Browsing by Subject "Distillation apparatus -- Automatic control"
Now showing 1 - 5 of 5
Results Per Page
Sort Options
Item Experimental demonstration of nonlinear model based control techniques on a lab-scale distillation column(Texas Tech University, 1991-05) Pandit, Hemant GopalNot availableItem Nonlinear model predictive distillation control using an extended neural Hammerstein model(Texas Tech University, 1998-05) Rangaratnam, BalachandranModel Predictive Control has been successfully applied in the chemical and petrochemical industries due to its intuitiveness and constraint handling capabilities. However most currently applied techniques use linear models that are valid only in the neighborhood of the operating point. Model predictive control using nonlinear models does have significant potential for efficient control over a wide operating range. This is particularly important for distillation control which is characterized by highly nonlinear, interactive and nonstationary behavior. The main challenge of nonlinear model predictive control is to develop accurate dynamic models. Phenomenological modeling is difficult, and computationally intensive. Hybrid models, that combine conventional identification techniques with alternative modeling approaches like neural networks, are favored because of their flexibility, computational efficiency, and ability to learn complex nonlinear mappings in a reasonable time. The Hammerstein modeling strategy simplifies the identification by separating the steady-state and transient components. In this project, an extended Hammerstein model was developed for use in a nonlinear model predictive control framework. The static nonlinear element of the Hammerstein model was modeled as a feed-forward neural network model, and the nonlinear dynamic element was identified as transfer function models with input-dependent adaptive dynamic parameters. Two distillation columns were modeled: a propylene-propane (C3) splitter operating at base case and at high purity and a toluene-xylene column. Steady-state and dynamic data were obtained from rigorous simulators developed previously. A dynamic model of the C3 splitter at base case using internally recurrent neural networks was also developed. Nonlinear model predictive control using the extended Hammerstein model was tested on dynamic simulations of each column. Nonlinear model predictive control using the recurrent dynamic model was tested on the C3 splitter at base case. The control performance was compared with that of PI controllers for each column for setpoint and disturbance rejection.Item On control of high relative volatility distillation columns(Texas Tech University, 1999-05) Duvall, Philip MarshallA detailed study of single-ended and dual-ended product composition control of four high relative volatility distillation columns, depropanizers with relative volatility ranging from 1.5 to 2.0, was conducted with emphasis on control configuration selection. Product impurities in the overhead and bottoms for the four designs ranged from 2.0 mol% (low purity) to 0.1 mol% (high purity). Rigorous, tray-to-tray steady state and dynamic simulations for four multi-component depropanizers were developed. The base case design (0.5 mol% impurity overhead and bottoms) was benchmarked against data from an industrial depropanizer. The simulations were used to compare nine different control configuration pairings, [L,V], [L,B], [L,V/B1, [L/D,V], [L/D,B], [L/D,V/B1, [L,V/B], [D,V/B], and [D,B], using Proportional-Integral (PI) control. Steady State Generic Model Control (GMC), and Dynamic Matrix Control (DMC). Controllers were tuned for setpoint changes and were tested for disturbance rejection performance using unmeasured feed composition changes (step and sinusoidal). All control studies were for unconstrained process control. For single-ended composition PI control, reflux (L) and reboiler duty (V) provided optimal control of the overhead and bottom compositions, respectively. For dual-ended composition PI control, the [L/D,V/B] configurations provided superior control for feed composition disturbances when compared to other configurations. Two-way decoupling of PI control for the [L,V] configuration resulted in significant control improvement in the [L,V] configuration for all product purity designs. The addition of feedforward control to the [L/D,V/B] configuration provided marginal, if any, improvement in depropanizer control when compared to the control performance of the [L/D,V/B] configuration without feedforward. Multi-Model Decoupled GMC (MMD-GMC) was introduced as a control technique to improve traditional GMC by providing dynamic compensation of the steady state targets to account for dynamic differences in vapor/liquid traffic in the column. The use of steady state MMD-GMC with the [L/D,V/B] configuration outperformed double ratio PI control for the low to mid purity designs (2.0 mol% to 0.5 mol% impurity in overhead and/or bottoms) but showed poor performance when compared to PI control for the high purity design (0.1 mol% impurity in overhead and bottoms). The [L,V] configuration combined with [2x2] Dynamic Matrix Control (DMC) control provided superior control performance for the low to mid purity depropanizer designs and outperformed double ratio PI control. For high purity distillation, DMC control performance using the [L,V] configuration was on par with the double ratio, dual-ended PI composition controller.Item Process model based control of distillation columns(Texas Tech University, 1988-12) Sinha, RupakNot availableItem Vacuum distillation control(Texas Tech University, 1998-12) Anderson, John JosephA detailed study of vacuum distillation column control implementations was performed with special emphasis placed on control configuration selection. Two vacuum separations were studied; toluene from xylene and ethylbenzene from styrene. Rigorous, dynamic simulations were developed for these two systems that incorporated varying tray-to-tray pressure drops and coupled, dynamic material and energy balances for each tray of the column. These columns were benchmarked against published data. For single-ended composition control, manipulating the reflux flowrate provided the best control of the overhead impurity for setpoint changes as well as for feed composition disturbance rejection. Bottom impurity control was best handled by ratioing the vapor boilup rate and the bottoms flowrate (boilup ratio, V/B). For dual-ended control, the [L,V] and [L,V/B] configurations provided better control of both product streams when the column has a reflux ratios near 1. In addition, the [L/D,V] and [L,V] configurations both provide good product impurity control especially when the bottom product stream is more valuable. These two configurations also performed best as the column's reflux ratio increased. Advanced control techniques such as decoupling and feedforward compensation were studied and decoupling was found to improved control performance on both product streams. Feedforward compensation improved configurations were ratio control was implemented (reflux ratio or boilup ratio) or when the process has slow dynamics. In addition, Dynamic Matrix Control (DMC) was applied to both the xylene/toluene columns and the styrene/ethylbenzene column. A [2x2] DMC controller was compared with decentralized PI controllers on several control configurations. For setpoint tracking, DMC improved control responses by decoupling control action on both ends of the column. For unmeasured feed composition disturbances, DMC did not have the control performance of PI as DMC lacked feedforward compensation for disturbances. In addition, DMC considered both product compositions as having equal importance. Finally, the minimum move suppression factors allowed by the DMC package were used which limited DMC performance for unmeasured disturbances. DMC does provide benefits on the styrene/ethylbenzene column by allowing the styrene composition to have a higher control priority. As a result, DMC performed comparably to PI for feed disturbances.