1 Control of maldistribution of flow in parallell heat exchangers Magnus G. Jacobsen, Sigurd Skogestad Nordic Process Controi workshop, Porsgrunn 29.-30.

Slides:



Advertisements
Similar presentations
Application of the Root-Locus Method to the Design and Sensitivity Analysis of Closed-Loop Thermoacoustic Engines C Mark Johnson.
Advertisements

Dynamic Behavior of Closed-Loop Control Systems
Distillation Modeling and Dynamics Distillation Part 2.
Entropy balance for Open Systems
The First Law of Thermodynamics
Lecture# 9 MASS AND ENERGY ANALYSIS OF CONTROL VOLUMES
Lec 13: Machines (except heat exchangers)
Specialization project 2012 Temperature control of an unstable chemical reactor By Ola Sæterli Hjetland Supervisors: Sigurd Skogestad, Krister Forsman.
CHE 185 – PROCESS CONTROL AND DYNAMICS PID CONTROL APPLIED TO MIMO PROCESSES.
1 Single-cycle mixed-fluid LNG (PRICO) process Part I: Optimal design Sigurd Skogestad & Jørgen Bauck Jensen Quatar, January 2009.
Modeling of Coupled Non linear Reactor Separator Systems Prof S.Pushpavanam Chemical Engineering Department Indian Institute of Technology Madras Chennai.
More General Transfer Function Models
A Vapor Power Cycle Boiler T Turbine Compressor (pump) Heat exchanger
1 Operation of heat pump cycles Jørgen Bauck Jensen & Sigurd Skogestad Department of Chemical Engineering Norwegian University of Science and Technology.
Chapter 3.1: Heat Exchanger Analysis Using LMTD method
1 Single-cycle mixed-fluid LNG (PRICO) process Part II: Optimal operation Sigurd Skogestad & Jørgen Bauck Jensen Quatar, January 2009.
THEORETICAL MODELS OF CHEMICAL PROCESSES
CHAPTER IV INPUT-OUTPUT MODELS AND TRANSFER FUNCTIONS
CHAPTER II PROCESS DYNAMICS AND MATHEMATICAL MODELING
Lapse Rates and Stability of the Atmosphere
1 Active constraint regions for optimal operation of a simple LNG process Magnus G. Jacobsen and Sigurd Skogestad Department of Chemical Engineering NTNU.
ME421 Heat Exchanger and Steam Generator Design Lecture Notes 6 Double-Pipe Heat Exchangers.
Process Operability Class Materials Copyright © Thomas Marlin 2013
1 Modelling, Operation and Control of an LNG Plant Jens Strandberg & Sigurd Skogestad Department of Chemical Engineering, Norwegian University of Science.
The First Law of Thermodynamics
Transfer Functions Chapter 4
Analysis of Disturbance P M V Subbarao Associate Professor Mechanical Engineering Department I I T Delhi Modeling of A Quasi-static Process in A Medium.
© 2014 Carl Lund, all rights reserved A First Course on Kinetics and Reaction Engineering Class 23.
1 Outline Skogestad procedure for control structure design I Top Down Step S1: Define operational objective (cost) and constraints Step S2: Identify degrees.
PSE and PROCESS CONTROL
A Vapor Power Cycle Boiler T Turbine Compressor (pump) Heat exchanger
Dynamic Response Characteristics of More Complicated Systems
1 Single-cycle mixed-fluid LNG (PRICO) process Part II: Optimal operation Sigurd Skogestad & Jørgen Bauck Jensen Qatar, January 2009.
ME421 Heat Exchanger Design
1 Self-Optimizing Control HDA case study S. Skogestad, May 2006 Thanks to Antonio Araújo.
Last Time Where did all these equations come from?
1 Active constraint regions for optimal operation of chemical processes Magnus Glosli Jacobsen PhD defense presentation November 18th, 2011.
© 2014 Carl Lund, all rights reserved A First Course on Kinetics and Reaction Engineering Class 23.
1 Single-cycle mixed-fluid LNG (PRICO) process Part I: Optimal design Sigurd Skogestad & Jørgen Bauck Jensen Qatar, January 2009.
© 2014 Carl Lund, all rights reserved A First Course on Kinetics and Reaction Engineering Class 29.
Chapter 4. Modelling and Analysis for Process Control
Abstract An important issue in control structure selection is plant ”stabilization”. The paper presents a way to select measurement combinations c as controlled.
Lecture 21-22: Sound Waves in Fluids Sound in ideal fluid Sound in real fluid. Attenuation of the sound waves 1.
1 Optimization of LNG plants: Challenges and strategies Magnus G. Jacobsen Sigurd Skogestad ESCAPE-21, May 31, 2011 Porto Carras, Chalkidiki, Greece.
1 Decentralized control Sigurd Skogestad Department of Chemical Engineering Norwegian University of Science and Tecnology (NTNU) Trondheim, Norway.
Transfer Functions Convenient representation of a linear, dynamic model. A transfer function (TF) relates one input and one output: The following terminology.
Control limitations for unstable plants
INSTRUMENTATION & CONTROL
The Second Law of Thermodynamics Entropy and Work Chapter 7c.
Lecture 2: Mathematical Modeling Copyright © Thomas Marlin 2013 The copyright holder provides a royalty-free license for use of this material at non-profit.
Aachen Status Report: CO 2 Cooling for the CMS Tracker at SLHC Lutz Feld, Waclaw Karpinski, Jennifer Merz and Michael Wlochal RWTH Aachen University, 1.
Feedforward Control ( 前馈控制 ) Dai Lian-Kui Shen Guo-jiang Institute of Industrial Control, Zhejiang University.
Feedforward Control ( 前馈控制 ) Liankui DAI Institute of Industrial Control, Zhejiang University, Hangzhou, P. R. China 2009/04/22.
Chapter 5 Part 2 Mass and Energy Analysis of Control Volumes Study Guide in PowerPoint to accompany Thermodynamics: An Engineering Approach, 8th edition.
Date of download: 9/22/2017 Copyright © ASME. All rights reserved.
Dynamic modelling of a Hex with phase change
CHAPTER IV INPUT-OUTPUT MODELS AND TRANSFER FUNCTIONS
Stian Aaltvedt Supervisors: Sigurd Skogestad Johannes Jäschke
Aachen Status Report: CO2 Cooling for the CMS Tracker at SLHC
ES 211: Thermodynamics Tutorial 5 & 6
Chapter 5 The First Law of Thermodynamics for Opened Systems
Decentralized control
Controllability of a Granulation Process
Phase Transition Example
Stability of Congestion Control Algorithms Using Control Theory with an application to XCP Ioannis Papadimitriou George Mavromatis.
Process Operability Class Materials Copyright © Thomas Marlin 2013
Decentralized control
Proportional Control and Disturbance Changes
12. Heat Exchangers Chemical engineering 170.
Outline Control structure design (plantwide control)
Presentation transcript:

1 Control of maldistribution of flow in parallell heat exchangers Magnus G. Jacobsen, Sigurd Skogestad Nordic Process Controi workshop, Porsgrunn jan 2009

2 Outline Motivation Problem description Model description Simulations Analysis Conclusions and future work Nordic Process Controi workshop, Porsgrunn jan 2009

3 Motivation The problem is observed in the LNG plant at Melkøya and is probably one of the reasons for suboptimal operation At the same time, it is an interesting control challenge; –Can we control the system so we have equal flows in the two branches, using only one input? –Obviously, we can achieve this for flows outside the instability area, but by design, we are close to or inside this area. Nordic Process Controi workshop, Porsgrunn jan 2009

4 Problem description When having a vaporizing fluid distributed between two or more parallels, we can have different flow rates in each parallell This leads to temperature gradients inside the exchanger. This causes increased wear on heat exchanger material, and disturbance on operation of surrounding units. We may also get liquid in one of the exit streams, this may damage compressors Nordic Process Controi workshop, Porsgrunn jan 2009

5

6 Simple model Heat is transferred from a source of constant temperature and aritmethic mean ΔT is used Gas phase is assumed ideal and heat capacities are assumed constant Hydrostatic contribution to pressure drop is neglected Flow = √(ρΔP/k) where k is a constant Data used are for water Nordic Process Controi workshop, Porsgrunn jan 2009

7 Simulations With one parallell, to establish how overall pressure drop, temperature and density change with increasing flow. Model is run to steady state for different values of inflow Next plots show: –Pressure drop as function of molar inflow –Outlet temperature as function of molar inflow –Outlet density as function of molar inflow Nordic Process Controi workshop, Porsgrunn jan 2009

8

9

10 Nordic Process Controi workshop, Porsgrunn jan 2009

11 Dynamic simulation with two parallell tanks Individual flows may vary, but total is constant Total flow divided by 2 lies inside instable region Inlet pressures are equal –Adds inlet pressure as algebraic variable (DAE system) Plots show: –Inflows –Outlet vapour fractions –Temperatures –Inlet pressure (common for both tanks) Nordic Process Controi workshop, Porsgrunn jan 2009

12 Nordic Process Controi workshop, Porsgrunn jan 2009

13 Nordic Process Controi workshop, Porsgrunn jan 2009

14 Nordic Process Controi workshop, Porsgrunn jan 2009

15 Nordic Process Controi workshop, Porsgrunn jan 2009

16 Nordic Process Controi workshop, Porsgrunn jan 2009

17 Linear analysis Linearization (using Simulink) of the model after 30 seconds shows that the initial operating conditions are unstable Linearization after 300 seconds shows that the final operating conditions are stable This indicates that if the hot side temperature is not changing, the system will stabilize at the operating point where maldistribution is present Nordic Process Controi workshop, Porsgrunn jan 2009

18 Controllability We have considered the following measured variables: Overall pressure drop (dP), difference in internal pressure (ΔP), difference in outlet temperature (ΔT ) Total feed rate n in is the manipulated variable At the instable operating point, dP is the only measurement out of those considered which changes with total flow (the others are zero) Nordic Process Controi workshop, Porsgrunn jan 2009

19 At both operating points, the linearized system is state controllable, that is, the controllability matrix has full rank However, because of the fast transition from the unstable to the stable region, control might be difficult For example, the transfer function from total flow to overall pressure drop has a duplicate real RHP zero at almost the same location as its RHP pole. –At the instable solution, this is the variable that is simplest to measure. Nordic Process Controi workshop, Porsgrunn jan 2009

20 Control alternatives Control inside the instable area is difficult anyway Multivariable control – will it be fast enough? Split-range SISO control? –Inside instable area, use feedback to control pressure difference –When ΔT is nonzero, reduce total flowrate temporarily But can we then go back to nominal operating conditions without going unstable? Nordic Process Controi workshop, Porsgrunn jan 2009

21 Remember: Nordic Process Controi workshop, Porsgrunn jan 2009

22 Conclusions and further work Control inside the unstable area is difficult (but probably possible) For the LNG plant it may be easier because of slower dynamics Next step will be trying out different control strategies Nordic Process Controi workshop, Porsgrunn jan 2009

23 See you in Lund in 2010!