Why Your Grid Is Already Smart (and How to Unlock its Data Treasure)

by John D. McDonald

Editor’s note: When does an embarrassment of riches turn into just an embarrassment? When utilities have a wealth of data and don’t take advantage of it. That’s the situation facing most utilities as the Smart Grid goes mainstream. The upcoming GridWeek Conference will feature a panel on this topic including John McDonald, GM Marketing of GE Energy T&D and an IEEE Fellow with 35 years of experience. Because this topic is so timely and so important, we asked John to provide an overview of the challenge and the best path forward. Scroll to the bottom for a link to the conference and John’s session.

You’re not truly wealthy if you have a hoard of money locked up in a vault with no key. Likewise, utilities are not truly empowered if they have hoards of data locked up inside their intelligent devices.

Every utility that has microprocessor-based devices and supervisory control and data acquisition (SCADA) technology has access to an incredibly detailed historical record. That operational and nonoperational data describes virtually every function and event occurring in the generation, transmission, and distribution systems.

Surprisingly, few utilities take advantage of this wealth of information. Why not? Many utilities don’t realize that (a) the data can be archived and (b) technology is already available to utilize that data. As a result, many utilities are short-changing their investment returns by failing to fully tap into the information from automated components and delivering it to decision makers.

To be sure, integrating disparate data sets is no simple matter. Although today’s SCADA systems use standardized protocols, the intelligent electronic devices (IEDs) in other automated systems use proprietary ASCII commands. You need a customized solution to communicate with each manufacturer’s IED.

Introducing the Data Mart

What is needed is a data mart – a server (or group of servers) to retrieve the data. Most automated systems have built-in archives known as historians. These local historians — which are linked to systems such as SCADA, substation automation, power plant distributed control systems, maintenance man­agement, outage management, and customer information sys­tems (see the accompanying figure) —are usually set up to store the data, but not to send it out.

The data mart solves that problem for a wide variety of devices. It accesses, stores, and integrates the data, processes it into usable information, and serves it up to applications and personnel in engineering, operations, and maintenance.

Envisioning Your Data Mart

You gain a tremendous advantage by including data mart requirements in the initial substation automation design. But you can still tap into a data mart’s power even if your automation is already completed. Despite the complexities, most vendors can retrofit their components with local data marts and cus­tomize them to provide data access to a larger corporate data mart.

Whatever your stage, keep in mind four key principles:

1.   Understand the three levels of integration and automation

2.   Establish three data paths

3.   Match users and data

4.   Implement the user-data matrix

1. Understand the three levels of integration and automation

Any automation project has three levels of integration and automation:

·       Implementation of IEDs

·       Integration of IEDs

·       Substation automation applications

Level 1: Implementation of IEDs. IEDs, which can monitor system conditions and provide both operational and nonoperational data, are of paramount importance to the data mart because they are found in nearly every piece of power system equipment. A utility might purchase and install IEDs in any or all of the following:

·       Protective relays

·       Voltage regulators

·       Transformer temperature monitors

·       Transformer tap position monitors

·       Transformer-dissolved gas analyzers.

Level 2: IED Integration. When it comes to IED integration, the main ingredient for success is to integrate both operational data (instantaneous values of voltage, current, and other data) and nonoperational data (logs of events and oscillography). Nonoperational data provide extremely valuable information that enables operations and engineering personnel to piece together the individual occurrences or conditions that led up to major events, such as outages or equipment failures.

Full IED integration means establishing integration data paths for both operational and nonoperational data. For example, when replacing electromechanical devices with IEDs, the IEDs can be integrated so that the operational data flows to the SCADA master and the nonoperational data to the corporate data mart.

Level 3: Substation Automation Applications. The third level is substation automation. This involves the deployment of substation and feeder operating functions and applications including SCADA, alarm processing, automatic load restoration, and volt/var control. The most common is SCADA, which primarily monitors operational data points.

2. Establish Three Data Paths

The second principle is establishing three data paths into and out of the substation:

·       Operational data

·       Nonoperational data

·       Remote access

Unless all three paths can be established, there is no point in implementing the corporate data mart. Without all three data streams, only a limited number of users can be served. Even then, they will be receiving a selective and incomplete picture of conditions and events.

Operational data path. The operational, data path lies between the substation integration and automation systems and the SCADA system. The SCADA can be programmed to scan automated devices in the substation every few seconds to provide a continuous feed of voltage, current, and other data.

Two factors must be considered: First, substation integration and automation systems must have the capability to interface with older SCADA systems and their proprietary protocols. Second, the communications infrastructure must have enough the bandwidth.

Nonoperational data path. The second data path, gaining access to the IED’s nonoperational data, is more of a challenge. Each different device typically operates with a different protocol. The data on this path is on-demand and nonperiodic, which complicates protocol issues. Depending on user requirements, nonoperational data can either be pushed by the substation or pulled by the data mart.

Remote access. The third path is remote access, which allows users outside the substation to access the IEDs. With proper security and access privileges, user can review device settings or change parameters, as well as downloading nonoperational IED data for analysis.

Often called pass-through or loop-through, this communication path is typically a dial-up phone line or dedicated fiber-optic connection. The user dials into a secure modem, which then calls the user back if his or her phone number is approved. The user then dials a code to specify which device the communications link should be established with. Data flow between the caller and the device is two way.

3. Match Data and Users

In addition to careful design of the technical architecture, successful data mart implementation has a third critical principle: identifying the end users and defining their data needs.

Unfortunately no industry standard approach has ever been developed, and the task is more difficult than might be expected.  However, the following process has worked well for at least one large utility:

1. Identify personnel who need information (decisionmakers).
2. Group users into informal clusters based on job function (not department).
3. Interview individuals to learn the type, frequency, accuracy, timeliness, and format of data they need.
4. Devise a matrix matching user groups to data requirements and describing how data is process and applied.

4. Put the Matrix into Action

Once you have a data-use matrix, you can move to the fourth and final principle: programming the system to connect users with the information they need. One method is a dashboard-type interface for each user group. You can configure the data mart to access the necessary files, integrate them into specified combinations, and deliver the output to the right user via the WAN.

A crucial step is determining which data the mart can process internally versus which data it should simply pass along to third-party applica­tions for processing. Not all data sets can be integrated and processed inside the data mart. You may new add-on applica­tion programs to output useful information.

Naturally, data security will be an important requirement of any system. Make data accessible on an as-needed basis only. Establish strict data security controls to ensure reliability and accuracy.  Establish a data mart firewall and multiple levels of user access with usernames and passwords appropriate for the user’s assignment.  Require data mart administrators to perform security audits and archive the data mart information on a regular basis.

A well-designed data mart can provide a wealth of uses.  Consider just a few examples:

• Automating maintenance: Maintenance work orders can be automatically generated based on equipment condition using actual parameters from the plant.
• Emissions monitoring: The warehouse can be linked to a continuous emission monitoring system, letting users access stored emissions data or generate emissions compliance reports.
• Data mining: Users can be set up with data mining tools to analyze historical data and guide them in making decisions to cut operating costs, reduce outages, and improve plant asset optimization.

When accurate, timely information regarding the performance of systems and equipment is avail­able to personnel throughout the enterprise, everyone starts making better decisions. In addition, the utility is able to maintain assets more effectively by planning equipment upgrades and realizing longer life spans for aging components. The data mart also eliminates duplication of data residing in multiple databases in utility users’ personal computers.

Utilities have a treasure trove of data in their intelligent devices. The data mart is the key to unlocking its value.

John McDonald is the General Manager of T&D Marketing for GE Energy. He is also an IEEE Fellow and the co-author of several seminal books on power engineering. It is his ability to look at both sides – the market side and the technology side – that caused us to recruit him as a contributor to SGN. This article was adapted from “Realizing the Power of Data Marts,” which he coauthored for the May/June 2007 issue of IEEE Power & Energy Magazine.

   Email John D. McDonald

   GridWeek conference site