Condition Based Maintenance of Critical Machinery Assets: An Intelligent Archite

Condition Based Maintenance of Critical Machinery Assets: An Intelligent Archite

Condition Based Maintenance of Critical Machinery Assets: An Intelligent Archite

Sholley, Diana, Features Reporter has reference to this Academic Journal, PHwiki organized this Journal Condition Based Maintenance of Critical Machinery Assets: An Intelligent Architecture Dr. George Vachtsevanos Georgia Institute of Technology School of Electrical in addition to Computer Engineering Atlanta GA 30332-0250 (404) 894-6252 Voice (404) 894-7583 Fax Presented at the Workshop on Automated Machinery Maintenance The University of Texas at Arlington July 17, 2003 Topical Outline Introduction – What is Condition Based Maintenance Elements of a CBM Architecture Example Demonstration An Intelligent Agent Based CBM Paradigm Future R&D Directions/Concluding Comments. Condition-Based Maintenance The Opportunity Condition Based Maintenance (CBM) promises to deliver improved maintainability in addition to operational availability of naval systems while reducing life-cycle costs The Challenge Prognostics is the Achilles heel of CBM systems – predicting the time to failure of critical machines requires new in addition to innovative methodologies that will effectively integrate diagnostic results with maintenance scheduling practices

Texas A&M University-San Antonio TX

This Particular University is Related to this Particular Journal

Condition Based Maintenance Objective Determine the “optimum” time to per as long as m maintenance Problem Definition A scheduling problem – schedule maintenance timing to meet specified objective criteria under certain constraints Condition Based Maintenance Major Objective Extend system life cycle as much as possible without endangering its integrity Enabling Technologies Various Optimization Tools Genetic Algorithms Evolutionary Computing A Maintenance Management Architecture Enabling Technologies Genetic Algorithms as long as Optimum Maintenance Scheduling Case-Based Reasoning in addition to Induction Cost-Benefit Analysis Studies

CBM Per as long as mance Assessment Objective: To assess the technical in addition to economic feasibility of various prognostic algorithms Technical Measures: Accuracy, Speed, Complexity, Scalability Overall Per as long as mance Measure: w1Accuracy + w2Complexity + w3Speed + (wi – weighting factors) Per as long as mance Assessment Matrix: Prognostics Objective Determine time window over which maintenance must be per as long as med without compromising the system’s operational integrity Prognostics Enabling Technologies Multi-Step Adaptive Kalman Filtering Auto-Regressive Moving Average Models Stochastic Auto-Regressive Integrated Moving Average Models (ARIMA) Forecasting by Pattern in addition to Cluster Search Variants Analysis Parameter Estimation Methods Others

Prognostics Enabling Technologies (cont’d) AI Techniques Case-Based Reasoning Intelligent Decision-Based Models Min-Max Graphs Petri Nets Soft Computing Tools Neural Nets Fuzzy Systems Neuro-Fuzzy PEDS Software System Architecture (St in addition to -alone) The Navy Centrifugal Chiller

Chiller Failure Modes Timing sequence(1)-No fault Detected Data Collection Feature Extraction 1. The timing sequence is managed by the Task Manager 2. Algorithm modules are started by FEATURE READY events 3. Each diagnostic module decides upon the presence or absence of a fault 4. The diagnostic modules report their conclusion to the database. 5. Each diagnostic module runs its routine in addition to responds back to the task manager. 6. Task manager receives the events in addition to decides which module or algorithm should be started. 7. The diagnostic decision (or No fault) is displayed on the GUI;GUI receives result from database. 8. All prognostic routines are initiated when a fault has been detected. FuzzyDS WNN t t1 t0 Start next cycle t2 No Fault Detected Prognostic routines will not run Mode Identification (Extract features For Diagnostics only) Timing sequence(2)—Fault Confirmed Data Collection Feature Extraction FuzzyDS WNN Fault confirmed Start prognosis DWNN CPNN t t1 t2 t3 Fault Detected By FuzzyDS or WNN Start confirmation Continue confirmation Fault confirmed t4 Mode Identification Extracts features as long as prognosis Collect data as long as prognosis t0

Timing sequence(3)—Fault not confirmed Data Collection Feature Extraction FuzzyDS WNN Start confirmation No fault detected here Start normal cycle again Fault Detected By FuzzyDS or WNN Continue confirmation t1 t0 t2 t3 t4 t Software Design Developing plat as long as m: Microsoft Visual C++ , Visual Basic, SQL server 2000, Access 2000. Software running under Windows NT plat as long as m Component based open system architecture. All the system components are implemented as Microsoft COM objects Event-based distributed communication. Capable of transmitting events at different priorities. Provides database as long as storage of collected data,configuration in as long as mation, diagnostic in addition to prognostic results. Operator User Interface Administrator User Interface Diagnostics Feature Extraction Prognostics ICAS Database Interface Features Repository Interface Tier 1 Tier 2 Tier 3 User Services Prognostic Services Persistent Services Conceptual Model of PEDS — 3-Tier Client-Server Architecture

Software Diagram Prognosis Diagnosis Feature extractor Data sampling module Sampling Data Start Sampling Data Task Manager Start Feature Extractor Enough Data Start Diagnosing Time Out Start Prognosing Save to Database Get Features From Database Save results to Database Wait as long as event Do Prognosis Get Features From Database Save results to Database Wait as long as event Do Diagnosis Get Data From Database Save Features to Database Wait as long as event Do Extraction N y y N Events Event Event Feature Ready N y Time Out Database relationships Mode Diagram Example Testbed: AC-Plant

Mode Identification Modes are characterized by the dynamics, set-points, in addition to controller. Modes switch due to events. Fuzzy Petri Net (Mode Changes Due to Events) Dynamics Classifier (Mode Due to Dynamics) Mode Decision Sensors Current Mode Fuzzy Petri Net Normal Mode 53 44 Events Characterized by Membership Functions If Chilled Water Inlet Temperature is above 53 degrees in addition to Chilled Water Outlet Temperature is above 44 degrees then Switch to Overload Mode other modes other modes Overload Mode Fuzzy Petri Net Simulation SENSORS Pre-rotation vane position Chilled water inlet temperature Chilled water outlet temperature MODE CHANGES Normal Load Mode Full Load Mode Overload Mode Fuzzy Petri net marking Sensors mode marking

Sholley, Diana Inland Valley Daily Bulletin Features Reporter

Feature Selection in addition to Extraction Motivation: Data driven diagnostic/prognostic algorithms require as long as fault detection in addition to classification a feature vector whose elements comprise the “best” fault signature indicators Intelligent distinguishability in addition to identifiability metrics must be defined as long as selecting the best features Time in addition to frequency domain analysis techniques must be employed to extract the selected features Sensors Pre- Processing Feature Extraction Fault Classification Prediction of Fault Evolution Data Rough Set Feature Selection Feature Extraction Preprocessing Rough Set Rule Generation Raw Database Overall Procedures as long as Feature Selection in addition to Diagnostic Rule Generation Feature Vector Table Diagnostic Rulebase PEDS Database Featurebase Feature Preparation Rough set theory is a popular data mining methodology, which provides mathematical methods to remove redundancies in addition to to discover hidden relationships in a database. Feature Extraction York Test Database Feature Preparation Available Measurements from York Test Preprocessing Preprocessing includes removing unreasonable objects from the database in addition to assigning the operational status. Fault Symptoms from ONR Report Heuristic Feature Pre-selection Raw Data Feature C in addition to idates Featurebase Fault Mode1 Time Feature 1 Feature 2 t1 t2 Decision 0 1 20.5 11.5 23.5 9.5

Feature Extraction Architecture Figure 1-2. Industrial Diagnostic/Prognostic Framework based on Data Mining Feature Selection. Rough Set Data MIner Data Calibration Classifier Designer Preprocessing GUI Feature Extractor Feature Selector Rule Generator Diagnostic/Prognostic Module Diagnostor Prognostor Feature Vector Table Feature Table Diagnostic Rule Table Featurebase Database Operator Raw data On-line Feature Values Historical Feature Value Feature Vectors Historical FeatureValues Feature Vectors Diagnostic Rules Sensor Suite Diagnostic Rules Classifier Parameters Diagnostic Results Prognostic Results Alarms,/ Reports Maintainance Actions The Diagnostic Module A Two-Prong Approach High-frequency failure modes (engine stall, etc.): The Wavelet Neural Net Approach Low-frequency events (Temperature, RPM sensor, etc.): The Fuzzy Logic Approach Failure Templates Fuzzify Features Inference Engine Fuzzy Rule Base (1) If symptom A is high & symptom B is low then failure mode is F1 (2) (Defuzzify) Failure Mode Preprocessing in addition to Feature Extraction Sensor Data Features

Concluding Remarks CBM/PHM are relatively new technologies – sufficient historical data are not available CBM benefits currently based on avoided costs Cost of on-board embedded diagnostics primarily associated with computing requirements Advances in prognostic technologies (embedded diagnostics, distributed architectures, etc.) in addition to lower hardware costs (sensors, computing, interfacing, etc.) promise to bring CBM system costs within 1-2% of a typical Army plat as long as m cost

Sholley, Diana Features Reporter

Sholley, Diana is from United States and they belong to Inland Valley Daily Bulletin and they are from  Ontario, United States got related to this Particular Journal. and Sholley, Diana deal with the subjects like Features/Lifestyle

Journal Ratings by Texas A&M University-San Antonio

This Particular Journal got reviewed and rated by Texas A&M University-San Antonio and short form of this particular Institution is TX and gave this Journal an Excellent Rating.