Condition-based Maintenance Regional Seminar Series Condition-based Maintenance Angela Rothweiler, PSE&G Gopal GopalKrishnan, OSIsoft July 17, 2009

Agenda What is Condition-based Maintenance (CBM) Get started today - Cytec use case Maintenance strategies you can validate before deployment Usage based – DTE Energy use case Condition-based - Basin Electric use case PSE&G use case - fully deployed and live since 2002 Sempra Energy use case – fully real-time enabled (2008) PI enabled maintenance for the enterprise Q & A

Agenda What is Condition-based Maintenance (CBM) Get started today - Cytec use case Maintenance strategies you can validate before deployment Usage based – DTE Energy use case Condition-based - Basin Electric use case PSE&G use case - fully deployed and live since 2002 Sempra Energy use case – fully real-time enabled (2008) PI enabled maintenance for the enterprise Q & A

What is Condition-based Maintenance Dept. of Defense quote “… CBM be implemented to improve maintenance agility and responsiveness, increase operational availability, and reduce life cycle total cost …” The challenging work in a CBM deployment is in the condition assessment, i.e. the analytics JIT1 maintenance2 based on monitored asset3 conditions WHAT Operating parameters (PI data) vibration, thermography, dissolved gas analysis… WHY Prevent failure, improve operations, reduce maintenance expenses… WHEN Depends on how it changes or degrades WHERE Critical assets WHO Both operations and maintenance HOW PI enabled operations & PI enabled maintenance 1JIT = just -in-time 2maintenance is not repair 3 include control loops

Agenda What is Condition-based Maintenance (CBM) Get started today - Cytec use case Maintenance strategies you can validate before deployment Usage based – DTE Energy use case Condition-based - Basin Electric use case PSE&G use case - fully deployed and live since 2002 Sempra Energy use case – fully real-time enabled (2008) PI enabled maintenance for the enterprise Q & A

Get started today Cytec Critical Variables Program ProcessBook multi-state symbol Etiam dignissim nibh. Ac quam adipiscing iaculis. Nam malesuada consequat neque.

Get started today Cytec Critical Variables Program - DataLink reports Targets met 95% percent of the time vs. 67% previously Steam savings $1MM plus Equipment is run as intended

Agenda What is Condition-based Maintenance (CBM) Get started today – Cytec use case Maintenance strategies you can validate before deployment Usage based – DTE Energy use case Condition-based - Basin Electric use case PSE&G use case - fully deployed and live since 2002 Sempra Energy use case – fully real-time enabled (2008) PI enabled maintenance for the enterprise Q & A

Types of maintenance Maintenance Corrective - this is after failure Preventive - this is before failure Calendar based - ignores equipment utilization pattern Condition based – aware of equipment conditions and utilization Usage based (via PI and analytics) Equipment condition based (via PI and analytics)

Service water pump – calendar based schedule ignores equipment usage pattern Pumps were off for an extended period, however the PM WO still went out Save 28 PM hours with usage based PM instead of calendar based

Usage based maintenance PI totalizer PI time-filtered conditional expressions (time-weighted and event-weighted) Run-hours Coal feed conveyor Pulverizer High pressure service water pumps Run-modes number of starts (peaker combustion turbine (CT) blades) number of trips Run-weight tonnage processed (mining industry) flow-rate (time-integral) converted to volume

Validate usage based strategy - Pulverizer Maintenance: 160 PM hours per pulverizer per year; 16 pulverizers per site, 6 sites Actual runhours: 80% based on PI data, implies a 20% savings Projected savings: 480+ PM hours (approx. $25,000 at $50/hr) per year per site

Validate usage based strategy - Coal conveyor Maintenance: 60 PM hours per conveyor per year; 22 conveyors per site, 6 sites Actual runhours: 25% based on PI data, implies a 75% savings Projected savings: 900+ PM hours (approx. $45,000 at $50/hr) per year per site

Equipment condition based strategies Failure or performance correlated to a slowly degrading metric Temperature Bearings Motor windings Pressure or DeltaP Heat-exchanger plugging Filters Vibration, amplitude (need to interpret along with operations data in PI) Instrument and transmitter drift Control loop health

Secondary air heater plugging Air heater tube plugging causes DeltaP (green line) to increase over several months and is a trigger for maintenance Ad hoc time-filter expressions as PI Calculations

Boiler (convection) tubes plugging Rapid rate of change of Delta P over several days is a trigger for maintenance Delta P (PI expression calculation, ad hoc plot)

Instrument drift – O2 analyzer Based on redundant dual sensors Delta approx. zero is GOOD Delta drifting from zero is NOT GOOD

Transmitter drift Boiler feedwater pump discharge pressure Based on redundant triple transmitters (PressA, PressB and PressC) XY Plot, PressA (X) vs. PressB (Y1), PressC(Y2) GOOD NOT GOOD

Firing rate control loop Boiler exit O2 set point: Approx. 3.2% Actual process value (green line): High variability from 1% to 5.5%

Firing rate control loop At purple crosshair, air (red) peaks when coal (yellow) dips causing O2 (green) to peak after 30-40 secs. At white crosshair, air (red) dips when coal (yellow) peaks causing O2 (green) to fall below 1% after a lag of 30-40secs, and so on…. Screen shows a 15 minute chart (each grid is 1 minute) for the firing rate control parameters – i.e. fuel, airflow, airflow to fuel ratio, and BoilerExitO2. At the purple cross-hair, fuel (yellow) has dipped, but air flow (red) has gone up and hence the air-to-fuel (blue) shows a sharp jump. After about 30-40s lag (width of each grid is 60s), the boiler exit O2 (green) shows a peak. At the white cross-hair, fuel (yellow) has peaked, but air flow (red) has dipped and correspondingly, the air-to-fuel (blue) shows a sharp dip. And, predictably, after about 30-40s lag, the boiler exit O2 (green) shows a sharp dip. This pattern, i.e. the air flow dipping when the fuel peaks and vice-versa, causes the O2 to vary from the lows of about 1% to highs of about 6%, even though the average value stays nominally around 3 to 3.5%. With such high variability, O2 is above 4.5% for about 35% of the time, and below 2.5% for about 30% of the time, during hours of full load i.e. > 200MW.

Agenda What is Condition-based Maintenance (CBM) Get started today - Cytec use case Maintenance strategies you can validate before deployment Usage based – DTE Energy use case Condition-based - Basin Electric use case PSE&G use case - fully deployed and live since 2002 Sempra Energy use case – fully real-time enabled (2008) PI aware Maintenance for the Enterprise Q & A

PSE&G use case

Agenda What is Condition-based Maintenance (CBM) Get started today - Cytec use case Maintenance strategies you can validate before deployment Usage based – DTE Energy use case Condition-based - Basin Electric use case PSE&G use case - fully deployed and live since 2002 Sempra Energy use case – fully real-time enabled (2008) PI enabled maintenance for the enterprise Q & A

How CBM PI System is used Sempra Energy use case Real-time CBM project is the first-of-its-kind in the industry – end-to-end from substation sensors to back-office analytics and visualization Collects and stores time-series data Controls communication from Enterprise level Gateway to PI Provides Analytics (ACE) and Visualizations (WebParts) Visualizes graphing, trending and analytics Maps data points automatically from Enterprise level Gateway Integrates with T&D Operations PI System Integrates with Legacy applications (Phase 2) Sends PI Notifications and Acknowledgement Serves as a CBM Model server

Dissolved Gas Analysis Sempra Energy use case Online DGA can be as frequent as one hour

Agenda What is Condition-based Maintenance (CBM) Get started today - Cytec use case Maintenance strategies you can validate before deployment Usage based – DTE Energy use case Condition-based - Basin Electric use case PSE&G use case - fully deployed and live since 2002 Sempra Utilities use case – fully real-time enabled PI enabled maintenance for the enterprise Q & A

Manual inputs – Operator rounds in PI Equipment inspection data collection specifically designed to help maintenance tasks (data not already in PI) Benefits: Proactive maintenance for increased MTBF (mean-time-between-failure) High HiHi Low LoLo Rate of Change > 3 deg in 7 days Compressor Discharge temperature STEP Source: www.aeec.com/conveyor/Belt_Cleaners/Vplow.aspx (retrieved Jan 2009) Operator Rounds: V-PLOW status on a coal conveyor belt

PI System Gateway to IBM Maximo server PI to Maximo data flow PI System Gateway to IBM Maximo server PI server Data or WorkOrder trigger PI Alerts Maximo Measurements Process Analysis Condition Assessment Meter Acknowledgment Work Order Number and Status. GAUGE CONTINUOUS CHARACERTISTIC Meter Acknowledgment Work Order Number and Status. OPERATING PARAMETERS Running or Idle Pressure Temperature Rpm Vibration Oil level Portal WORK ORDER PI Process Book and Web based RtWebparts Integrated view of Operations and Maintenance EQUIPMENT 28

Common equipment names in PI and Maximo PI Module DB Maximo Equipment Drilldown

Combined Maintenance and Operations Microsoft SharePoint - Browser View - shows the alarm history for one or more equipments and the actual measurements generating the alarm. WO details (history and their statuses) for PUMP1 are also listed. CBM Alarms Equipment listing with CBM Alarm points WO details from Maximo Operations history PI RLINK-Maximo

PI enabled maintenance for the enterprise Hyperlink launches the demo only during a live session Maximo_Asset_Registry and WorkOrder_Demo

Agenda What is Condition-based Maintenance (CBM) Get started today - Cytec use case Maintenance strategies you can validate before deployment Usage based – DTE Energy use case Condition-based - Basin Electric use case PSE&G use case - fully deployed and live since 2002 Sempra Utilities use case – fully real-time enabled PI enabled maintenance for the enterprise Q & A

Other resources PRODUCTS and FEATURES PI Manual Logger for Operator rounds Advanced Calculation Engine PI Notification PI Web parts PI ProcessBook PI DataLink PI TimeFilter, PI Expressions, PI Performance Equations PI integration with IBM Maximo (Maintenance) PI web services (to be released later in 2009) Search for “condition based” on OSIsoft website: http://www.osisoft.com/Templates/search.aspx?searchtext=condition+based

Other resources Using PI to Back - Test Usage and Condition Based Maintenance Strategies to predict Quantifiable Benefits Prior to Deployment in Asset Management (Sebastien Cournoyer, DTE Energy, Laurence Hruby, Basin Electric Power Cooperative, Gopal GopalKrishnan, OSIsoft) – PPT Using Basic PI Tools to Implement a Critical Variables Program that is Inexpensive and Easy to Maintain, Cytec Industries Inc. http://www.osisoft.com/templates/item-abstract.aspx?id=2521&terms=cytec Substation Inspection and Condition Based Maintenance using PI http://www.osisoft.com/templates/item-abstract.aspx?id=2651&terms=PSE&G SDG&E PI-based Substation Real-Time Condition Based Maintenance System (Neal Bartek, Subburaman Sankaran, SDG&E) – PPT