CaltechTHESIS
  A Caltech Library Service

Design and control studies on the fluid catalytic cracking process

Citation

Huq, Iftikhar (1998) Design and control studies on the fluid catalytic cracking process. Dissertation (Ph.D.), California Institute of Technology. http://resolver.caltech.edu/CaltechETD:etd-02012008-111240

Abstract

NOTE: Text or symbols not renderable in plain ASCII are indicated by [...]. Abstract is included in .pdf document. Fluid Catalytic Cracking (FCC) units are widely used in the oil refining industry to crack low value hydrocarbons into a range of higher value hydrocarbons including gasoline. Because of its feed processing flexibility, the FCC process is considered a primary conversion unit in an integrated refinery and optimal FCC operation can have a significant impact on the refinery profitability. Because of its important role in the refinery, the FCC process has recently attracted renewed interest both from academia and industry, with the main emphasis on better understanding and operating the process. In general, several possible alternatives can usually be postulated or envisioned to realize operational improvements, and the critical task is then to rank order these alternatives. Rank ordering the available alternatives requires the utilization of suitable: 1. process models describing each alternative, 2. tools to quantitatively assess the operational performance achievable with each alternative. "Operational performance" here refers to the quality of closed loop process response. The different alternatives have often been analyzed and compared through extensive simulations. However, because of stringent environmental and economic requirements, most industrial plants, including the FCC process have evolved into highly complex and integrated processes and ranking through simulation might be infeasible. In ranking design alternatives through simulation, two crucial factors must be considered if the conclusions drawn from such studies are to be reasonably reliable: • Detailed case studies need to be conducted for all the alternatives. These case studies typically require considering the effect of a wide range of disturbances, inputs, operating conditions, model structures and parameters, control structures and parameters. It might be necessary to examine many hundreds if not thousands of possible alternatives. Moreover, the effect of many other aspects of practical operation such as plant/model mismatch, process constraints, sensor/actuator failures, etc. on the operational performance must also be considered for each case, and makes the assessment significantly more involved. • While qualitative "good/bad" indicators might have been sufficient in the past, it is becoming more important to quantitatively rank order the available alternatives in todays highly competitive environment. Indeed, choosing a "slightly better" alternative over a "good" alternative can have a significant impact on the sustained profitability of the process. Because of the large number of possible cases that must be considered, contradictory conclusions can often be drawn when using such an approach to rank the available alternatives. Such a simulation approach is therefore infeasible except possibly for the simplest of cases. This suggests the utilization of more efficient screening tools to effectively assess the many design alternatives possible for improved process operation while accounting for the practical considerations just discussed. With respect to the requirements (1) and (2) above for ranking the possible alternatives, we note that: 1. FCC models: The renewed interest in improved FCC operation is partially reflected by the recent publication in the open literature of two first principles models claiming to describe the essential features of the FCC process dynamics (Arbel, Huang, Rinard, Shinnar and Sapre 1995, McFarlane, Reineman, Bartee and Georgakis 1993). 2. Tools for operational performance assessment: The structured singular value (SSV) framework introduced by Doyle (1982) provides a quantitative measure of the achievable operational performance, explicitly accounts for plant/model mismatch, and also permits the approximate incorporation of many of the practical requirements indicated above into the analysis. Utilizing the two first principles FCC models and the structured singular value framework, several aspects of the design and control of FCC processes have been addressed in this thesis, with special emphasis on quantitatively ranking possible design/control options from an operational performance improvement viewpoint: 1. Model IV FCC units differ from other cracking units in that model IV FCC units do not have slide valves in the catalyst circulation lines to enable direct control of catalyst circulation rate through the unit. Reducing fluctuations in catalyst circulation rate is found to significantly improve closed loop performance of this FCC unit. Based on this process insight: (1) an operational modification: operating the regenerator with the catalyst overflow weir always flooded, and (2) a design modification: installing slide valves in the catalyst circulation U- tubes are studied and shown to reduce catalyst flow fluctuations and improve the operational performance of the FCC unit. It is interesting to note that essentially all new FCC units and revamped older FCC units typically include both the above two modifications. To reflect this current FCC operation it is assumed for all the subsequent analysis in this thesis that the above two modifications are already incorporated into the FCC design. 2. A quantitative rank ordering of possible control structures for FCC operation in both partial combustion (PC) mode and complete combustion (CC) mode operation is undertaken. It is argued that intentional transitions between the PC and CC modes is rare. For FCC units operated within design specifications, 2 x 2 control structures are found to be sufficient for effective regulatory control. In particular, it is found that high coking feeds cannot be processed in CC mode operation as the safe upper limit on regenerator temperature can be violated. For PC mode regulation, riser temperature [...] and regenerator dense bed temperature [...] are the most suitable choice for controlled variables. For CC mode regulation, [...] are the most suitable choice for controlled variables, where [...] is the flue gas [...] molar concentration. If higher coking feeds are to be processed, the FCC would need to be refitted with additional process units, as discussed in Point 4 below. For both PC and CC modes, using feed temperature [...] and combustion air rate [...] as manipulated variables is found to provide operational performance levels comparable to those achieved with more conventional control structures typically used in industry (using catalyst circulation rate [...] and combustion air rate). That feed temperature can be an effective manipulated variable can have significant implications for the improved operation of older FCCs (e.g. the Model IV) where catalyst circulation rate cannot be directly manipulated. 3. Decentralized regulatory control of the FCC unit operating in PC mode is examined in detail. Decentralized PI controllers are considered, as they are most often used in the industrial setting. It is found that (1) Decentralized PI controllers can provide satisfactory regulation over the normal FCC operating range. (2) A common regulatory control strategy for PC mode operation is to use decentralized PI controllers for the pairing [...](notation in Point 2). It is shown that "tightening" the two PI controllers in fact reduces the process interactions and improves operational performance. This is counterintuitive - typically, detuning the controllers is expected to reduce process interactions and improve performance. (3) Certain FCC operating points are unstable. In the absence of control action, the FCC process would either drift to a stable operating point at a higher regenerator temperature or wind down to the cold state (negligible regenerator combustion and reactor cracking). It is shown that reasonable transitions to unstable operating points can be effectively handled with an appropriately tuned decentralized PI controller. (4) A theoretical development that provides insight into possible changes in the sign of the steady—state and infinite frequency relative gain array (RGA) elements is also presented. It is found to be related to the non—minimal phase behavior of one or more of the following (a) the overall plant, (b) the individual loop transfer functions, (c) the remaining system if some controlled variable and the associated manipulated variable is removed. Stable plants for which pairing for performance can correspond to negative steady—state RGA pairings are thus identified. 4. There is a growing need in the oil refining industry to process higher coking feeds in FCC units. However, FCC units operated in the CC mode might not be able to process sufficiently high coking feeds (as discussed in Point 2 above), and would have to be refitted to accommodate this greater feed processing flexibility. The most significant operating limitation is found to be the upper limit on regenerator temperature. The two possible refitting options considered here to alleviate the impact of this constraint are: (1) converting to partial combustion mode (PC mode) operation by installing a CO boiler and (2) extending complete combustion mode (CCE mode) operation by installing a catalyst cooler. The main objective is to compare the steady—state and dynamic characteristics of these two refitting options. From a steady state viewpoint, it is found that installing the catalyst cooler has a distinct advantage over installing the CO boiler when processing high coking feeds - because the regenerated catalyst is always clean burned. In particular, the product yields as well as overall conversions are typically higher. A comparison of dynamic characteristics suggests that while a simple decentralized control strategy provides satisfactory performance over a wide operating regime in PC mode, a more sophisticated control strategy might be required to achieve tight control over a comparable operating range in CCE mode with the catalyst cooler.

Item Type:Thesis (Dissertation (Ph.D.))
Subject Keywords:fluid catalytic cracking process
Degree Grantor:California Institute of Technology
Division:Chemistry and Chemical Engineering
Major Option:Chemical Engineering
Thesis Availability:Public (worldwide access)
Research Advisor(s):
  • Morari, Manfred
Thesis Committee:
  • Unknown, Unknown
Defense Date:13 October 1997
Record Number:CaltechETD:etd-02012008-111240
Persistent URL:http://resolver.caltech.edu/CaltechETD:etd-02012008-111240
Default Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:446
Collection:CaltechTHESIS
Deposited By: Imported from ETD-db
Deposited On:20 Feb 2008
Last Modified:26 Dec 2012 02:29

Thesis Files

[img]
Preview
PDF (Huq_i_1998.pdf) - Final Version
See Usage Policy.

10Mb

Repository Staff Only: item control page