WP4b Activities M13M24 - PowerPoint PPT Presentation

1 / 17
About This Presentation
Title:

WP4b Activities M13M24

Description:

Partners: UNIDO, UVA, TUB, US, LTH, UCL. WP-Leaders: Sebastian Engell, Olaf Stursberg (now TUM) ... problems are based on engineering insight and reduction of ... – PowerPoint PPT presentation

Number of Views:76
Avg rating:3.0/5.0
Slides: 18
Provided by: ManuelR150
Category:

less

Transcript and Presenter's Notes

Title: WP4b Activities M13M24


1
WP4b Activities M13-M24
Partners UNIDO, UVA, TUB, US, LTH,
UCL WP-Leaders Sebastian Engell, Olaf Stursberg
(now TUM)
  • Predominantly work on the case studies
  • Challenging, industrially relevant problems
  • Significant continuous dynamics as well as
    discrete-event control
  • Safety-related discrete controls in processing
    plants
  • Critical for economic performance and plant
    life-time
  • Interaction of discrete logic with continuous
    dynamics
  • Not restricted to shutdown logic
  • Large transitions involving continuous and
    discrete controls
  • No systematic design procedures available
  • Degree of automation low, but automation highly
    desirable
  • Very complex problems, beyond the state of the art

2
Case Studies Overview
  • Reactor-tank example (UCL)
  • Derived from PVC plant at Solvay
  • Simple dynamics, periodic online scheduling
    problem
  • Batch plant (Magdeburg/ Berlin)
  • Reactor-filter system derived from color pigment
    production
  • Complex kinetics and interactions, batch process
  • Evaporator system (Dortmund/ Bayer Tech. Serv.)
  • Complex dynamics modeled by hybrid DAE-system
  • Ramp-up and safety logic verification problems
  • Fuel cell system (Sevilla)
  • Complex real laboratory system
  • Verification of control logic to prevent
    undesirable states

3
Case Studies Overview (2)
  • Sugar Plant Crystallization Section (Valladolid)
  • Real complex industrial process
  • Full and simplified simulation models available
  • Start-up and optimization problem
  • Open plate reactor (Lund/ Alfa-Laval)
  • System described by nonlinear PDEs
  • Startup problem
  • Laboratory plant (Valladolid)
  • Mixed batch/ continuous process
  • Scheduling and level/temperature control
  • Refrigeration system (Danfoss) (new)
  • Nonlinear model, switched operation
  • Feedback control and optimization of the
    operating point

4
Discontinuous Plant (TUB)
  • 3 products P1, P2, P3 have to be produced in
    given amounts.
  • Order of filtering operations matters.
  • Each reaction Ai?Pi is accompanied by a parallel
    reaction Ai?Wi
  • Concentrations of Wi are constrained.
  • First order reaction kinetics
  • Discrete control inputs reactor feed rates,
    filter allocations
  • Continuous control inputs heating and cooling
    rates
  • Safety constraints (bounds for temperatures)
  • Goal minimal operating cost (energy, material
    and processing time)

5
Evaporator System
Goal Start up in minimum time and with minimum
steam consumption
Counter-current flow liquid-vapor
3 components mixture A, H2O, Alcohol
?A
Continuous Inputs Vc, V1, V2, V3, V4 and VV2 0
100 Discrete Inputs VVS,1 and VVS,2
0 or 1
6
Evaporator Hybrid Process Model
  • Discrete dynamics
  • 4 locations per evaporator ? complete system 64
    locations
  • Combination of the heating and evaporation states
  • 2 discretely switched valves
  • Continuous dynamics
  • Nonlinear DAE-system
  • 17 state variables for one evaporator (4
    differential, 13 algebraic)
  • Total mass of alcohol in the evaporator
  • Flow rate of vapor into a heat exchanger
  • 6 continuously actuated valves

7
SRDC-CS Discretely Controlled Evaporator
  • Only the last evaporator is considered
  • The process is close to stationary operation
  • Inputs and outputs are held constant
  • Discretely switched redundant valve assemblies
  • Closed (0 )
  • Open (stationary setting)
  • Steam pipe PV

0 / 68.1
0 / 14.9
0 / 56.3
  • Assumption occurrence of two errors
  • Blocking of the valve assemblies in open or
    closed position
  • Obstruction of the pipe PV
  • Automatic error correction after 130 seconds
    (redundancy)

8
Open Plate Reactor (Lund/ Alfa Laval)
9
Open Plate Reactor pilot plants
10
Modeling and Start-up Problem Formulation
  • Detailed nonlinear pde model from first
    principles
  • The plate reactor is described as an ideal
    tubular reactor
  • Multiple equilibrium points, rapid transitions
  • Discretized to a system of ODEs using finite
    differences
  • Operation rules
  • Avoid accumulating unreacted chemicals
  • Avoid too high temperatures at hot spots

11
Supermarket Refrigeration System (Danfoss)
  • Task
  • Control the temperature Tair inside each display
    case between 2ºC and 5ºC
  • Keep Psuc below 1.7 (1.9) bars during day-time
    (night-time) operation
  • Avoid frequent compressor on/off
  • Minimize energy consumption
  • Control variables Disturbances
  • On/off valves - Amount and type of goods
  • On/off compressors - Heat transfer from
    surroundings

12
Problems with the Current Control Scheme
Synchronization
Large variations of the suction pressure
Many switches in compressor capacity
13
Case Studies Participation
  • Reactor-tank example ? 4 contributions
  • Batch plant ? 2 contributions
  • Evaporator system startup ? 2 contributions
  • Evaporator system verification ? 1 contribution
  • Fuel cell system ? 1 contribution
  • Sugar Plant Crystallization Section ? 1
    contribution
  • Open plate reactor ? 2 contributions
  • Laboratory plant ? 2 contributions
  • Refrigeration system ? 3 contributions

14
Lessons Learned
  • Case Studies
  • Reasonable set of case studies from simple to
    very complex
  • Highlight different aspects of practical control
    problems
  • Demanding test-cases for hybrid control
  • High industrial relevance, confirmed by advisors
  • Separation of concerns (hierarchy) and efficient
    parameterizations enable the (sub-)optimal
    solution of real-world problems.
  • All successful solutions of the more complex
    problems are based on engineering insight and
    reduction of the solution space.
  • Tool usage
  • Modelling and simulation environments
    optimizersEcosimPro / SQP-solver, gPROMS
  • MATLAB
  • MATLAB-tools under development
  • Need for the development of better methods and
    tools!

15
WP4b Meetings and Events in M13-M24
? Periodic WP meeting at the University of
Seville, December 16, 2005 Discussion of the
progress of work, planning for the upcoming
period. ? Session on the activities of WP4b and
overview presentation at the CEmACS/ HYCON
Workshop in Lund, June 1, 2006. ? Special
session on the results obtained within WP4b,
organized for the ADHS conference, Alghero, June
8, 2006. ? Periodic WP meeting during the ADHS
conference Planning of the post-conference
workshop discussion of the decisions required
for M4b.2.3 / M4b.4.2. ? Post-conference workshop
at the PSE/ESCAPE conference in Garmisch,
July 13, 2006. ? Brief WP meeting after the
post-conference workshop Planning of the
activities in M25-M42
16
WP4b - Milestones and Deliverables M13-M24
?
?
?
?
submitted in September 2006, since the
solution of the verification tasks turned out to
be significantly more challenging as expected.
17
Dissemination
  • Workshop in Lund WP 4b session and plenary
    presentation, discussion with industrial advisors
  • Invited Session at ADHS 2006
  • Post-conference Workshop at ESCAPE 16 / PSE 2006
  • Invited session proposal submitted to ACC 2007
  • 22 conference papers and contributions to books
    - 3 journal papers
Write a Comment
User Comments (0)
About PowerShow.com