ensuring the accuracy of the entire unit's operational calculations and adapting to the dynamic operational requirements of the antimony-free polyester esterification process. This model accurately reflects the impact of flow field pulsations on the droplet breakup and aggregation state, making the predicted Sauter mean diameter closer to actual operating conditions. This significantly improves the prediction accuracy and control effect of the entire gas-liquid separation and condensation recovery unit in the antimony-free polyester esterification process.

[0043] This application further proposes that the separation efficiency solution module includes a soft-sensor derivation sub-model, a critical flooding gas velocity correction model, and an actual separation efficiency calculation model, specifically configured as follows: The soft-sensor derivation sub-model is used to calculate the total liquid phase mass flow rate entering the gas-liquid separator. Its established gas-liquid entrainment mapping relationship is positively correlated with the total hot-side gas phase mass flow rate and the local gas phase dynamic pressure, and negatively correlated with the gas-liquid interface surface tension. The critical flooding gas velocity correction model is established based on the standard flooding constant of the separator internals, and a derating correction is performed by introducing a dynamic pulsation characteristic variable characterizing the secondary entrainment sensitivity of the liquid film. The dynamic pulsation characteristic variable is modified from high... The ratio of the root mean square of the frequency fluctuation to the arithmetic mean determines the efficiency. The actual separation efficiency calculation model is constructed as a truncated probability distribution function. When the apparent gas velocity inside the separator is lower than the corrected critical flooding velocity, the separation efficiency exhibits an exponential asymptotic growth relationship with the droplet size ratio (the ratio of the Sauter mean diameter to the critical cut-off diameter) and the velocity margin (the difference between the actual apparent velocity and the critical flooding velocity). When the velocity is equal to or higher than the critical flooding velocity, the separation efficiency reaches zero. The total uncaptured liquid phase mass flow rate is the escape droplet mass flow rate to the condenser. The calculation formulas for the soft-sensor derivation sub-model, the actual gas-liquid separation efficiency, and the escape droplet mass flow rate are as follows:

[0044]

[0045]

[0046]

[0047]

[0048] in: This indicates the total liquid mass flow rate entering the gas-liquid separator; This indicates the total mass flow rate of the hot-side gas phase; This represents a dimensionless entrainment of empirical constants; Indicates the density of the gas phase; Indicates the apparent velocity in the gas phase; Indicates the hydraulic diameter of the separator inlet; Indicates surface tension; Indicates the entrainment index; This represents the critical flooding velocity after considering fluctuation correction; Indicates the standard solution's flooding constant; This represents the secondary entrainment sensitivity calibration coefficient; The root mean square of the high-frequency fluctuations in the instantaneous differential pressure signal sequence; This represents the arithmetic mean of the instantaneous differential pressure signal sequence; Indicates the density of the liquid phase; This indicates the actual gas-liquid separation efficiency; Indicates the maximum design separation efficiency; This represents the Sauter mean diameter of the entrained droplets; Indicates the critical cutting particle size of the internal component; This indicates the apparent gas velocity inside the separator; This represents the mass flow rate of the escaping droplet.

[0049] In the above scheme, the soft-measurement extrapolation sub-model for the total liquid mass flow rate entering the gas-liquid separator is used to estimate the total liquid mass flow rate entering the gas-liquid separator using other easily obtainable process parameters without the need for direct measurement by physical sensors. This model can be integrated into the device's control system, acquiring parameters such as the total gas mass flow rate, gas density, apparent gas velocity, separator inlet hydraulic diameter, and surface tension in real time, and performing calculations using preset dimensionless entrainment empirical constants and entrainment exponents. Alternatively, this model can run as an independent software module in a host computer or distributed control system (DCS), acquiring the required parameters and performing calculations through data communication with field instruments. The calculation formula for the critical flooding gas velocity after considering fluctuation correction is used to calculate the critical flooding gas velocity of the gas-liquid separator considering the influence of high-frequency flow field fluctuations, i.e., the maximum gas velocity at which the gas-liquid separator can operate stably without flooding. This calculation can be executed by the processor within the separation efficiency solution module, receiving the high-frequency fluctuation root mean square and arithmetic mean of the instantaneous differential pressure signal sequence in real time, and dynamically calculating based on preset standard flooding constants, secondary entrainment sensitivity calibration coefficients, liquid phase density, and gas phase density. Alternatively, this formula can be encoded as an executable program and deployed in a field-programmable gate array (FPGA) or application-specific integrated circuit (ASIC) to achieve high-speed, real-time correction of the critical flooding gas velocity. The formula for calculating the actual gas-liquid separation efficiency is used to dynamically evaluate the actual separation effect of the gas-liquid separator on droplets based on the Sauter mean diameter of the entrained droplets, the apparent gas velocity inside the separator, and the corrected critical flooding gas velocity. This calculation logic can be embedded in the algorithm of the separation efficiency solution module, comparing the current apparent gas velocity inside the separator with the calculated critical flooding gas velocity, and selecting the appropriate formula for calculation based on different conditions. Alternatively, this separation efficiency calculation can be performed as a subroutine, periodically called in the main control program to ensure real-time updates of the separation efficiency, and the calculation speed can be optimized using lookup tables or interpolation methods. The formula for calculating the escape droplet mass flow rate is used to calculate the total mass flow rate of droplets that fail to be captured by the gas-liquid separator and thus escape into the subsequent condensation system. This calculation is directly based on the multiplication of the total liquid phase mass flow rate entering the gas-liquid separator and the actual gas-liquid separation efficiency, and can be executed immediately after the separation efficiency calculation in the separation efficiency solution module's calculation flow. Alternatively, the result of this escape droplet mass flow rate calculation can be used as a key output, passed through a data interface to the heat transfer attenuation prediction module as an input parameter for predicting the dynamic coking and fouling thermal resistance of the condenser. Standard liquid flooding constant. Maximum design separation efficiency and the critical cutting particle size of internal features These are the mechanical structural parameters or design limit performance parameters that come with the separator and condenser at the factory, provided by the equipment manufacturer.

[0050] This application's solution introduces a soft-measurement derivation sub-model of total liquid phase mass flow rate, combining the total mass flow rate of the hot-side gas phase, the ratio of gas flow pressure to surface tension, and dimensionless empirical parameters to calculate the total liquid phase flow rate. This eliminates the need for additional hardware measurement devices and closely aligns with the entrainment characteristics of two-phase flow, providing an accurate foundation for subsequent calculations. Based on this, the solution utilizes the high-frequency fluctuation characteristics of instantaneous differential pressure extracted by the flow field feature prediction module—specifically, the ratio of the root mean square of the high-frequency fluctuation to the average differential pressure—to dynamically correct the critical flooding velocity. A secondary entrainment sensitivity calibration coefficient is introduced to accommodate the impact of fluctuations on flooding under different operating conditions. This solves the problem that traditional fixed critical flooding velocity cannot reflect the dynamic damage of flow field fluctuations to liquid film stability and flooding critical conditions, enabling the acquisition of a critical flooding velocity that conforms to actual fluctuation conditions. Furthermore, by combining the Sauter mean diameter of entrained droplets obtained from the flow field feature prediction module and the corrected critical flooding velocity, the actual gas-liquid separation efficiency is calculated based on the relationship between the actual flow velocity and the critical flooding velocity. When the apparent gas velocity inside the separator is lower than the critical flooding velocity, the separation efficiency is calculated by combining the maximum design separation efficiency, the ratio of the Sauter mean diameter to the critical cut-off particle size of the internal components, and the difference between the flow velocity and the critical flooding velocity. When the flow velocity reaches or exceeds the critical flooding velocity, the separation efficiency is considered to be zero. This calculation method conforms to the actual laws of the gas-liquid separation process, accurately reflects the influence of droplet size and flooding state on the separation effect, and obtains a separation efficiency that closely matches reality. Finally, the escape droplet mass flow rate is calculated from the total liquid phase mass flow rate and the actual gas-liquid separation efficiency, which can accurately quantify the mass of droplets penetrating the separator and entering the condenser. This provides accurate input for the subsequent heat transfer attenuation prediction module to predict the dynamic coking fouling thermal resistance of the condenser, thereby ensuring that the subsequent calculations and optimizations of the entire system conform to actual operating conditions.

[0051] As a specific implementation, the aforementioned separation efficiency solution module can be implemented as an embedded controller, such as a high-performance industrial-grade programmable logic controller (PLC) or a field control unit of a distributed control system (DCS). This controller interacts with the flow field characteristic prediction module via industrial Ethernet or fieldbus (such as Modbus TCP / IP) to acquire the high-frequency root mean square of the instantaneous differential pressure signal sequence in real time. and arithmetic mean And the Sotter average diameter of the entrained droplets Simultaneously, the controller is also connected to a gas phase flow meter, a pressure sensor, and a temperature sensor to obtain the total mass flow rate of the hot-side gas phase. gas phase density Apparent velocity in the gas phase Separator inlet hydraulic diameter Surface tension Liquid phase density and apparent gas flow rate inside the separator Parameters such as these are pre-stored within the controller, along with dimensionless entrainment empirical constants. Entrainment Index Standard solution constant Secondary entrainment sensitivity calibration coefficient Maximum design separation efficiency and the critical cutting particle size of internal features Parameters. The dimensionless entrainment empirical constant. With entrainment index The specific values ​​are determined in advance by the following method: before the device is put into industrial operation, a scaled-down cold experimental platform for esterification reaction gas-liquid entrainment is built to collect multiple sets of different total gas phase mass flow rates. Local dynamic pressure and surface tension Measured data on actual liquid phase entrainment under specific conditions; subsequently, a system was established based on... As the dependent variable, with dimensionless terms For a log-linear regression model with independent variables, the constants that satisfy the specific internal geometry of the separator can be determined by fitting the model using the least squares method. With index And pre-stored. The controller periodically performs computational tasks, first calculating using the soft-sensor inference sub-model. Then, it is dynamically corrected based on the characteristics of high-frequency fluctuations. Then calculate Finally, it was concluded that These calculation results are sent in real time to the heat transfer attenuation prediction module via a data interface as its input. For example, = 10 kg / s, constant = 0.01, entrainment index = 0.5. Liquid phase density = 1000 kg / m 3 Flow rate inside the separator = 1.5 m / s, internal parameters = 0.15 m / s, = 1.5, = 0.98, = 150 m. Calculate Weberian terms ,but .calculate : ;because < (1.5 < 4.029), calculate the actual separation efficiency. : ; Right now ; Calculate the escape quantity: .

[0052] Through the above technical solution, this application can obtain accurate gas-liquid separation results that closely match actual operating conditions, providing a reliable input basis for subsequent condenser coking prediction and comprehensive energy efficiency optimization. This solution, by introducing a soft-measurement derivation sub-model of total liquid phase mass flow rate, avoids the difficulty of directly measuring liquid phase flow rate in complex two-phase flows, improving the convenience and economy of data acquisition. Simultaneously, by dynamically correcting the critical flooding velocity, the device can respond in real time to the influence of high-frequency flow field fluctuations on the flooding critical condition, effectively avoiding the flooding risk caused by traditional fixed critical values ​​and ensuring the stable operation of the separator. Calculating the actual separation efficiency by combining the Sauter mean diameter and the corrected critical flooding velocity can more accurately reflect the combined influence of droplet size distribution and flow field dynamics on the separation effect, improving the accuracy of separation efficiency assessment. Finally, the accurately calculated escape droplet mass flow rate provides more reliable and accurate input data for subsequent condenser coking prediction and comprehensive energy efficiency optimization, thereby improving the prediction accuracy and optimization effect of the entire device and enabling it to better adapt to the fluctuating operating conditions of the antimony-free polyester esterification process.

[0053] This application further proposes that, in the heat transfer attenuation prediction module, the evolution model of the dynamic coking fouling thermal resistance of the condenser is constructed as an asymptotically saturated function evolving over time; the evolution model is derived from the dynamic equilibrium relationship between the fouling deposition rate term and the airflow shear stripping rate term; wherein, the fouling deposition rate term is positively coupled with the antimony-free catalyst activity temperature correction coefficient, the escaped droplet mass flow rate, and the oligomer mass fraction; the airflow shear stripping rate term is positively correlated with the square of the apparent gas velocity inside the separator and the gas phase density; the real-time total heat transfer coefficient is calculated based on the thermal resistance of the initial design total heat transfer coefficient and the dynamic coking fouling thermal resistance through the principle of series thermal resistance superposition; the evolution model formulas for the dynamic coking fouling thermal resistance of the condenser and the real-time total heat transfer coefficient of the condenser are as follows:

[0054]

[0055]

[0056] in: Indicates the dynamic thermal resistance of coking fouling; This represents the temperature correction factor for the activity of antimony-free catalysts. This indicates the probability of oligomer adhesion to the wall; Indicates the mass flow rate of the escaping droplet; Indicates the mass fraction of oligomers in the droplet; Indicates the airflow shear stripping and scrubbing coefficient; Indicates the density of the gas phase; This indicates the apparent gas velocity inside the separator; Indicates the effective heat exchange surface area; Indicates the density of oligomerized dirt; Indicates the thermal conductivity of oligomeric fouling; Indicates the system's cumulative uptime; Indicates the real-time overall heat transfer coefficient; This represents the overall heat transfer coefficient in the initial design.

[0057] Specifically, dynamic coking fouling thermal resistance This refers to the additional thermal resistance generated on the heat exchange surface of the condenser due to oligomer coking. Its value changes dynamically with system operating time, directly reflecting the impact of the fouling layer on heat transfer efficiency. This thermal resistance can be obtained by back-calculating heat transfer data through online monitoring of parameters such as the inlet and outlet temperatures and flow rates of the heat exchanger, or by predicting it through a physicochemical model. (Antimony-free catalyst activity temperature correction coefficient) This coefficient is used to quantify the effect of antimony-free catalysts on the coking reactivity of oligomers at different temperatures, ensuring that the model accurately reflects the coking tendency of antimony-free catalysts at actual operating temperatures. This coefficient can be obtained by fitting laboratory small-scale or pilot-scale data with kinetic models such as the Arrhenius equation, or by using empirical formulas or looking up tables. Oligomer adhesion probability to the wall. This represents the probability that escaped oligomer droplets will contact the condenser wall and successfully adhere to form fouling, reflecting the initiation conditions and efficiency of oligomer coking. In practical engineering applications, to avoid complex surface chemical microscopic calculations and reduce the real-time control load on the system, this probability... The preferred method is to obtain the offline empirical calibration constant; in specific implementation, for a specific condenser heat exchanger tube material (such as 316L stainless steel) and a specific antimony-free catalyst system, the dynamic fouling deposition rate within a specific temperature range is measured through preliminary plate-mounted experiments for back-calculation, and the result is pre-stored in the control system in the form of a lookup table. Escape droplet mass flow rate This refers to the total mass flow rate of droplets escaping from the gas-liquid separator and entering the condenser. It is the source of coking material and directly affects the coking rate. This flow rate can be calculated by the upstream separation efficiency solution module or by combining online droplet size distribution measurement with gas phase flow rate. The mass fraction of oligomers in the droplets... The content of oligomer components with a tendency to coke in the escaped droplets quantifies the "coking potential" of the escaped droplets. For offline acquisition of this mass fraction, droplets can be collected periodically by installing an isokinetic sampling probe in the condenser inlet pipe section, followed by determination using gel permeation chromatography (GPC) or high-performance liquid chromatography (HPLC). For real-time online acquisition, a soft measurement model is preferred, which extracts easily measurable macroscopic process variables such as the real-time temperature of the esterification reactor, system vacuum, and reaction liquid level (residence time), and dynamically estimates these variables in real time based on a pre-trained partial least squares regression (PLS) or artificial neural network (ANN) algorithm. The value is determined to meet the real-time requirements of the control system while ensuring accuracy. Airflow shear stripping and scouring coefficient. This coefficient characterizes the shear force exerted by the airflow on the formed fouling layer. This shear force facilitates the removal of fouling and reflects the removal mechanism in the dynamic equilibrium of the fouling layer. This coefficient can be calculated using fluid dynamics simulations of wall shear stress and calibrated using experimental data, or determined using empirical formulas or semi-empirical models. Gas phase density. This is the mass-to-volume ratio of the gas phase within the condenser, affecting the momentum and shear force of the airflow, and consequently, the formation and removal of fouling. This density can be calculated by measuring temperature and pressure using online sensors and applying the equation of state, or obtained by looking up a table. The apparent gas velocity inside the separator... This refers to the average velocity of the gas passing through the internal cross-section of the separator, which affects the shearing and stripping effect of the gas flow on the fouling layer. This velocity can be measured by a flow meter to measure the gas phase flow rate and calculated based on the separator's cross-sectional area, or it can be obtained through fluid dynamics simulation. Effective heat transfer surface area. This refers to the actual surface area involved in heat exchange within the condenser. It is a fundamental geometric parameter for heat transfer calculations, typically determined based on condenser design drawings or specifications, or obtained through geometric measurements. (Oligomer fouling density) This is the mass-to-volume ratio of the formed fouling layer, affecting its physical properties and thermal resistance calculations. This density can be obtained through laboratory analysis of fouling samples, or selected based on empirical values ​​or literature data. The thermal conductivity of oligomeric fouling... This coefficient represents the ability of the dirt layer to conduct heat, directly determining its thermal resistance. This coefficient can be estimated by experimentally measuring the thermal conductivity of a dirt sample or based on the composition and structure of the dirt. System cumulative operating time. This is the total time elapsed since the device started operating until the current moment. It is the time variable in the dynamic coking fouling thermal resistance model, reflecting the fouling accumulation process, and is obtained through a system timer or operating records. Real-time total heat transfer coefficient. This coefficient represents the actual heat transfer capacity of the condenser at the current moment and is a key parameter for evaluating condenser performance and optimizing energy efficiency. It can be derived in real-time by monitoring parameters such as the inlet and outlet fluid temperatures and flow rates of the condenser, combined with heat transfer calculations, or calculated using this model. (Initial design overall heat transfer coefficient) It is the design heat transfer coefficient of the condenser in a clean and dirt-free state. It serves as a benchmark value to measure the impact of dirt on heat transfer performance. It is usually determined based on the condenser's design parameters and heat transfer calculations, or calibrated using measured data during the initial operation of new equipment.

[0058] The solution in this application introduces dynamic coking fouling thermal resistance. The evolutionary model can accurately quantify the dynamic process of coking of escaped oligomers on the heat exchange surface of the condenser in antimony-free polyester processes. This model comprehensively considers the coking formation mechanism (such as the antimony-free catalyst activity temperature correction coefficient). Oligomer adhesion probability escape droplet mass flow rate Mass fraction of oligomers in droplets ) and dirt removal mechanisms (such as airflow shear stripping scrubbing coefficient) gas phase density Apparent gas velocity inside the separator (and included the system's cumulative runtime) This key time variable accurately reflects the dynamic equilibrium process of coking fouling as it grows over time and is simultaneously sheared off by airflow. Based on this, the real-time overall heat transfer coefficient... The calculation formula is based on the initial design overall heat transfer coefficient. and dynamic coking fouling thermal resistance This enables real-time updates to the actual heat transfer capacity of the condenser. This dynamic and accurate prediction mechanism overcomes the problem that traditional static fouling thermal resistance calculations do not match actual dynamic changes, providing accurate and reliable real-time heat transfer performance data for subsequent integrated energy efficiency optimization modules. This allows the entire device to more effectively address the heat transfer attenuation problem caused by coking in the antimony-free polyester process, ensuring the stability of system operation and the accuracy of optimized control.

[0059] As a specific implementation method, during device operation, the heat transfer attenuation prediction module can continuously receive the escape droplet mass flow rate calculated by the upstream separation efficiency solution module. (For example Meanwhile, the antimony-free catalyst activity temperature correction factor. Based on the current reaction temperature and catalyst type, this can be obtained by consulting a preset temperature correction curve or empirical formula (e.g., = 1.2). Oligomer adhesion probability to the wall. and the mass fraction of oligomers in the droplets Estimates can be made based on process conditions and material properties using a pre-established database or soft sensor model (e.g.) = 0.05, = 0.1). Airflow shear stripping and scrubbing coefficient gas phase density Apparent gas velocity inside the separator Effective heat exchange surface area Oligopolymer dirt density and the thermal conductivity of oligomer fouling These parameters can be set or obtained in real time based on the device's design parameters, online sensor measurements, or empirical values ​​(e.g., the washing coefficient). = 0.002, gas phase density =1.0 kg / m 3 heat exchange area =50 m 2 Dirt density = 1200 kg / m 3 thermal conductivity = 0.2 W / (m·K)). System cumulative running time The data is recorded in real time by an internal system timer (e.g., t = 7200 s (2 hours)). Once all necessary parameters are ready, the heat transfer attenuation prediction module can use the aforementioned evolution model of dynamic coking fouling thermal resistance to calculate the current dynamic coking fouling thermal resistance at each time step. Constant terms are combined for calculation: sedimentation factor. Peeling resistance factor .calculate : Subsequently, the overall heat transfer coefficient of the condenser was considered in conjunction with the initial design. (For example = 500 W / (m 2 ·K), by using the evolution model of the real-time total heat transfer coefficient, the real-time total heat transfer coefficient of the condenser is calculated and updated. . This is updated in real time. The value will be passed to the subsequent integrated energy efficiency optimization module as an important input for its optimization control decisions.

[0060] Through the above technical solution, this application provides a precise and dynamic quantitative calculation model to address the engineering bottleneck of dynamic heat transfer attenuation caused by coking of escaped oligomers in antimony-free polyester processes. This model not only accurately reflects the dynamic evolution of coking fouling over operating time, avoiding the limitations of traditional static fouling thermal resistance calculations, but also updates the actual total heat transfer coefficient of the condenser in real time, providing accurate and reliable basic parameters for subsequent energy efficiency optimization and control. This enables the equipment to more effectively manage the heat transfer performance of the condenser, thereby ensuring the hue and molecular weight distribution of antimony-free polyester products and improving the operational stability and economic benefits of the entire esterification system.

[0061] This application further proposes that, in the comprehensive energy efficiency optimization module, the comprehensive net energy efficiency benefit objective function is constructed as a difference model between the economic benefit item of target product recovery and the total system operating cost item; the economic benefit item is linearly positively correlated with the actual gas-liquid separation efficiency; the total system operating cost item includes the circulating water pump power consumption item and the induced draft fan power consumption item, wherein the circulating water pump power consumption item is positively correlated with the cube of the cooling water control flow rate, and the induced draft fan power consumption item incorporates the attenuation degree of the real-time total heat transfer coefficient as a nonlinear back pressure penalty factor; in each control cycle, the comprehensive energy efficiency optimization module uses the cooling water control flow rate as the independent variable and adopts a numerical iterative optimization algorithm to calculate the gradient of the total derivative of the objective function, continuously updating the independent variable along the gradient ascent direction until the objective function converges to the extreme value, thereby outputting the optimal cooling water control flow rate; the comprehensive net energy efficiency benefit objective function and its numerical iterative optimization control law formula are as follows:

[0062]

[0063]

[0064] in: This represents the objective function for overall net energy efficiency benefits across the entire region. This represents the economic value conversion factor for ethylene glycol; This indicates the total liquid mass flow rate; Indicates the actual separation efficiency; This indicates the mass fraction of ethylene glycol in the recovered liquid phase; This represents the conversion factor for electricity costs; Indicates the resistance coefficient of the cooling pipes; This represents the independent variable of cooling water mass flow rate; Indicates the density of cooling water; Indicates the reference efficiency of the water pump; This represents the total mass flow rate of the gas phase. Represents the arithmetic mean; Indicates the density of the gas phase; Indicates the reference efficiency of the induced draft fan; Indicates the overall heat transfer coefficient in the initial design; Indicates the real-time overall heat transfer coefficient; Indicates the system back pressure penalty index; Indicates the first Cooling water flow control in the iterative step; Indicates the first Cooling water flow control in the iterative step; Indicates the adaptive search step size; This represents the gradient of the total derivative of the objective function with respect to cooling water flow rate. Electricity cost conversion factor. And the economic value conversion factor of ethylene glycol This is a commercial and market price constant, directly input by the system operator based on the daily electricity price and the market price of chemical products. Cooling pipe resistance coefficient. (dimensions are) The pump's reference efficiency is calculated by incorporating the pipeline's geometric parameters and can be derived using classical fluid dynamics formulas (such as Darcy's formula). and induced draft fan reference efficiency All quantities are defined by the equipment nameplate.

[0065] To ensure the feasibility of the above-mentioned overall net energy efficiency benefit objective function in industrial settings, the core economic settlement parameter in the formula, the mass fraction of ethylene glycol in the recovered liquid phase, is addressed. This device is equipped with specific physical sensing methods: preferably, an online tuning fork density meter or online refractometer is installed at the drain pipe at the bottom of the gas-liquid separator or at the outlet of the subsequent collection tank. Since the actual recovered liquid phase is mainly a binary mixture of ethylene glycol and water, the system, based on the built-in binary system density-concentration correlation curve or refractive index-concentration correlation curve, uses real-time readings of density or refractive index and temperature compensation signals to look up and calculate the current ethylene glycol mass fraction online. As a backup or periodic calibration method, manual sampling and offline analysis using gas chromatography (GC) are employed to ensure the long-term accuracy of this economic benefit calculation benchmark.

[0066] In this scheme, the overall net energy efficiency benefit objective function is... Aimed at quantifying the economic benefits of the entire gas-liquid separation and condensation recovery process, this objective function comprehensively considers the revenue from product recovery and the energy costs incurred in equipment operation. This objective function can be integrated as a software module into a distributed control system (DCS) or advanced process control (APC) system, receiving process data and performing calculations in real time. Furthermore, it can be embedded into dedicated industrial controllers or programmable logic controllers (PLCs) with sufficient computing power, enabling direct integration with process sensors and actuators. This objective function provides a comprehensive performance evaluation metric, going beyond single-objective optimization (e.g., maximizing recovery rate or minimizing energy consumption only), thus achieving a balance of economic benefits.

[0067] Numerical Iterative Optimization Control Law It is a tool for finding the mass flow rate of cooling water. An algorithm for finding the optimal value that maximizes the aforementioned objective function. This control law can be implemented in programming languages ​​such as Python, MATLAB, or C++ using standard numerical optimization libraries and integrated into existing control systems. Alternatively, it can be implemented using specialized optimization solvers provided in commercial APC software packages, which are designed for real-time industrial applications. This control law provides a systematic and adaptive mechanism that can adjust the cooling water flow rate in real time, ensuring that the system operates close to its economic optimum under various conditions.

[0068] Real-time total heat transfer coefficient As a penalty factor for fan power consumption, this coefficient means that when condenser coking leads to a decrease in heat transfer efficiency, it will affect the energy consumption term in the objective function, especially the fan power consumption component. When the efficiency is reduced, the power consumption of the fan will increase accordingly, thus "penalizing" the reduced heat transfer efficiency. The calculations can be performed by a dedicated module that continuously monitors the heat exchanger's operating parameters (e.g., inlet and outlet temperatures, flow rate) and applies heat transfer correlations for calculation. Alternatively, Estimation can also be performed using soft sensing techniques or machine learning models trained on historical operating data, providing robust real-time estimates even in the presence of noise in sensor data. This feature directly links the degradation of heat exchanger physical performance (coking) to the economic objective function, incentivizing optimization algorithms to find operating points that mitigate or effectively compensate for the effects of coking. The system back pressure penalty index... The preset empirical weighting tuning index is tuned according to the following engineering tuning principles: Based on the actual characteristic curves of the induced draft fan under different coking conditions, correlation registration is performed. When the on-site process is extremely sensitive to vacuum fluctuations caused by condenser coking, the tuning principle is improved. The value of (usually the range of values ​​is) By artificially amplifying the penalty weight of fan power consumption in the objective function, the control law is driven to issue larger cooling water flow commands during optimization iterations to strongly suppress coking; conversely, when the system prioritizes energy saving and emission reduction of the water pump, the flow rate can be appropriately reduced. The value of .

[0069] This application's solution achieves adaptive optimized operation of the gas-liquid separation and condensation recovery unit in the antimony-free polyester esterification process by deeply coupling dynamic process parameters with economic benefits. Specifically, the solution introduces a comprehensive net energy efficiency benefit objective function across the entire domain into the comprehensive energy efficiency optimization module. Numerical iterative optimization control law.

[0070] During device operation, the flow field feature prediction module continuously acquires the instantaneous differential pressure signal at the gas-liquid separator inlet and extracts its high-frequency fluctuation characteristics. Combined with fluid physical properties, it predicts the Sauter mean diameter of entrained droplets. Subsequently, the separation efficiency calculation module calculates the total liquid flow rate entering the separator based on the ratio of gas flow pressure to surface tension. The critical flooding velocity is dynamically corrected based on high-frequency fluctuation characteristics, and the actual gas-liquid separation efficiency is calculated by combining the Sauter mean diameter. These dynamically updated and The value is transmitted in real time to the integrated energy efficiency optimization module as a key input to the product recovery benefit term in the objective function, thereby accurately quantifying the recovery value of ethylene glycol products.

[0071] Meanwhile, the heat transfer attenuation prediction module predicts the dynamic coking fouling thermal resistance of the condenser based on the escape droplet mass flow rate, characteristic activity parameters of the antimony-free catalyst, and intrinsic fouling properties, and updates its real-time overall heat transfer coefficient. This real-time overall heat transfer coefficient Subsequently, it was incorporated into the overall net energy efficiency benefit objective function of the comprehensive energy efficiency optimization module, serving as a core component of the wind turbine power consumption penalty factor. When coking occurs in the condenser... When the efficiency decreases, the wind turbine power consumption penalty term in the objective function will increase accordingly, thus reflecting the additional energy cost caused by the reduction in heat transfer efficiency.

[0072] After receiving the aforementioned dynamic parameters, the integrated energy efficiency optimization module constructs and updates the overall net energy efficiency benefit objective function in real time. This objective function comprehensively balances the benefits of product recovery with the total energy consumption, including water pump and fan power consumption (considering heat transfer attenuation penalties). To achieve optimal operation, this module calculates the impact of the objective function on cooling water flow rate. The total derivative gradient is obtained, and a numerical iterative optimization control law is used to adaptively search the search step size. Continuously adjust cooling water flow rate This iterative process enables the system to converge rapidly in the direction of increasing net benefits, dynamically finding the optimal cooling water control flow command that maximizes overall net energy efficiency benefits.

[0073] Through the aforementioned mechanism, the proposed solution deeply couples flow field dynamics, separation efficiency, heat transfer attenuation, and equipment energy consumption to form a closed-loop adaptive optimization system. This enables the device to adjust its operating strategy in real time when facing complex dynamic conditions in the esterification boiling system, such as high-frequency fluctuations in the flow field, secondary droplet breakage, demister flooding, and condenser coking. This effectively balances product recovery and energy consumption, thus avoiding energy waste and system instability issues that may occur under traditional fixed-parameter control.

[0074] The following is a specific example to illustrate this. Assume that during the operation of the antimony-free polyester esterification process, the current operating condition's independent variable is the cooling water flow rate. = 20 kg / s, revenue and energy consumption conversion factor = 5.0, = 0.9, = 0.02, pipe resistance parameter Wind turbine reference energy consumption factor Punishment Index = 2.

[0075] In this iteration step, the microprocessor calculates the current running efficiency. And the objective function is estimated using the central difference method. The total derivative (assuming the gradient is calculated) According to the control law, let the adaptive search step size be... =0.5:

[0076] The controller then updates the opening command of the cooling water regulating valve, increasing the cooling water flow rate to 20.425 kg / s. The above cycle is repeated in the next control cycle (e.g., after 10 seconds) until the gradient approaches zero, at which point the system locks in the globally optimal flow rate, achieving an adaptive balance of energy efficiency.

[0077] This integrated energy efficiency optimization module can be deployed on an industrial control computer, which interacts with field sensors, actuators, and the DCS system via industrial Ethernet. The optimization algorithm can be implemented using a program developed in Python or C++, leveraging its powerful numerical computing capabilities for real-time gradient calculation and iterative optimization. The final output of the optimal cooling water control flow command is sent to the frequency converter of the cooling water pump via the DCS, thereby precisely adjusting the cooling water flow.

[0078] Through the above technical solution, this application effectively solves the technical problem that traditional devices cannot balance product recovery benefits and equipment energy consumption under dynamic operating conditions, easily leading to energy waste and system instability. Specifically, by constructing a comprehensive net energy efficiency benefit objective function and introducing the real-time total heat transfer coefficient as a fan power consumption penalty factor, this solution achieves deep coupling and quantification of product recovery benefits and power equipment energy consumption. This enables the system to comprehensively evaluate the economic benefits under different operating strategies, rather than simply pursuing high recovery rates or low energy consumption.

[0079] Furthermore, by employing a numerical iterative optimization control law, this scheme can dynamically adjust the cooling water flow rate based on real-time changes in flow field characteristics, separation efficiency, and condenser coking conditions. This adaptive optimization mechanism enables the device to continuously find the optimal operating point when facing dynamic disturbances such as pressure pulsation, secondary droplet breakage, and condenser coking in complex esterification boiling systems, thereby maximizing overall net energy efficiency gains.

[0080] Therefore, this solution not only ensures that the antimony-free polyester esterification process maintains a high product recovery rate under dynamic operating conditions, but also significantly reduces the additional fan and pump energy consumption caused by heat transfer attenuation, avoiding energy waste. This greatly improves the economic efficiency and stability of the equipment operation, guarantees the quality of antimony-free polyester products, and provides key technical support for the industrial application of the new antimony-free polyester esterification process.

[0081] In some of the embodiments described above in this application, core calculation processes such as prediction of the Sotter average diameter of entrained droplets and correction of critical flooding gas velocity are proposed. High-frequency pulsating interaction calibration coefficients and secondary entrainment sensitivity calibration coefficients are used. Empirical calibration parameters such as airflow shear stripping and scouring coefficients are also used for the calculation of fouling thermal resistance. However, in the implementation process, under different antimony-free polyester esterification conditions, the flow field fluctuation characteristics, secondary entrainment laws, and actual parameters of the coking process will change with the operating conditions. Fixed empirical parameters cannot match the changes in actual operating conditions, which will lead to a decrease in the accuracy of entrained droplet diameter prediction, separation efficiency calculation, and coking fouling thermal resistance prediction. This will affect the accuracy of subsequent optimal cooling water flow rate calculation and make it impossible to guarantee that the device will operate stably in the optimal energy efficiency state for a long time.

[0082] In this regard, this application further proposes that the device also includes a parameter adaptive calibration module. This module is a functional unit specifically designed for calibrating and updating key model parameters. It can be a software module integrated into an industrial control system (ICS) or a distributed control system (DCS), or it can be a standalone embedded controller. The parameter adaptive calibration module is used to acquire the instantaneous differential pressure signal characteristics and actual separation and capture data under offline operating conditions. This process aims to provide initial calibration data for the model parameters based on actual operating conditions. Specifically, during the device commissioning phase or under specific test conditions, the instantaneous differential pressure signal at the gas-liquid separator inlet can be acquired using a high-precision pressure sensor. Simultaneously, physical sampling and analysis methods (e.g., isodynamic sampling combined with particle size analysis or mass weighing) can be used to acquire the actual separated and captured droplet size distribution and mass flow rate data. The parameter adaptive calibration module derives dimensionless correction coefficients using a nonlinear regression fitting method. High-frequency pulsation interaction calibration coefficient and secondary entrainment sensitivity calibration coefficient This nonlinear regression fitting method can handle the complex nonlinear relationship between parameters and actual measured data, thereby accurately determining the initial values ​​of these key parameters. For example, optimization methods such as the Levenberg-Marquardt algorithm and the Gauss-Newton algorithm can be used to solve the problem by minimizing the error between the model's predicted values ​​and the offline measured data. The parameter adaptive calibration module corrects the actual heat transfer coefficient online during device operation using the gradient descent method. This online correction mechanism ensures that the parameters adapt to constantly changing operating conditions and fouling accumulation during unit operation. The actual heat transfer coefficient can be calculated by real-time monitoring of the fluid temperature and flow rate at the condenser inlet and outlet, and then comparing this actual value with the model prediction using the gradient descent method (e.g., by calculating the error function). (Dynamically adjust the partial derivatives and iteratively update along the negative gradient direction) The value of is used to ensure that the model predictions are consistent with the actual situation.

[0083] This application's solution introduces a parameter adaptive calibration module, implementing a hierarchical calibration strategy for key empirical parameters, combining offline initial calibration with online dynamic correction. Before the device is put into operation, the parameter adaptive calibration module first acquires the instantaneous differential pressure signal characteristics and actual separation and capture data under offline operating conditions. These data reflect the real flow field characteristics and separation effect under specific device structure and process conditions. Based on this offline data, the module uses a nonlinear regression fitting method to accurately determine the calibration coefficients for high-frequency pulsating interaction. and secondary entrainment sensitivity calibration coefficient The initial values ​​were then provided to the flow field characteristic prediction module and the separation efficiency solution module, enabling the prediction of the Sauter mean diameter of entrained droplets and the calculation of the actual gas-liquid separation efficiency to more accurately reflect actual operating conditions. During device operation, the parameter adaptive calibration module continuously monitored the actual operating status of the condenser and, based on the deviation between the actual heat transfer coefficient and the model prediction, used the gradient descent method to correct the airflow shear stripping and scouring coefficient online. This online correction mechanism ensures that the heat transfer attenuation prediction module can dynamically and accurately predict the thermal resistance of condenser coking fouling and its impact on the real-time overall heat transfer coefficient. Through this combination of offline initial calibration and online dynamic correction, the parameter adaptive calibration module provides continuous high-precision parameter support for the flow field characteristic prediction module, the separation efficiency solution module, and the heat transfer attenuation prediction module. This allows the prediction model of the entire device to always closely match actual operating conditions, thereby ensuring the accuracy of the integrated energy efficiency optimization module in constructing the objective function and solving the optimal cooling water control flow command, ultimately achieving long-term stable and efficient operation of the device under complex and variable operating conditions.

[0084] As a specific implementation, the parameter adaptive calibration module can be implemented by an industrial-grade computer system equipped with a high-performance processor and data storage unit. During the offline calibration phase, the system interacts with field sensors (such as differential pressure transmitters and flow meters) and laboratory analytical equipment (for droplet size and mass analysis) to acquire instantaneous differential pressure signal sequences and corresponding actual separation and capture data. For example, multiple data acquisitions under different operating conditions can be performed during the initial startup of the device or during specific process tests. Subsequently, the computer system runs a pre-set nonlinear regression fitting program, which can process the acquired data based on statistical methods such as least squares or maximum likelihood estimation to calculate the high-frequency pulsation interaction calibration coefficients. and secondary entrainment sensitivity calibration coefficient The specific values ​​are stored and used as initial parameters for the flow field characteristic prediction module and the separation efficiency solution module. During online operation, the parameter adaptive calibration module continuously receives operating data from the condenser via a real-time data interface, including cooling water inlet and outlet temperatures, process fluid inlet and outlet temperatures, and corresponding flow rates. Based on this real-time data, the system calculates the actual total heat transfer coefficient of the condenser. Simultaneously, the heat transfer attenuation prediction module will adjust the current... The system predicts a total heat transfer coefficient. The adaptive parameter calibration module compares these two heat transfer coefficients and, using a gradient descent algorithm, periodically (e.g., every few minutes or hours) adjusts the coefficients based on a preset step size and convergence criterion. The value is calculated until the deviation between the predicted and actual values ​​reaches an acceptable range. For example, the gradient descent algorithm can calculate the objective function (such as the squared difference between the actual and predicted heat transfer coefficients) on the curve. The gradient is calculated and updated along the negative gradient direction. This dynamic adjustment ensures It can adapt to the actual situation of dirt accumulation on the condenser surface, thereby maintaining the accuracy of heat transfer attenuation prediction.

[0085] Through the above technical solution, this application effectively solves the problem that traditional fixed empirical parameters cannot match the complex and variable operating conditions of antimony-free polyester esterification processes. The parameter adaptive calibration module ensures the calibration coefficients for high-frequency pulsating interactions by combining offline initial calibration with online dynamic correction. Secondary entrainment sensitivity calibration coefficient and airflow shear stripping and scrubbing coefficient This significantly improves the accuracy and real-time adaptability of key parameters such as entrained droplet average diameter prediction, actual gas-liquid separation efficiency calculation, and dynamic coking and fouling thermal resistance prediction of condensers. Because the accuracy of the core prediction model is guaranteed, the integrated energy efficiency optimization module can make decisions based on more reliable data, thereby accurately solving and outputting the optimal cooling water control flow command, avoiding control deviations and energy efficiency losses caused by inaccurate parameters.

[0086] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A gas-liquid separation and condensation recovery device for antimony-free polyester esterification process, characterized in that, include: Flow field feature prediction module: acquires the instantaneous differential pressure signal at the inlet of the gas-liquid separator and extracts its high-frequency fluctuation characteristics. Combined with fluid physical parameters, it predicts the Sauter mean diameter of entrained droplets. Separation efficiency calculation module: Calculates the total liquid flow rate entering the separator based on the ratio of gas flow pressure to surface tension, dynamically corrects the critical flooding velocity according to the high-frequency fluctuation characteristics, and calculates the actual gas-liquid separation efficiency and the mass flow rate of the escaped liquid droplets escaping to the condenser by combining the Sotter mean diameter. Heat transfer attenuation prediction module: Based on the escape droplet mass flow rate, characteristic activity parameters of antimony-free catalyst and intrinsic properties of fouling, predict the dynamic coking fouling thermal resistance of condenser and update its real-time total heat transfer coefficient. The comprehensive energy efficiency optimization module constructs a comprehensive net energy efficiency benefit objective function that includes product recovery benefits and equipment power consumption with heat transfer coefficient decay penalty. By calculating the gradient of the total derivative of the objective function with respect to cooling water flow rate, a gradient-based numerical iterative optimization algorithm is used to continuously update the independent variables along the gradient ascent direction to solve the problem and output the optimal cooling water control flow rate command.

2. The gas-liquid separation and condensation recovery device for antimony-free polyester esterification process according to claim 1, characterized in that, In the flow field feature prediction module, the prediction model for the Sauter mean diameter of the entrained droplets includes a static hydrodynamic baseline term and a dynamic pulsation correction term. The static fluid dynamics benchmark term characterizes the initial breakup effect of the basic shear force of the flow field on the droplets. It is positively correlated with the surface tension of the gas-liquid interface and negatively correlated with the gas phase dynamic pressure and Reynolds number. The dynamic pulsation correction term characterizes the secondary aerodynamic fragmentation effect of pressure pulsation on droplets. It is constructed as a decreasing function of the ratio of the root mean square of the high-frequency fluctuations of the instantaneous differential pressure signal to the arithmetic mean, and is used to reduce and correct the static hydrodynamic reference term.

3. The gas-liquid separation and condensation recovery device for antimony-free polyester esterification process according to claim 2, characterized in that, The separation efficiency calculation module includes a soft-sensor derivation sub-model, a critical flooding gas velocity correction model, and an actual separation efficiency calculation model, with the following specific configuration: The soft measurement extrapolation sub-model is used to calculate the total liquid mass flow rate entering the gas-liquid separator. The established gas-liquid entrainment mapping relationship is positively correlated with the total gas mass flow rate on the hot side and the local dynamic pressure of the gas phase, and negatively correlated with the surface tension of the gas-liquid interface. The critical flooding velocity correction model is established based on the standard flooding constant of the separator internals, and is corrected by introducing a dynamic pulsation characteristic variable that characterizes the secondary entrainment sensitivity of the liquid film. The dynamic pulsation characteristic variable is determined by the ratio of the root mean square of the high-frequency fluctuation to the arithmetic mean. The actual separation efficiency calculation model is constructed as a truncated probability distribution function. When the apparent gas velocity inside the separator is lower than the corrected critical flooding velocity, the separation efficiency is positively correlated with the droplet size ratio and the velocity margin, showing an exponential asymptotic growth relationship. When it is equal to or higher than the critical flooding velocity, the separation efficiency is zero. The total uncaptured liquid phase mass flow rate is the escape droplet mass flow rate to the condenser.

4. The gas-liquid separation and condensation recovery device for antimony-free polyester esterification process according to claim 3, characterized in that, In the heat transfer attenuation prediction module, the evolution model of the dynamic coking fouling thermal resistance of the condenser is constructed as an asymptotic saturation function that evolves over time. The evolution model is derived from the dynamic balance between the fouling deposition rate term and the airflow shear stripping rate term. The fouling deposition rate term is positively coupled with the antimony-free catalyst activity temperature correction coefficient, the escaped droplet mass flow rate, and the oligomer mass fraction; the airflow shear stripping rate term is positively correlated with the square of the apparent gas velocity inside the separator and the gas phase density. The real-time total heat transfer coefficient is calculated based on the thermal resistance of the initial design total heat transfer coefficient and the thermal resistance of the dynamic coking fouling through the principle of superposition of series thermal resistances.

5. The gas-liquid separation and condensation recovery device for antimony-free polyester esterification process according to claim 4, characterized in that, In the comprehensive energy efficiency optimization module, the comprehensive net energy efficiency benefit objective function is constructed as a model of the difference between the economic benefit item of the target product recovery and the total system operating cost item; The economic benefit item is linearly positively correlated with the actual gas-liquid separation efficiency; the total system operating cost item includes the circulating water pump power consumption item and the induced draft fan power consumption item, wherein the circulating water pump power consumption item is positively correlated with the cube of the cooling water control flow rate, and the induced draft fan power consumption item introduces the attenuation degree of the real-time total heat transfer coefficient as a nonlinear back pressure penalty factor. The integrated energy efficiency optimization module uses the cooling water control flow rate as the independent variable in each control cycle. It employs a numerical iterative optimization algorithm to calculate the gradient of the total derivative of the objective function and continuously updates the independent variable along the gradient ascent direction until the objective function converges to the extreme value, thereby outputting the optimal cooling water control flow rate.

6. The gas-liquid separation and condensation recovery device for antimony-free polyester esterification process according to claim 2 or 3, characterized in that, It also includes a parameter adaptive calibration module, which is used to acquire the instantaneous differential pressure signal characteristics and actual separated and captured data under offline operating conditions, and obtain dimensionless correction coefficients, high-frequency pulsation interaction calibration coefficients and secondary entrainment sensitivity calibration coefficients by nonlinear regression fitting, and correct them online by gradient descent method according to the deviation of the actual heat transfer coefficient during device operation.

Images