Data-driven Methods for Monitoring, Fault Diagnosis, Control and Optimization John MacGregor Ali Cinar ProSensus, Inc. Illinois Institute of Technology McMaster University
Feb 22, 2016
Data-driven Methods for Monitoring, Fault Diagnosis,
Control and Optimization
John MacGregor Ali CinarProSensus, Inc. Illinois Institute of TechnologyMcMaster University
Overview• An overall theme: Making use of historical plant data
• Empirical models • Optimization• Control
• Monitoring and fault diagnosis• Fault tolerant control
John MacGregor
Ali Cinar
Models• Mechanistic
– Structure from theory / Parameters from data– Advantages are well known– Problems:
• Assumptions that may be poor; theory for many y’s not known• May not incorporate many of measured variables • Examples: Y’s or X’s that are images or PAT sensors
• Empirical– Structure and parameters from data– Advantages are again well known:– Problems:
• Structure is often imposed and unrealistic, no interpretability nor any causality
Latent Variable Models - Concepts
Latent variable spaceMeasured variables
TX
t1
t2
(c) 2004-2010, ProSensus, Inc.
Summary statistics: T2 and SPE
Latent variable regression modelsTwo data matrices: X and Y
Symmetric in X and Y• No hypothesized relation between X and Y• Both X and Y are functions of the latent variables, T• Choice of what is X and Y depends upon objectives
X = T PT + E Y = T CT + F
TX Y
(c) 2004-2010, ProSensus, Inc.
Why Latent Variable Models?
• Low dimensional models – Define the space containing most of the information
• Simultaneously model both the X and Y spaces– Model structure truly determined by the data– This makes models unique and interpretable– Provides causal models in the low dimensional LV space
• Allows for active use of the model (eg. optimization)– Allows for
• easy handling of missing data• Easy detection of abnormal observations (*)
• Other regression methods (MLR, ANN, etc) do not share these advantages when using historical data. – Non-unique, uninterpretable, non-causal
Optimization in Latent Variable Spaces• For active use of model, must have causality
– Active use optimization / control / diagnosis• Historical plant data generally do not contain causal
information on individual variables– Nor will any model built from these data
• But latent variable models do provide causality in the low dimensional LV space (t1, t2, …)– Y = TCT X= TPT (t’s define Y and X)– T = XW* (To change T must move combinations of x’s)
• Optimization in low dimensional LV space– Then X and Y obtained from LV’s
• Illustrate concept with 2 industrial examples
Optimization: Injection molding process• GE water systems (2003)
– Polyurethane film manufacture very sensitive to humidity, temperature and raw material variations
– Operators periodically readjusted the process largely by trial and error
• Inject ~50 parts; measure ~10 quality variables; make adjustments– Injection velocity profiles, timing sequences, etc.
• Iterate until within specification– Provided a good set of data for LV modeling
• Nonlinear PLS model – 20 raw material properties; 26 process variables; 10 quality variables
• Models for both Y and Variance (Y)
– Constraints:• Humidity , temp and raw material properties constrained to
their currently measured values• SPE < ϵ; T2 < T2
99% These ensure validity of model
– Applied only when multivariate control limits violated– Results:
• Readjustment in one step• Improved quality• Reduced scrap• Operational since 2004
- 1 0
- 5
0
5
- 7 - 6 - 5 - 4 - 3 - 2 - 1 0 1 2 3 4 5 6 7
t[2 ]
23
24
25
2826
20
2930
21
27
39 37
363538
34
)(ˆ)(ˆ)(ˆ)(ˆ 21
4,..2,1,
newnewT
newdesT
newdest
tyQtytyyQtyyMina
anew
Optimization: Injection molding process
Optimization of a batch polymerization
Pilot plant data (Air Products & Chemicals)
Z X Y
Recipe & Initial Conditions Variable Trajectories
End Properties (13)
time
variables
batc
hes
• Very high dimensional optimization problem
• Easily solved in low dimensional LV space
Optimization for new product quality • Constraints or desired values are specified for the 13 y’s• Minimize batch duration• Optimization done in the three dimensional LV space
4
3
22
1
2
432
21
ˆˆ
)(),,ˆ,ˆ(
ˆ)(ˆ)(ˆ3,2,1
,
ayDaxCaT
aSPEtPLSTSPExy
xqSPEqTqtyyQtyy Tnewdes
Tnewdes
tMina
anew
Multiple solutions for Z, X-all satisfying the y specifications, but with different batch times.
Supervisory MPC of Batch Processes• Objective: Control final product quality
– Product quality only measured upon completion of batch– Control problem is thus one of
1. Predicting final quality from all the initial and evolving data2. Making optimal mid-course corrections at several decision
points during the batch (QP)3. Different objectives at each decision point
– PLS models have been shown to be ideal for modeling batch trajectory data and predicting final quality• Build from historical batch data• Plus some DOE runs at the decision points• Closed-loop identification used for subsequent implementations
Supervisory MPC of Batch Processes• Commercial systems in food industry
> 100,000 batches controlled> 99.9% up-time> 50% reduction in std dev of all final quality attributes- 20-40% increases in productivity
-0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.40
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
SP
with ABC
without ABC
Final quality attribute
• LV models also allow MSPC monitoring throughout the batch.
• This helps make controller robust to faults – e.g. wireless temp sensor failure – default controllers.
Summary (first part)
• Latent variable models provide powerful ways to use historical operating data – Can make use of all measured variables– Provide unique, interpretable models for analysis– Provide causality in the LV space for optimization,
control • Industrial examples used to illustrate this
– Provide monitoring and diagnosis capabilities (next part)
Implementation and Automation of Process Supervision
• Many variations of PCA: PCA, MBPCA, DPCA, …• Many techniques: PCA, PLS, Independent Component
Analysis (ICA), …• The Irish potato famine - single kind of potato (Lumper)
Diversity provides robustness• Develop a SPM, FD and control system that uses many
alternate techniques– How to decide which technique works better for a given
situation? Add a management layer– How to improve decision-making with experience? Use
distributed AI
Adaptive, Decentralized Process Supervision
Develop an agent-based monitoring, faultdetection, diagnosis, and control system to:– Coordinate alternative techniques for
reliable and accurate fault detection, diagnosis (FDD) and control
– Improve performance via: ocontext-dependent performance
assessment and decision-makingomulti-level learningoadaptation
Distributed Artificial Intelligence• Implement with Agent-Based Systems (ABS)• Decision-making is decentralized and divided
into hierarchical layers• Agents:
– are autonomous software entities– observe their environment – act on their environment according to predefined
rules/algorithms– may adapt by changing their rules/interpretation
based on their environmental conditions at run time
MADCABS: Monitoring Analysis Diagnosis and Control with Agent-Based Systems
• MADCABS is built using a hierarchical layout, with physical communication layer, process supervision layer and agent management layer
Sim u lato r
output: si mulatedsystem data
example: solver .dl l i nour case P la n t (re a l o r
s im u la te d )
P r o c e ss su p e r v is io n an d c o n tro lag e n ts .
T his laye r n ee d s p er fo rm an c ee v alu atio n .
M an ag e m e n t L aye r fo r ag e n tp er fo rm an c e e v a lu atio n an d
p r io r ity ass ig n m e n t
M A D C A B S' P h ysica lL a yer
S e nsors
A c tua tors
Preprocessing, monitoring, diagnosis, control
Collection of raw data from plantBasic information flow:
Mapping control actions back to processEvaluation of technique and agentperformance
Process Monitoring
• Statistical process monitoring techniques used– Principal component analysis (PCA)– Dynamic PCA (DPCA)– Multi-block PCA (MB-PCA)
Process
Calculates the performances: Accumulated performances of fault detection agents are summed to find the total performance of the monitoring agent
Builds new statistical models: When all monitoring agents are performing badly or the process operating mode changes.
Monitoring Organizer
Process
Fault Detection• Fault detection agents are the monitoring
statistics for PCA, DPCA and MB-PCA– Hotelling’s T2 and SPE statistics
Fault detection organizer
1. PCA_SPE2. PCA_T2
3. MB-PCA_SPE4. MB-PCA_T2
5. DPCA_SPE6. DPCA_T2
Fault detection agents
Gives out-of-control signals based on the consensus formed between fault detection agents
Observes performances of fault detection agents under different fault magnitudes and keeps history
Triggers diagnosis agent
Fault Diagnosis
Database
Diagnosis Training Agent
Diagnosis Agent
Diagnosis Manager
Fault Detection Organizer
Process
Consensus Fault Decision
Fault Identification
Agents
Fault Diagnosis: Identification techniques
• Contribution Plots– Variable
contributions to monitoring statistics T2 and SPE
• Fishers Discriminant Analysis (FDA)
• Partial Least Squares Discriminant Analysis (PLSDA)
Observation
SPE SPE = eeT
For an out-of-control observation:Squared Prediction Error SPE = ∑ej , j = 1, …, number of variables N
X1 X3 X5
Variable Contributions
Fault Type 1
Fault Type 2
Variance between classesVariance within classes
Classify new observation based on the closeness to the existing clusters
max
Fault Type 1
Fault Type 2
Classify new observation based on the class membership
y = BPLS x
1s
1s
Fault Type 1
Fault Type 2
X Y
Fault Diagnosis (Identification) Agents
Process
Identification (Discrimination)
Agents
FDA PLSDA
Contribution Map
Estimator
Project the new fault data on the model and
determine the most likely fault class
PCA_SPE [X1,X4,X7]PCA_T2 [X1,X4]MB-PCA_SPE [X1,X4]MBPCA_T2
[X1,X4,X7,X8]DPCA_SPE [X1,X4,X6]DPCA_T2 [X1,X4]
[X1,X4]Fault Signature for F1
Contribution Maps[X1,X4] : [F1, F1, …]
Agent Performance Management Layer
Performance Evaluation:• Record the performance of the agent and the
state metrics that define the state of the system when that performance is observed.
[State Metrici, i=1,…,I ] = f (performance)
• Compare the current state of the system to recorded states, and estimate the performance of the agent for the current state based on its performance for similar states in history.
State Metric 2
New Data Point:What would the performances
of each agent be for this state?
State Metric 1
Agent A
Agent B
Agent C
d1 d2
d3For each agent:- Identify performances at closest state points.- Obtain a performance estimate for the current state point by interpolation.
P1 P2
P3
Pestimate
Pest,A
Pest,C
Pest,B
Performance History Space
Diagnosis Performance History• Record:
– Fault signature• Fault signatures are the process variables significantly
contributing to the inflation of the monitoring statistic• Fault signatures are available once the fault is detected
– Performance of the agent for that fault signature• Performance is recorded only after diagnosis is confirmed
• Use the history to find: – Agents that are the best performers for the current fault
signature.
Diagnosis agent uses the estimated performances of fault identification agents for the potential fault to form the consensus diagnosis decision
Adaptation:Performance-Based Consensus Analysis- Agents update their built-in knowledge and
methods they use- Discriminant agents update their models
with current dataAdaptive
FDAAdaptivePLSDA
Contribution Map Estimator
Over time, after a diagnosis decision is confirmed for a fault type, the misclassifications are used to update the models of the adaptive instances
Fault-tolerant Control StructuresPl
ant o
r sim
ulat
or
System Identification
PID control
Set of Controllers
MPC Control
Controller Performance Assessment
Monitoring and Diagnosis
Single centralized control system
Decentralized control: . Local coordinated MPCs. Local MPCs integrated with local FDD modules using ABS
Summary/Conclusions • Latent variable models provide powerful ways to
use historical operating data • Data-driven methods are well-suited for
distributed process supervision• Learning and adaptation in monitoring, FDD and
control enable fault- tolerant control• MADCABS provides an environment for adaptive
fault diagnosis and fault-tolerant control• There are alternative approaches – Vive la
difference!
Acknowledgements
• IIT & ANL:• Fouad Teymour• Cindy Hood• Michael North• Arsun Artel• Inanc Birol• David Mendoza• Sinem Perk• QuanMin Shao• Derya Tetiker• Eric Tatara• Cenk Undey
Financial Support by National Science Foundation CTS-0325378 of the ITR program.
• McMaster University & ProSensus
• Many of my former grad students at McMaster
• The ProSensus team