Transcript History of modern control - Department of Chemical Engineering

Computers have had a profound
impact on
- automatic control
- automation
- manufacturing
Other uses:
- mathematical modeling
- simulation
1
Changing Manufacturing
Requirements
Good 

Fast  any two out of three before 1990,
Cheap 
and
Green: new addition for 21st Century
2
Feedback Control Is Basic Building
Block (Since 1950s)
Output
Setpoint
Feedback
Controller
Updated
Process State
Process
Inputs
Process
Observer
Process
Outputs
Quality
Measurement
Feedback
Information
3
Beginnings of Advanced Process
Control (APC)
• First usage of APC was in guidance and control
of aircraft/satellites.
• Due to complexity of these systems, PID control
was inadequate.
• Digital computer control was required for
analysis of the differential equations.
1957 – Sputnik launching
USSR/USA competition in control technology
(Maximum vs. Minimum Principle)
4
1960s/1970s – a split developed between
“modern” and “classical” control camps
• Time domain vs. frequency domain
• Optimization vs. PID tuning
• Automatic control became an
interdisciplinary field.
• PID control was still dominant in process
industries.
5
6
Gap Between Control Theory and
Practice
• Explosion of information since 1960s
• “You can get 80% of the profit with 20% of
the effort”.
• “What can go wrong will go wrong”.
7
“The author has been reading the
chemical process control literature for over
25 years and in his opinion the vast
majority of papers contained little or no
material useful in the daily practice of
control engineering”. (ca. 1986)
8
Why APC Has Not Been Used
• There are very few pilot installations for
testing control algorithms.
• Proprietary processes and great variety of
processes prevent technology transfer.
• Engineers design safe self-regulatory
processes – then use large inventories
and blend products.
9
• You can’t make any money with APC.
• Inter-disciplinary problem – knowledge required
includes control theory, engineering, advanced
math, statistics.
• Small yield for effort – plants have other
problems to solve that will give more significant
increase in production, yield, quality, etc.
• Math model of process required in process
control – not easy to get for some processes.
10
Three Types of Control (ca. 1975)
1. Feedback
2. Feedforward
3. Divine intervention
11
Major Developments Influencing Growth of
APC Since 1970s
•
•
•
•
•
Energy crisis
Distributed control hardware
Environmental restrictions
Quality control (international competition)
Computing speed
12
Computers (as of 1960)
Maximum
Average
Monthly
Rental
(1960 $)
Core
Storage
Capacity
(in 1000
bits)
Add
Time
(Microsec)
Read
Cards
Per
Min
IBM-7090
55,000
160
0.004
250
CDC-1604
34,000
32
0.005
1300
DEC-PDP1
2,200
4
0.010
(Tape
Input)
13
14
Key to Better Controllers?
• Better mathematical models and
instrumentation
• Key concept in new generation of
feedback controllers – they are
“Model-Based”
• Tuning based on optimization criteria
rather than frequency response but model
accuracy is a requirement
15
Model Predictive Control (MPC)
• Most widely used multivariable control algorithm
in chemical process industries
• Makes explicit use of process model (related to
Kalman filter)
• Control actions obtained from on-line
optimization (QP)
• Handles process variable constraints
• Unifies treatment of load, set-point changes
• Many commercial packages
16
17
Desirable Closed Loop
Performance
•
•
•
•
•
Tight control about a set point
Fast, smooth set point changes
Insensitivity to model errors
Insensitivity to plant changes
Ease of on-line tuning
18
Early Ideas About MPC
One technique for obtaining a feedback
controller synthesis from knowledge of openloop controllers is to measure the current control
process state and then compute very rapidly for
the open-loop control function. The first portion
of this function is then used during a short time
interval, after which a new measurement of the
process state is made and a new open-loop
control function is computed for this new
measurement. The procedure is then repeated.
Lee and Markus (1967)
Foundation of Optimal Control Theory
19
What is Coming Next?
• Faster hardware – MPC of units with more than
10 inputs and 10 outputs is already established
in industrial practice. Larger MPC
implementations and faster sample rates will
probably accompany faster computing.
• Better MPC algorithms – Improved algorithms
could easily have more impact than the
improved hardware for the next several years.
20
• MPC on the DCS - What will be the ratio of
PID to MPC loops if this happens?
• Nonlinear Models - When will control
based on nonlinear models become part of
industrial practice? The nonlinear MPC
theory and algorithms are improving
steadily as are nonlinear model
identification technologies.
21
22
23
Figure 19.1 The five levels of process control and optimization in
manufacturing. Time scales are shown for each level.
24
Supply Chain Management
•
•
•
•
•
•
Anticipate customer requirements
Commit to customer orders
Procure new materials
Allocate production capacity
Schedule production
Schedule delivery
25
Process Dynamic Modeling
Approaches
•
•
•
•
•
Empirical
Semi-empirical
Theoretical/fundamental
Flowsheet simulator
Nonlinear/linear
26
Nonlinear Response
Reboiler
duty fixed
27
Some Quotes about Modeling
• All models are wrong but some are useful.
• It is much easier to prove a model wrong
than prove it right.
• It is better to be approximately right than
exactly wrong.
28
• Models that accurately represent the plant
over the full operating region are
necessary.
• Very high computer speeds are required.
- Dynamic models will need to be run at
50-500 times real time to meet
application objectives.
29
Improved Instrumentation and
Control Technologies
• Nonlinear model predictive control
• Process and controller monitoring, fault
detection
• Estimation and inferential control
• Identification and adaptive control
• Plantwide control (design vs. control)
• Process sensors
• Microfabricated instrumentation
• Information and data handling
30
Process Control – 21st Century
• Factory of the future
-
B.S. engineer/operator
Nonlinear programming
Self-tuning controllers
Data reconciliation, filtering
Artificial intelligence
31
Future DCS Operator
•
•
•
•
Requests simulation optimization runs
Analyzes/implements control moves
Makes decisions to improve profits
Maintenance scheduling, shutdown
planning
• Analogy to airline pilot (process unit $ >
airplane $)
32
33
34
Manufacturing and Operations in
the Future
• Operations are guided by complete
information, i.e., integration of sales,
marketing, manufacturing, supply, and
R&D data accomplished.
• Data and information flow in a seamless
fashion along the whole supply chain from
raw materials suppliers through all the
steps of manufacturing operations to the
customer.
35
• Computer networks with wireless
communication capability connect all
components of the supply chain.
• Individuals on a need to know basis will
have instantaneous reliable access to
data, information, and decision-support
tools that will help them to do their job
regardless of their geographical location.
36