AmerenUE System Load Snapshot:

Since the massive August 2003 northeast power outage, there has been a renewed emphasis placed on reliable power delivery by utilities and regulatory bodies. Today leading utilities across the country are seamlessly integrating operations, planning, customer service, tariffs and sales with real-time utility information systems to leverage utility infrastructure assets.

The challenge
Although traditional distribution systems were designed with excess capacity built-in, customer growth and limited validation of theoretical planning assumptions are forcing systems to operate at or above capacity. Public opposition to new transmission grids leads to a lengthy siting process, enforcing the need for better understanding and utilization of existing systems and better planning on new expansion on the distribution system. In fact, some utilities have reported seeing some residential consumption increase at rates as high as 4 percent annually. As the cost and demand of adding new transmission and distribution lines keeps growing, utilities are forced to seek innovative ways to reduce or optimize their current grid utilization.

The solution
St. Louis-based AmerenUE serves as a prime example of a utility that has realized the operational value of its AMR-derived data. With over 1.4 million customers and 95,000 distribution transformers under the Landis+Gyr fixed network system, AmerenUE is one of the largest fixed network AMR systems in the United States. While AmerenUE initially justified its fixed network AMR implementation in 1995 on the more traditional approaches of daily, accurate meter reads, management believed they could further use the Landis+Gyr system to provide data that would enhance their distribution operations and planning. Through the use of an innovative System Load Snapshot (SLS) function of the system, AmerenUE is better able to determine whether or not its transformers are being used efficiently. As it happens, the decision to go with a fixed network was correct.

In the past, AmerenUE estimated loading on distribution transformers with a program that was based on a sample of monthly customer use against profile metering at transformers. The methodology incorporated time-dependent load measurements obtained from a small set of profile-metered locations at randomly selected transformers. This profile data was used to establish a correlation between peak demand and monthly use, which in turn was used to estimate peak transformer load. The approach produced conservative (i.e., high) estimates of transformer loading that incorporated significant levels of uncertainty. These shortcomings made it difficult to reliably identify overloaded or underutilized distribution transformers. In today’s era of process improvements, budget optimization, and capital spending limitations, an opportunity exists to require more quantifiable information to make better design decisions.

How does System Load Snapshot work?
Through the use of the Landis+Gyr SLS application, AmerenUE began optimizing their distribution network in 1998. Using historical system peak loading data, AmerenUE engineers select three periods of the day, during summer and winter seasons, to measure when they expect the system peak to occur. The Landis+Gyr fixed network is easily configured for the seasonal time periods to perform additional time-sensitive system-wide load readings. With this system-wide, time synchronized information, AmerenUE is able to measure the actual peak load at each distribution transformer in their system throughout the entire year. As system peak periods change, the time periods of interest are easily modified.

AmerenUE first identifies the day(s) and time they expect the system peak load will occur, though Landis+Gyr can go back a day if the peak occurred before it was expected. Via the SLS application, the Landis+Gyr system is configured to collect and report load data for all the automated meters under the network. The SLS application provides the user with daily customer usage, once captured, and in Ameren’s case, provides three hourly averages of demand data for all consumption meters during system peak.

The SLS application uses unique reading features within the Landis+Gyr system to store and collect the meter demand data for the entire population of automated electric meters. The individual meter demands are then aggregated for use in distribution transformer load monitoring, switching analysis, and circuit loading simulation models. Through further analysis, AmerenUE is able to identify and correct both overloaded and underutilized assets, to optimize the distribution network, and to reduce losses. Instead of relying on a relatively small sample of customer data to determine the load curve for the entire population, the SLS application provides the flexibility to analyze customer loading patterns on a transformer-by-transformer basis in a highly accurate manner. Benefits are not limited to simply identifying overloaded or underutilized assets, but also expand to include performing more realistic load flow analysis through the use of true load data, thereby improving system planning, operations, and expenses.

The results
On July 29, 1999, the maximum temperature recorded in St. Louis was 102°F, creating a record peak load for the AmerenUE control area based on previous assumptions. To validate these loads and assumptions the SLS data was captured and analyzed to generate the Transformer Load Management (TLM) report. The results of the TLM report indicated that, of the entire population of 95K single-phase transformers, 229 transformers exceeded AmerenUE’s design criteria while the median load on all transformers was 58.3 percent.

With hundreds of millions of dollars in distribution transformers, improper transformer utilization can mean a significant and unnecessary cost. The System Load Snapshot program helps AmerenUE avoid such losses, and the utility expects significant cost savings as the program continues. Andrew Sugg from Ameren Services says “We are able to track where every MW on the system went instead of estimating. We are able to get on-peak data for substations and feeders where SCADA cannot be justified.” To initially determine the accuracy of AMR’s data, Ameren engineers compared data from the company’s Supervisory Control and Data Acquisition (SCADA) system with meter data gained from snapshots. AMR data proved far more accurate. In one study, the fixed network-based SLS data differed from SCADA data 8 percent to 23 percent, while numbers from statistical modeling veered 22 percent to 46 percent from SCADA data.

Practical example
Traditional TLM systems use a monthly value since most meter reading systems cannot provide more than a monthly read and predetermined customer diversity values to predict transformer loading. In one case, the traditional model predicted a transformer load in excess of Ameren engineering practices. Prior to the implementation of SLS, that transformer configuration would have been changed by either adding a larger one or by adding a new one. Having multiple snapshots in a given season has also allowed engineers to determine that the actual loading was well under design guidelines and did not require any changes. Being able to view customer loading at more discrete time periods, AmerenUE was able to avoid unnecessary customer interruptions, labor to change the transformer and prematurely changing an operational asset.

Through the analysis of the actual data provided by the SLS application, AmerenUE was able to make comparisons with existing TLM models. Mark Konya, of AmerenUE Services, said that AmerenUE maintained a different transformer load management program prior to initiating the AMR-based System Load Snapshot program. However, the technology used for that previous effort had become outdated and the program was providing less-than-accurate results. “We felt a greatly improved transformer load management program would allow us to accrue significant savings that we otherwise wouldn’t be able to tap into,” he said. In 1999, the year Ameren first used AMR data in transformer load analysis, the company found that of the 336 transformers replaced by AmerenUE the first year the SLS data was collected, 137 transformer replacements could have been avoided. That number of transformers represented savings in the hundreds of thousands. On top of monetary savings, transformer replacements come with a customer service cost, because replacing transformers without cause puts customers out of service needlessly. On the flip side, overloading transformers puts the equipment at a higher risk of failure, which also means hours of interrupted service for utility customers.

As the utility has become more comfortable with the information it receives from the system, more applications for the data have been identified. AmerenUE has made its individual transformer load data available online to field engineering personnel.