Constraint Based Diagnosis and Recovery Assessment for Interconnected Chemical Processes

Diagnosis and recovery in an interconnected process require compatible representations of fault evidence and feasible interventions. We specify a constraint-based architecture, Constraint-based Diagnosis and Recovery (CDDR), that uses a signed process model to generate and rank diagnostic candidates and a steady-state model to assess recovery options. Inclusion-minimal diagnoses are distinguished from minimum-cardinality candidates, and recovery feasibility is treated as an operational assessment rather than a test of whether a diagnosis is true. A reported synthetic campaign comprises an eight-unit plant, 18 measured variables, 16 constraints and 1,000 episodes across five seeds. Reported top-1 unit-plus-fault-type accuracy is 0.573 for CDDR, 0.475 for univariate threshold monitoring, 0.246 for causal reachability and 0.149 for Principal Component Analysis (PCA) contribution analysis. CDDR has a reported false-alarm fraction of 0.015. Recovery validity is reported as 0.5556 for CDDR, 0.3788 for an alarm-driven comparator and 0.8430 for an oracle, with different scoring denominators. These are descriptive results pending verification of episode records, scoring rules and analysis code. CDDR performs poorly on fouling drift (top-1 0.011) and has the lowest accuracy at doubled observation noise (0.033). The campaign does not validate optimal intervention selection, transient safety, performance on the Tennessee Eastman (TE) process or industrial deployment.

Authors

Institutions

Publication Details

Journal
Systemic Analytics
Published
2026-11-12
DOI
https://doi.org/10.31181/sa44202696
Primary Topic
Fault Detection and Control Systems
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Constraint Based Diagnosis and Recovery Assessment for Interconnected Chemical Processes

Wolfgang H. Müller, Haruna Mavakumba Kefas, Marzieh Lamtar Gholipoor
Systemic Analytics
Fault Detection and Control Systems
article

Constraint Based Diagnosis and Recovery Assessment for Interconnected Chemical Processes

Wolfgang H. Müller, Haruna Mavakumba Kefas, Marzieh Lamtar Gholipoor
article en

Abstract

Diagnosis and recovery in an interconnected process require compatible representations of fault evidence and feasible interventions. We specify a constraint-based architecture, Constraint-based Diagnosis and Recovery (CDDR), that uses a signed process model to generate and rank diagnostic candidates and a steady-state model to assess recovery options. Inclusion-minimal diagnoses are distinguished from minimum-cardinality candidates, and recovery feasibility is treated as an operational assessment rather than a test of whether a diagnosis is true. A reported synthetic campaign comprises an eight-unit plant, 18 measured variables, 16 constraints and 1,000 episodes across five seeds. Reported top-1 unit-plus-fault-type accuracy is 0.573 for CDDR, 0.475 for univariate threshold monitoring, 0.246 for causal reachability and 0.149 for Principal Component Analysis (PCA) contribution analysis. CDDR has a reported false-alarm fraction of 0.015. Recovery validity is reported as 0.5556 for CDDR, 0.3788 for an alarm-driven comparator and 0.8430 for an oracle, with different scoring denominators. These are descriptive results pending verification of episode records, scoring rules and analysis code. CDDR performs poorly on fouling drift (top-1 0.011) and has the lowest accuracy at doubled observation noise (0.033). The campaign does not validate optimal intervention selection, transient safety, performance on the Tennessee Eastman (TE) process or industrial deployment.

Systemic AnalyticsVol. 4(4)
Islamic Azad University Quchan (IR), Modibbo Adama University of Technology (NG), Technical University of Munich (DE)
Fault Detection and Control Systems
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon