robust design via operability

McMASTERU N I V E R S I T Y

1280 Main Street West, Hamilton, ON, Canada L8S 4L7

Thomas E. MarlinDepartment of Chemical Engineering

And McMaster Advanced Control Consortiumwww.macc.mcmaster.ca

Copyright © 2007 by Thomas Marlin

ROBUST DESIGN VIA OPERABILITY

57th CSChE Conference

Session on Process Design and Analysis

October 29, 2007, Edmonton, Alberta

GOALS OF THE PRESENTATION

• Provide a (not the) definition of operability

- A taxonomy of issues

• Demonstrate that operability is not a collection of “tricks”

- Improves teaching of fundamentals- Motivates students to learn/apply

• Whet your appetite, refer to WEB site for expanded coverage – with workshops

• Suggest that instructors share teaching materials to improve our understanding and to reduce teaching load

Too little time!

FC1

TC1

TC2

T10

T12

T11

T13

fuel

LC1

L2

LAHLAL

F4

To flare

The process must be

OPERABLE!

I think that I agree.But, what do you mean by

OPERABLE?

WE DO NOT HAVE A CONSISTENT UNDERSTANDING OF THE ISSUES INVOLVED IN OPERABILITY

We have recently encountered a communication difficulty, so Process Operability = Robust Design

FC1

TC1

TC2

T10

T12

T11

T13

fuel

LC1

L2

LAHLAL

F4

To flare

OPERABILITY BECOMES ESSENTIAL WHEN WE CONSIDER REALISTIC VARIABILITY AND

UNCERTAINTY

Feed composition

Temperature

Production rate Reaction kinetics

Pressure drop

Efficiency

This is not obvious, it

must be learned

FC1

TC1

TC2

T10

T12

T11

T13

fuel

LC1

L2

LAHLAL

F4

To flare

We must introduce appropriate equipment to ensure that a process is operable

OPERABILITY CAN NOT BE “ADDED-ON” AT THE END OF THE DESIGN

“The principle sins of flowsheets used for economic evaluation are sins of omission … frequently omitted items include storage tanks, surge tanks, duplicated equipment (for reliability), startup equipment, emergency safety equipment, ..” (Valle-Riestra, 1983).

CRITERIA FOR SELECTING OPERABILITY TOPICS

• Generic issues that apply to wide range of systems

• Reinforce fundamental principles

• Introduce common issues and solutions in engineering practice

- Not comprehensive for any particular process

- Demonstrates principles, students can expand to other issues and technical solutions

- Leading to respect for making a real physical systemfunction

• Prepares for performing a major project

Key Operability issues

1. Operating window

2. Flexibility/controllability

3. Reliability

4. Safety & equipment protection

5. Operation during transitions

6. Dynamic Performance

7. Efficiency

8. Monitoring & diagnosis

FC1

LC1

IS THIS PROCESS OPERABLE?

Feed tank

Feed-effluent exchanger

Packed catalyst bed, exothermic reactions

Key Operability Topics

1. Operating window


3. Reliability




7. Efficiency



1. Operating window


3. Reliability




7. Efficiency


Key Challenge

How do we enable students to solve complex problems with multiple

objectives?

• Provide superstructure for knowledge

• Tie to professional skills: problem solving, group skills,report writing, oral presentation, etc.

• Link to basics

• Show compelling, practical examples (within the students grasp)

• Encourage (require) students to investigate, discover and evaluate issues in real processes

1

2

3

15

16

17LC-1

LC-3

LC-2

Feed drum attenuatescompositiondisturbances

Averaging level controlattenuates flow ratedisturbances

dP-1

dP-2

To flare

T5

T6

TC-7

AC-1

LAHLAL

LAHLAL

PAH

PC-1

P3

F3F4

F7

F8

F9

Feed flow rate andcompositiondisturbances

PV-3

TAL

T10

L4


1. Operating window


3. Reliability




7. Efficiency


OPERATING WINDOWMaximum fluid flow

Net positive suction head

Feed rate and composition

changes

What determines the window?

Maximum & minimum vapor rate


1. Operating window


3. Reliability




7. Efficiency


FLEXIBILITY AND CONTROLLABILITY

CW

NC

PC ??

Where must we provide adjustable equipment to maintain the process in safe and profitable operation?

Control pressure?

Principle: Q = U A (T) Which to influence?


1. Operating window


3. Reliability




7. Efficiency

8. Monitoring & diagnosis Not recommended!

CW

PC

LC

NC


Q = U A (T) Which to influence?

T too high


1. Operating window


3. Reliability




7. Efficiency



Q = U A (T) Which to influence?

CWPC

LC

NC

Fully open

Recommended!


1. Operating window


3. Reliability




7. Efficiency


FC1

TC1

TC2

T10

T12

T11

T13

fuel

LC1

L2

LAHLAL

F4

To flare

RELIABILITY

* Copyright by CCPS/American Institute of Chemical Engineers and copied with the permission of AIChE

What aspects contribute to reliability?


1. Operating window


3. Reliability




7. Efficiency


RELIABILITY

FC1

TC1

TC2

T10

T12

T11

T13

fuel

LC1

L2

LAHLAL

F4

Parallel pumps

Diversity

By-pass valves

Hot spot


1. Operating window


3. Reliability




7. Efficiency


SAFETY & EQUIP. PROTECTION

ALARMS

Safety Instrumented System

RELIEF

CONTAINMENT

EMERGENCY/COMMUNITY RESPONSE

BPCS

Layers of Safety Protection

• Concentrate on the first four layers

• Stay close to the process


1. Operating window


3. Reliability




7. Efficiency


FC1

TC1

TC2

T10

T12

T11

T13

fuel

LC1

L2

LAHLAL

F4

To flare



What aspects contribute to safety?


1. Operating window


3. Reliability




7. Efficiency


FC1

TC1

TC2

T10

T12

T11

T13

fuel

LC1

L2

LAHLAL

F4

To flare


Fail closed valve

Alarm

SIS for low flow rate

Safety Relief *


Temperature control


1. Operating window


3. Reliability




7. Efficiency



HAZOP method provides a structured manner for safety analysis (using qualitative analysis)

Note the specific location Unit: ________Fired heater____________________ Node:___air pipe after compressor and valve_ Parameter: __Pressure____________________

Guide Word

Deviation Cause Consequence Action

Stoppage of power to motor or turbine turning the compressor

Uncombusted fuel in the fire box – danger of explosion Uncombusted fuel – wasted fuel

SIS based on the rotation of motor shaft *

Break of coupling between motor and compressor

“

SIS based on rotation of compressor shaft*

Failure of compressor, e.g., breakage of blades

“ (plus danger from

flying metal)

Closure of air valve due to failure

“ Fail open valve

Any of the above “ SIS that measures the flow of air after the pipe and activates the shutdown if the flow if too low

lower

Low pressure in the fuel pipe node

Closure of air valve due to poor decision by operator

“ Air flow controller with ratio to fuel flow

fuel

air

feed

product

FC

TC

Principles

Engineering practice

MONITORING & DIAGNOSISKey Operability issues

1. Operating window


3. Reliability




7. Efficiency


FC1

TC1

TC2

T10

T12

T11

T13

fuel

LC1

L2

LAHLAL

F4

TC20

TY15

>

• What is important?

• How frequently must we monitor?

• What should we measure?

• What should we calculate?


1. Operating window


3. Reliability




7. Efficiency


MONITORING & DIAGNOSISRapid decisions – made by operating personnel

FC1

TC1

TC2

T10

T12

T11

T13

fuel

LC1

L2

LAHLAL

F4

TC20

TY15

>

Alarms

P

T

P

T

T

Extra sensors for diagnosis

P


1. Operating window


3. Reliability




7. Efficiency


MONITORING & DIAGNOSISLonger-term decisions – made by engineers

FC1

TC1

TC2

T10

T12

T11

T13

fuel

LC1

L2

LAHLAL

F4

TC20

TY15

>

MONITOR

• Heater efficiency

• Reactor conversion and selectivity

• Material balance

• Exchanger fouling

• Pressure drops through system

• Time each pump in serviceSensors +Calculations!


1. Operating window


3. Reliability




7. Efficiency


MONITORING & DIAGNOSIS

Feed tank

FC1

P3

V300

TC3

T4

Fueloil

F2

T7

Producttank

C.W.

F7

Air

Intake

FC5

Time

Temp

Fuel flow

Feed rate

Process Trouble Shooting

1. Engage

2. Define

3. Explore

4. Plan

5. Implement

6. Evaluate

Systematic Problem Solving Method!


1. Operating window


3. Reliability




7. Efficiency


MONITORING & DIAGNOSIS

Hypothesis Initial information Diagnostic ActionT sensor drift Neutral Check with temperature at

exit of reactorFuel valve is stuck open Disprove

(Temperature wouldincrease)

Place flow controller inmanual and make small

change to controller outputFeed rate causing T decrease

(TC too slow)Disprove

(previous changes werecontrolled)

________________

Feed tank

FC1

P3

V3

00

TC3

T4

Fueloil

F2

T7

Producttank

C.W.

F7

Air

Intake

FC5

L1

L2

P1

P3

T5

T6

Time

Temp

Fuel flow

Feed rate

High level in V-30

Sensor error

Delta pressure sensor calibrated incorrectly (reading higher level than actually exists)

Connection point (tap) blocked/corroded (level measurement is constant)

Too much liquid into the tank

TC-5 poorly tuned (oscillating and creating too much vapor overhead)

Steam valve fails open (unsafe)

Too little liquid leaving the tank

Poor feedback control

Magnitude of feedback controller gain (Kc) is too small

Valve malfunction

Reflux or product flow valve failed closed (safe)

Pump malfunction

Vortex (unlikely with high level)

Caviation

Power loss (motor failure or coupling break)

Increased feed rate (level controller will lower level in time)

Increased propane feed (level controller will lower level in time)

Distillate liquid product valve saturation

Reflux or product flow valve stuck , not responding to control signal


1. Operating window


3. Reliability




7. Efficiency


FISHBONE DIAGRAM FORMONITORING & DIAGNOSIS

Student “Triad” Groups for Trouble Shooting Tutorials

Monitors the TS’ers method and provides feedback after exercise

Applies the method and verbalizes thinking process

Has studied the case and provides responses to diagnostic actions (but no hints)

Art work by D. Woods

INSTRUCTOR’S EXPERIENCES

Ammonia reactor and separation loop Milk powder evaporators and fluid bed

drier Municipal water purification plant Desalination plant by reverse osmosis Ice cream production

Penicillin production (reactor and separation)

Refrigeration and cooling tower plant Boiler feed water treatment and storage Kraft pulp digester Wine production

• Any problem-based teaching style will likely satisfy needs

• Heavy load to develop

• Operability involves generic topics that are applicable to essentially any process (would have to be modified for product design). See sample projects from previous years.

STUDENTS’ EXPERIENCES

The Ugly To much work!

The Good• Students enjoyed the problem solving tasks (HAZOP and

Trouble Shooting)

• Defined diverse projects & found good operability issues

The Bad• Difficulty recognizing causes of variability and

uncertainty

• Challenge to “work backwards”: effect cause

• Needed to build experience with qualitative process analysis

A PROPOSAL FOR Ch.E. INSTRUCTORS• Integrate Robust design/process operability in the

capstone design course

• No instructor has experience with all issues

• Limited educational material is available that is accessible to undergraduates

Proposal to establish a portal for robust design/ process operability, with educational materials for public use

Expanded copy of white paper and power point lessons are available atwww.pc-education.mcmaster.ca/

A RECOMMENDATION FOR PRIOR COURSES

• Typical courses need to introduce causes of variability and uncertainty.

• Solutions should ensure operability (at least in s-s)

Variability in production rate: 70-110% of base design

0.01

0.1

1

0 2 4 6 8 10

(CA

0-C

c)/C

A0

time (min)

Uncertainty in Rxn kinetics

L

TP Determine

the reactor

volume

A RECOMMENDATION FOR PRIOR COURSES

• Courses need to indicate that processes are changed to achieve desired conditions.

• Engineers should be able to analyze the process qualitatively

We want to increase the feed flow rate by 10% and maintain the conversion unchanged.

What do we do?

L

TP

NO NEW PRINCIPLES, but Applications, Problem Solving, and Integration are

Unique

Processfundamentals

Design a CSTR,

V, T, F, …

Monitor the Reactor

Performance(conversion, yield, ..)

Troubleshoot CSTR(Why is the temperature oscillating?)

Safety HAZOP Analysis(What if pressure is high?)

Focus on fundamentals

We would appreciate comments, criticism, suggestions at any time, now

or by email.


1. Operating window


3. Reliability




7. Efficiency


OPERABILITY IN UNDERGRADUATE EDUCATION


1. Operating window


3. Reliability




7. Efficiency


OPERATING WIDOW

A B

-rA = k0 e -E/RT CA

feasible

infeasible

infeasible

Shape of op. Window?

Solvent

T

A

Reactant

Coolant

•Depends of variability

•Depends on flexibility

•Not rectangular!

Principles: Size equipment!

OPERATING WIDOW


1. Operating window


3. Reliability




7. Efficiency


RELIABILITY

A)

B)

C)

7290900 3 .).( AR

92709001123 .).( BR

97009001132 .. CR

Increased reliability with increased complexity and cost. We have added redundancy with parallel paths.

No redundancy

System-level redundancy

Module-level redundancy

Probability that the process will perform its function properly (one “path” must function)


1. Operating window


3. Reliability




7. Efficiency



FC1

P3

V30

0

TC3

T4

Fueloil

F2

AirIntake

FC5

Learning goals are process principles of SIS: measurements, logic and actions

Feed flow > minAir flow > minTemp < max

F100

F210

T305

…..

Fuel valve

Stack damper

……..

s

fc


1. Operating window


3. Reliability




7. Efficiency


OPERATION DURING TRANSITIONS

Steady-state processes

• Start-up and shutdown

• Regeneration

• Short runs with frequent switches

• Load following (highly variable demand)

Unsteady-state processes

• Batch

Special equipment and procedures (controls) are required during transition.

Equipment capacity must satisfy peak demand, not daily or batch average!


1. Operating window


3. Reliability




7. Efficiency


OPERATION DURING TRANSITIONS

FC1

TC1

TC2

T10

T12

T11

T13

fuel

LC1

L2

LAHLAL

F4

To flare

Heating required during start-up

Disposition of effluent during start-up

(recycle, flare, storage, …)

Extra sensor with large range


1. Operating window


3. Reliability




7. Efficiency


DYNAMIC PERFORMANCE

Rapid compensation for disturbances and timely changes to set points – the process side of process control

AC1

LC

Reduced disturbances

AC +

Feedforward

TC

CascadeAC

Faster feedback


1. Operating window


3. Reliability




7. Efficiency


DYNAMIC PERFORMANCE

FC1

TC1

TC2

T10

T12

T11

T13

fuel

LC1

L2

LAHLAL

F4

To flare

By-pass gives fast feedback dynamics

Inventory to smooth flow, composition and temperature

disturbances

With safe, smooth operation making consistently high product quality, is there more to do? YES!


1. Operating window


3. Reliability




7. Efficiency


EFFICIENCY & PROFIT

Operability requires extra capacity for extreme situations. We can take advantage during most times to increase profit.

Safety

Product quality

Production rate, etc

optim

izat

ion

Set

point


1. Operating window


3. Reliability




7. Efficiency


EFFICIENCY & PROFIT

FC1

TC1

TC2

T10

T12

T11

T13

fuel

LC1

L2

LAHLAL

F4

TC20

TY15

>

Operate at highest conversion without excess temperature

Minimize heating using fuel, while keeping by-pass open

robust design via operability

Documents