PC-Based Control Comes of Age

Twice in the summer of 1996 the electrical grid in the Western U.S. collapsed and caused a massive loss of productivity. In 2003, the Northeast was confronted by a similar collapse which affected millions of people. In an industry that provides unparalleled levels of system uptime, the magnitude of these system failures shocked many. The 1996 blackouts led to an extensive investigation of the events leading up to the grid collapse. Evaluations of sub-system performances led to engineering design changes to improve system response. As with most complex systems, there was not a single fix that would miraculously correct the problem. Instead, many ideas for improvements were developed, evaluated and implemented.During the evaluations of the incident in the Pacific Northwest region (where the system initially started to collapse on both occasions), it was determined that control system performance at four hydroelectric power generating stations needed improvement.These generating stations are located in areas that are strategically important to the electrical transmission system because of proximity to major transmission system interties. Their location allows these stations to be a first line of defense during an electrical disturbance, but only if the units perform as required.

Control System Upgrades Critical

The generating stations' original control systems were antiquated. Although still performing as originally implemented (supported, as all control systems are, by the efforts of a very talented group of technicians), it was apparent that they would not be a good platform for improving system performance. This was especially true with respect to the integration of new technologies.

Faced with the need for vastly improved levels of performance, a decision was made to implement a new plant-wide control system. Automated Control Systems, Inc. was brought in to evaluate needs and develop the new control systems.

Project goals were clear and comprehensive. The facility's management believed that extensive use of COTS (commercial off the shelf) technology would be the way to extend system life, and provide best-of-breed product use. The utility's engineers insisted on a high level of fault tolerance, and a common code base across the system. The challenge was to evaluate a wide variety of products and design a control system that would meet these goals.

Because of the technology's low cost, high performance, high reliability, freedom to mix and match components, and inherent ability to modify the hardware without a huge price penalty, PC-based controls were considered to be an attractive, cost-effective choice. The larger question was whether or not PC-based controls would provide the required reliability and flexibility.

A Dam Fine Start Up

A system engineer views the control system via HMI screens

during startup at John Day Dam.

The same assessment criteria applied to proprietary components was used to evaluate the PC-based technology. These criteria included:

Does the function and design of the device meet the requirements of the application?

What are the failure states of the solution and how do the failures impact safety and the process?

What is the track-record of the technology in similar applications?

As with any important project, the importance of proper testing and evaluation cannot be stressed enough. Because many control systems look good on paper, a proper test and evaluation plan designed by an experienced professional to highlight product differences is essential.

In the areas of direct equipment and system control (where the controller has the responsibility of making decisions that impact equipment protection or profitability), we needed a platform built from the ground up for mission critical functions.

In supervisory functions such as SCADA, plant resource allocation, historical logging, reporting, and operator interface; we needed to balance reliability with advanced functionality.

Evaluation Results

After extensive evaluation; a PC-based control system, an HMI, and an I/O system from three different vendors were selected. This best-of-breed approach allowed us to pick the leading product in each category.

Controller:

SoftPLC (http://www.softplc.com/) was selected as the controller supplier for a number of reasons. Its control software is compatible with the Allen-Bradley PLC-5 instruction set and communication protocols, it easily and reliably integrates with other vendor technologies, has virtually unlimited program and data capacities, and has an extensible instruction set that allows the user to develop advanced functionality in C/C++ and JAVA. Project personnel used the extensible instruction to implement advanced control calculations and communication protocols to a number of other computing systems.

PLC-Based Control Rejected

Because we needed to program in multiple languages in addition to relay ladder logic in order to implement advanced functions such as a Modbus Master communications PLCs were not considered a viable control technology. While some PLCs offer appropriate communication software functionality, the nature of proprietary hardware can make adding additional communication ports expensive or even impossible.

We also needed to keep the SoftPLC clock synchronized to a satellite clock for certain top-of-the-hour reporting requirements. This extension to the control's instruction set was done in Java, a language not easily accommodated with a PLC.

There are also many other advanced control functions that were not feasible with PLC-based controls. We implemented plant wide control functions including allocation of generating units and reporting on equipment capacity with a C++ control engine that utilized an XML system model to control how the software worked at each site. The XML system modeling has been a great design. Finally, PLC-based control was rejected because the flexibility to use best of breed I/O that would meet strict IEEE surge-withstand criteria was needed.

I/O

MTL IO-95 (http://www.mtl-most.com/) was selected as the I/O for the system. Tests include IEEE-472/ANSI C.37.90.1 in which the I/O is subjected to spikes of several thousand volts. The project has over 30,000 I/O points currently installed, and will eventually have about 38,000 I/O points.

HMI/SCADA

GE-Fanuc's iFix (http://www.gefanucautomation.com/ ) HMI/SCADA was selected because it offered several major advantages over other PC-based SCADA packages, including high-performance integration with custom application programs, a client/server architecture that scales easily from hundreds to tens of thousands of points, and the use of Visual Basic (VBA) as the scripting language.

We used VBA to add right mouse click support on a data point to bring up point-specific information, to allow operators to query the alarm history of the point, and to bring up historical trend information.

Active X graphics objects enabled the addition of a quantity viewer, allowing station operators to navigate through the application logic in an intuitive manner. Easy relational database and Active-X integration let these users leverage current control system investments and now allows for improvements.

The GE-Fanuc real-time database proved to be very flexible and allows for many of the functions necessary in larger HMI/SCADA systems such as the ability to enable/disable alarming on the fly and manually override points.

An Unqualified Sucess

Since its implementation the entire control system has been a huge success. The original scope of the project grew from upgrading the original four hydroelectric facilities to encompass eight more for a total of 12. By the end of 2005 almost 17% of all US hydroelectric capacity will be controlled by this system.

The first control systems installed have been running for almost five years and have surpassed expectations for uptime, usability, flexibility and operational readiness. The operator and technician acceptance of the system has been tremendous.

One of the keys to the project's success was correctly specifying the controller, I/O, and HMI technologies. It sounds simple, but without the right control system components no matter what level of integration effort is applied, it wouldn't have resulted in the high-performance control system that was created for this project.

We feel that our integration expertise brought out the best in the selected products and helped to lead the client towards an optimal solution. But the project would not have been a success without the client's clear vision, domain expertise, and commitment to quality.

Dan Perrier is a system integrator in Washington State. Visit his website www.automation-software.com.