Pamela D. McNeil, DTE Energy-MichCon, Detroit, Michigan;

and Heathcliff Howland, OSIsoft, Inc., San Leandro, California

The physical-energy infrastructure enables natural gas flow from source to consumer. A software infrastructure for managing time-series data and events enables information delivery from a pipeline company’s asset base to its business systems.

In an era of increasing compliance complexity and pressure to increase reliability and throughput, gas pipeline companies are bridging the gap that exists between process automation systems and business decision and reporting systems.

Michigan Consolidated Gas Co. (MichCon) did so by mastering the collection, storage, presentation, and analysis of high-fidelity fleet-performance data through application of massively scalable industrial data management software.

This entailed the 2008 deployment of enterprise-class historian software at one MichCon compressor station and at their Gas Control headquarters. The software continuously records real-time operating data from nearly every piece of station equipment, including compressor engines and auxiliaries; dehy; and flow and gas-quality measurement systems. It also records operating data from supervisory control and data acquisition (SCADA) and other critical pipeline applications at Gas Control headquarters. Data is accessed using desktop and web-based tools to drive energy management, asset optimization, reliability, and compliance initiatives.

The complexity and importance of environmental health and safety (EH&S) compliance and a shrinking pool of industry talent challenge today’s natural gas pipeline transmission, storage, and distribution companies to develop innovative strategies for optimizing pipeline operations. As a result, many pipeline companies are launching system- and fleet-wide performance optimization programs as they define their vision of “the pipeline of the future.” This vision typically includes initiatives like performance centers, compressor-engine lifecycle optimization, LAUF reduction, real-time carbon accounting and energy management.

Success hinges on a common system requirement – managing the collection, storage, and flow of high-resolution equipment and system operating data. Innovative pipeline companies find they cannot afford to let operating data simply scroll out of automation and control systems. They are investing in “operations information supply infrastructure” to connect user and business applications with the equipment and system operating insight needed for competitive advantage.

Based on the success of such a system within the fossil-power generation and electric-distribution operations divisions of MichCon’s parent company DTE Energy, MichCon launched a technology evaluation program to assess the impact this type of infrastructure would have on gas transmission and storage operations. It started by asking one question: If SCADA, measurement, station equipment process, analysis, modeling, and forecasting data were saved and made available as either high-resolution data history, “role-based” information views, KPIs, or notifications across the organization, what gets done better? In other words, what would be the benefit to the business?

MichCon chose the PI System from OSIsoft, Inc. as the software infrastructure for managing collection, storage, and presentation of system and equipment operating data. The company gave particular attention to identifying how it could leverage the PI Infrastructure to address the following strategic objectives:

  • Increase system reliability
  • Maximize system capacity
  • Increase asset efficiency
  • Ensure regulatory compliance with less overhead
  • Reduce IT total cost of ownership (TCO).

MichCon’s evaluation of PI involved a six-month field acceptance test (FAT). Implementation was in three phases. Phase I included installation of a plant PI System at MichCon’s Belle River Station; Phase II installation of a SCADA PI System at MichCon’s gas control headquarters; and Phase III installation of software to push data from the station and gas control PI Systems to MichCon’s SAP portal. Thus, operations data could be viewed by users via the web — in the form of summary displays, trends, KPI scorecards, dashboards, and tables — alongside other important information like maintenance records. Implementation of each phase took four to five days.

Facing challenges

In general, the challenges facing pipeline companies fall into four categories: managing fuel consumption and LAUF; managing fleet reliability in the absence of traditional OEMs; reducing health, safety, security, and environmental compliance risk; and workforce constraints.

By looking at publicly available financial and operations information that FERC requires interstate natural gas pipeline companies to file, the magnitude of these challenges is evident. For example, expense-line items in FERC Form 2 reports for the top 20 interstate pipeline companies (by total gas deliveries) shows that in 2006:

  • Dollars spent on the expense item “gas for compressor station fuel” were an order of magnitude more than any other single O&M expense.
  • Dollars spent on “gas for compressor station fuel” ranged from a low of 43% to a high of 64% of total O&M expenditures.

For the larger pipeline companies, fuel cost can be as high as $200 MM/year. Some companies pass this cost on to the shipper, but some are looking at voluntarily going off “fuel-tracker” agreements in an effort to become more competitive.

A look at reported LAUF gas is equally insightful. The market value of LAUF on an annualized basis can amount to tens of millions of dollars for many companies. Reducing LAUF is becoming a focus for most transmission, storage, and distribution companies. Again this is due primarily to the increase in gas cost – but it is also tied to GHG compliance risk around fugitive emissions.

Increasingly, pipeline companies are running compressor-engine fleets over a wider operating range to deliver gas into markets that fluctuate like never before. For some companies this impacts machine reliability. With most of these engines decades old and the OEMs no longer existing, the problem of maintaining fleet reliability is becoming a critical one.

In its 2007 study, “Securing our future: developing the next workforce, an analysis of risk and recommended strategies for the natural gas pipeline industry,” INGAA reports:

“More than half of managers report that present skill development for technical employees is inadequate. Seventy-eight percent of managers expect workforce issues to have a significant impact on the company capability within five years, with loss of knowledge mentioned most frequently. Some report impending retirements of ‘irreplaceable’ employees. Half of managers feel their companies’ knowledge management processes are insufficient to capture knowledge from retiring employees.”

Operational data visibility

Data is a measured process variable derived from an integrated instrument source, or manually acquired via operator effort. The data format is tag (point name), measurement time stamp, and measurement value. This information is acquired as frequently as the instrumentation can present new time-stamped values. Collection of data in real time, storage of historical values, and provision of a uniform, easy-to-use means of investigating data is fundamental to creation of operational data visibility.

Operational visibility is a critical step in addressing the industry challenges described above. For many pipeline companies, MichCon included, the challenge is getting a single version of continuous, quality data having sufficient resolution onto the corporate network and into the hands of specialists. To do this reliably, scalably, securely, and in a way that minimizes IT total cost of ownership (TCO) requires an enterprise data infrastructure.

The fundamental purpose is to connect every person and application within an organization’s value chain to information needed to continuously improve asset performance. MichCon considered the following core capabilities when choosing the PI System:

  1. Connect to and automatically collect equipment operating data from every automation, control, and instrument system used in the field, at gas control, or on the web.
  2. Efficiently process data streams as they are collected so that hundreds of thousands of data “snapshots” can be archived every second.
  3. Efficiently process requests for data from users and applications so that hundreds of thousands of archived data points can be retrieved every second.
  4. Scale to handle millions of tags worth of data streams.
  5. Maintain decades of data history on-line for ready access via desktop and web-based client tools.
  6. Connect to critical business applications like ERP, CMMS, forecasting, simulation, and portals.
  7. Deploy and derive benefits from software without requiring development expertise or custom coding.
  8. Include tools to support analysis and notifications, turning massive amounts of collected data into information for delivery “by exception” to users and applications.

Project description

During Phase I of the project a plant PI System was installed at MichCon’s Belle River natural gas storage facility. The client-server system comprises three key software components running on computers at the station: the PI Interfaces, PI Server, and PI Clients. Belle River station’s Process Control Network (PCN) uses Allen-Bradley programmable controllers (PLCs) and RSLinx for data access. PI-OPC interface software was installed on a single workstation-class PC on the PCN to connect to RSLinx and acquire data from 12,000 station I/O points.

Station data collected in real-time includes all reciprocating and turbine compressor engine, auxiliary machinery, and dehydration system parameters and diagnostics; yard valve status; and pressure, temperature, and flow points. Points are mapped to PI tags and real-time values are archived as frequently as required for accurate data representation on the PI Server.

PI Server software was installed on a single server-class computer on the CDN at the station and securely networked to the PI Interface computer to enable data flow. To allow users at the station to access real-time and historical data on the PI Server, PI Client software was installed on as many engineering, operations, and management desktops as possible. Client software included PI ProcessBook for trend and display building and PI DataLink – the Microsoft Excel Add-In – for analysis and reporting. Additional PI Interfaces are being added to connect to data sources not integrated with station automation. These include station flow-measurement remote terminal units (RTUs) for gas measurement station “send-out” flow rates, pressure, temperature, and quality, and to portable Windrock compressor engine analyzers.

Phase II of the project included installation of a central PI System at MichCon’s Gas Control Headquarters. The purpose was to provide:

  • A central repository for data from SCADA and other critical pipeline applications — including simulation, forecasting, nominations, etc. — to support engineering, operations, and management groups.
  • A central repository for critical station data replicated from the Belle River station PI Server. Ultimately each station will have a dedicated PI Server. Replicating the most important station data on the central server provides “one-stop shopping” for critical equipment data and enables easier configuration of a centralized performance center.
  • A central connection point for routing equipment performance data into MichCon’s business portal (SAP) and other business software systems.

Phase III of the project included installation and configuration of OSIsoft software to link and correlate PI and SAP data that can then be displayed in MichCon’s SAP Portal. Merging pipeline equipment and system operating data with the business system is a powerful tool that can underpin far-reaching corporate initiatives like enterprise energy consumption management.

Implementation and “as-left” specifics

The Belle River PI System was installed in four days in April 2008. Work included installation and configuration of the PI Interface, PI Server and PI Clients on station computers across two isolated networks. Upon completion, values from 12,000 equipment I/O points were being collected at sample rates configured by MichCon ranging from 1 second for faster points, like turbocharger RPM, to 60 seconds for slower-moving station temperature and pressure measurement points. PI data is accessible at the station and to users on the corporate network. During implementation, sample reports and displays were built as ‘go-bys’ and personnel received introductory training. A short time later, personnel used PI Client tools to build production reports, displays, and equipment performance KPIs.

Installation of the PI System at MichCon Gas Control took four days in late April. Work included implementation of the PI Interface to MichCon's Telvent OASyS SCADA system, the PI Server on MichCon’s corporate network, and the PI Client software on PCs used by Gas Control, Gas Planning, and other departments. 3,600 OASyS analog tags, 1,500 digital tags, and 300 rate tags were mapped to PI and configured to collect data at sample rates ranging from one second to 60 seconds. SCADA data history and real-time values are now accessible from the desktops of every user capable of connecting to the PI Server on the MichCon corporate network.

Training was conducted so that MichCon personnel could use SCADA data in the PI System and also take over system-management tasks. Following this baseline installation, MichCon personnel installed an additional PI Interface to pull distribution-system pressure data from the Access database front-end of the Mercury RTU system into PI. Tools were configured to streamline porting of distribution system pressures and SCADA output to the system modeling application used by the Gas Planning department.

Phase III installation, of OSIsoft's Business Package for SAP Portal, called iViews, was performed remotely over the course of three days in July 2008.

Project results

MichCon’s objective was to promote continuous operations improvement by capturing instantaneous operating data made available by instrumentation and control systems. MichCon believes that giving the organization access to instantaneous and historical operating data promotes a more proactive operating mode. As one MichCon employee put it in a recent project review session, “It’s hard to believe that in the 21st Century the only way to know what is going on is to go into the control room.”

Some general concepts for driving operations improvement using fleet-data visibility include as follows.

  1. Proactive management of asset performance:
    • Measure and compare actual performance against baseline as defined by equipment nameplate, warranty, performance testing, or a dynamic fleet average.
    • Prospect high fidelity operating data history for root-cause analysis, discovering fleet performance trends and identifying opportunities for fleet optimization.
    • Have assets inform the business when their performance can be improved by pushing KPI’s into business systems automatically.
  2. Improve equipment reliability by:
    • Shift from calendar-based strategy to condition-based maintenance strategy.
    • Use role-based dashboards and notifications for early warning of likely failure.
  3. Implement “performance center” program based on what DTE Energy created for fossil generation and electric operations.
  4. Reduce fleet “cost of compression” by optimizing compressor engine utilization.
    • Role-based dashboards and notifications for early warning of “out-of-envelope” operation.
    • Deploy consistent asset performance measures fleet and system wide.
  5. Reduce regulatory compliance cost by optimizing administration of engine emissions data.
  6. Streamline and automate critical business processes that currently rely on manual setup, data entry, distribution, and data archiving.
  7. Capture, store and enable access to important business data currently being calculated and “stranded” in spreadsheets and critical applications.

Since MichCon’s compressor-engine control and station-level SCADA systems were not designed to continuously record or transport high-resolution equipment operating data, PI is used to automatically collect a high-fidelity continuous record of compressor-engine performance data. This automatic process data mapping is the foundation for performance optimization of the MichCon compressor engine fleet.

For example, each engine has ideal performance curves over a range of loads for “best fuel.” MichCon intends to use PI to continuously compare actual engine operation against best fuel targets and KPIs that indicate deviation from plan. The difference between actual and plan represents fuel savings potential. The same concept applies to health monitoring of other engine subsystems and control loops. The goal is to measure the real-time cost of these deviations against plan for the entire fleet, displaying the KPIs inside web-based “scorecards” on the SAP Portal.

Typical pipeline capital improvement projects include purchase of new horsepower, retrofit engine controls, and dehydration systems. If this equipment is being provided “turnkey” by a team of contract service vendors and OEMs, performance guarantees and warranty restrictions initialize after commissioning. MichCon intends to use PI as a start-up and commissioning tool and for monitoring equipment performance against guarantees. Access to high-resolution historical operating data is a powerful tool for negotiating warranty claims, OEM service contract invoices, and issues involving poor performance against guarantees. During commissioning, PI will expedite start-up troubleshooting related to new installation “teething” issues.

Condition-based maintenance is a critical component of MichCon’s continuous improvement vision. One challenge for MichCon in moving from scheduled maintenance to condition-based was getting equipment data to flow into a central system so that predictive analysis algorithms could be applied. MichCon plans to configure analysis rules using the PI System’s Advanced Calculation Engine and display the results on maintenance dashboard portal pages on the MichCon SAP portal. Analysis rules will trigger PI Notifications so that personnel are informed of equipment condition on an exception basis.

Acknowledgment

Based on a paper presented at the Gas Machinery Conference, Oct. 6-8, 2008, held in Albuquerque, New Mexico.