(494g) Optimization of Pre-Treatment Control Strategy for Thermal Cracking Hydrocarbon Feedstock

Thermal cracking of hydrocarbons is the major process in petrochemical industry for light olefin production. Such process converts hydrocarbon feedstock into more valuable products, by means of highly endothermic reactions. The performance of thermal cracking processes is influenced to a great extent by the feedstock composition and degree of saturation, as the product yield depends on the conversion level and extent of reaction.

In cracking furnaces, the hydrocarbon stream is first heated by flue gas and mixed with steam, to the initial cracking temperature (500 to 650oC). The feed stream is further heated up to 850oC in a fired tubular reactor, under controlled residence time, temperature profile and partial pressure. The hydrocarbons in the feedstock are cracked into smaller molecules in a short reaction time varying from 0.4 to 1 s. The process stream is then immediately cooled down in a transfer line exchanger (TLE), in order to stop further cracking (Karimzadeh et al., 2008).

For the satisfactory cracking of a given feedstock, a specific steam/hydrocarbon ratio (DS/HC) is required. In case of ethylene and propylene production, for which the feedstock consists of ethane and/or propane, better conversions are reached for steam/ hydrocarbon ratios varying from 0.2 to 0.5 (Chen et al., 1997).

As the steam/hydrocarbon ratio directly affects the cracking process, it must be kept in an appropriate value in order to reduce coke formation, hydrocarbon partial pressure and furnace residence time. The control strategy aims to accomplish the following targets: a) increase the production campaign; b) reduce the thermal load; c) increase the cracking severity and selectivity; d) increase the product yield.

In this research, the influence of feed properties on product yield of an industrial thermal cracker has been investigated. A multiple regression model was developed to identify the best operating conditions for the saturation process currently in use. The model was also used to improve the saturation control strategy. The study was performed for the Ethylene and Propylene production lines of a major Brazilian Petrochemical Plant.

PROCESS DESCRIPTION

The desired steam/hydrocarbon ratio is obtained by submitting the feedstock to a controlled humidification process in saturator tower. The feedstock comprises a mixture of fresh and recycled hydrocarbon flows, resulting from fractionation processes. The combined stream is preheated before feeding the thermal crackers.

The saturator encompasses a two-stage column fed by hydrocarbon feedstock flow, make-up water and process water flows. Saturation is obtained by the intimate contact between the process water and hydrocarbon gas stream. Previously to its injection into the column, the water flow is flashed in recirculation line, partially vaporizing and saturating the hydrocarbon flow. The make-up water consists of process water condensates, which are previously treated to integrate the recirculation water flow system. The Saturator bottom flow is pumped to the recirculation water heater, where the latent heat lost during vaporization is compensated by low-pressure vapour steam.

High pressure superheated steam is employed to vaporized the process water to thermal crackers as well as to complement the dilution vapour fed in the Saturators. The process water flow rate and inlet temperature must be controlled in order to obtain an adequate saturation.

OPTIMIZATION OF PROCESS CONTROL STRATEGY

Satisfactory control of the saturation systems is of great importance to improve the thermal cracking process efficiency. The process control strategy aims to guarantee a continuous and controlled saturated hydrocarbon flow into cracking furnace, under specific DS/HC.

The saturation control strategy currently in use in the plant is based on the temperature control at the outlet flow on top of the Saturator. Such temperature is a function of the feedstock molecular weight, operating pressure and DS/HC ratio setpoint. The saturation ratio setpoint is manually adjusted as a function of the feedstock flow rate.

The control strategy employs the signal of a pressure transmitter (PT) located on top of the Saturator, the DS/HC setpoint value (FFIC) and the feedstock molecular weight (AN/AHS) to calculate the temperature setpoint at the Saturator top outlet stream (TIC). The temperature controller on top of the Saturator (TC) compares the calculated setpoint to the actual measured temperature (TI) and actuates in the recirculation water flow rate (FC).

The DS/HC ratio is recalculated by the inverse correlation, based on the top temperature and pressure signals, and updated in the control loop. The ratio control strategy is indirectly obtained by controlling the top temperature, whose setpoint is calculated based on thermodynamic phase equilibrium.

The proposed optimisation approach consists to develop a new mathematical model built up from process simulations, and determines the most effective control variables. The model development involved the generation of thermodynamic equilibrium data, model definition and parameter estimation. Statistical techniques were used to validate the proposed model. Effectiveness of the manipulated variables was evaluated by means of static sensitivity analyses, in order to identify the variables that most influence the top temperature. In this contribution, saturation process of both ethane and propane feedstock have been studied. The control strategy was directly implemented in the DCS regulatory control level.

MATHEMATICAL MODEL

The proposed model correlation involves the variables required to define thermodynamic phase equilibrium: operating temperature, TT, and pressure, P, hydrocarbon molecular weight, Mw, and steam/hydrocarbon ratio, DS/HC.

Thermodynamic equilibrium data for saturated ethane and propane under the desired operating pressure conditions are not usually available in open literature. Therefore, simulation of the thermodynamic phase equilibrium was conducted under the conditions of temperature, pressure and compositions of interest, in order to generate such data.

Simulations of the saturation process were conducted in a heated pressurized drum fed with hydrocarbon and make-up water. Model assumptions consider total vaporization of both flows.

The multivariable regression model proposed for the generated thermodynamic equilibrium data constitutes a quadratic polynomial approach that takes into account interaction effects.

MODEL VALIDATION

Statistical techniques of parameter estimation and confidence interval estimation were used to assess the multiple regression model adequacies.

A sensitivity analysis of the critical process variables were conducted, regarding to the main estimated and measured parameters used in the Saturator control model. Such study aims to help evaluating the actual control strategy and to support the implementation of a more suitable control strategy.

Sensitivity analysis was based on a saturation process, simulated in a flash drum with three inlet flows: pure Ethane, make-up water and flashed process water. Adiabatic operation was assumed.

Sensitivity analysis of the Saturator top temperature regarding the analyzed operational parameters shown that the recirculation water temperature is the most critical variable. The great influence of such variable indicates its inclusion in the hydrocarbon/steam ratio control loop, as this ratio is correlated to the temperature by the regression model.

In the current strategy, the top temperature controller (TC) actuates directly in the recirculation water flow rate (FIC). The water temperature is only used as a set point for the low-pressure steam flow rate (FS) feeding the heat exchange, and therefore, the water temperature is not part of the DS/HC ratio control loop.

Analysis of the historical data trends corroborates the obtained results. The behavior of the parameters in a period of seven days shown that variations in the ethane feed temperature, due to reduction of the load flow rate, reflect in the top temperature, due to changes in the ratio set point. Those variations are promptly followed by changes in the water temperature. The immediate response of the water temperature is due to operator intervention, which continuously adjusts the water temperature set point according to the saturation behavior.

The recirculation water flow rate, whose influence is smaller, does not follow the top temperature variations. This influence may be observed in the historical data trend, which shown that significant differences can be detected between the flow rate curve, set by the controller, and the top temperature response.

CONCLUSIONS

Optimisation of the hydrocarbon feedstock saturation has been proposed to enhance set point tracking and disturbance rejection of the control strategy currently in use. The proposed approach consists to develop a new mathematical model built up from process simulations and to determine the most effective control variables. Static sensitivity analysis led to the identification of recirculation water temperature as the most important control variable.

The adjustment obtained by using the proposed model is considered satisfactory, as the mean error of the controller, estimated by analyses of the historical data, was approximately 0.1 oC. No further adjustment is required, as the mean error of the model is already inferior to the one of the controller.

Process data show that the optimised control strategy provided an enhanced performance of the cracking furnaces of Ethylene and Propylene Production Plants, as the desirable hydrocarbon feedstock saturation level has been achieved.

Chen, Q., Schweitzer, E., Van Den Oosterkamp, P. Berger, R., Smet, C., Marin, G. (1997). Oxidative Pyrolysis of Ethane. Ind. Eng. Chem. Res., (36), 3248-3251.

De Schepper, S., Heynderickx, G., Marin, G. (2009). Modeling the evaporation of a hydrocarbon feedstock in the convection section of a steam cracker. Computers and Chemical Engineering, (33), 122-132.

Karimzadeh, R., Godini, H., Ghashghaee, M. (2008). Flowsheeting of steam cracking furnaces. Chemical Engineering Research and Design, doi:10.1016/j.cherd. 2008.07.009. Unpublished.