Managing outages: scale and granularity matter
100,000 `last gasps' can overwhelm, veteran says
One veteran of operations and outage management at ConEd in New York told me this week that storm-related outages initiate a classic pattern of response from customers.
The first day, people make the best of the situation, often having "hurricane parties" and such. The second day, the novelty wears off and they become a bit testy. By the third day, the aggravation of an extended outage generates anger. That anger is compounded when customers get inaccurate estimates of restoration time.
This much is not news. But in an era when utilities are trying to become more responsive to their customers—in part by providing estimated times to restoration of service—and because various grid modernization initiatives are sold on the promise of less-frequent, shorter outages, sub-par performance in this area merits attention.
Two recent, related pieces we've done may provide background:
Now that the various systems that detect, analyze and address outages are being transitioned from legacy systems that required manual data inputs and mainframes to modern mash-ups of newer systems, expectations are rising.
As Kevin Hennelly, principal consultant, DNV KEMA, and a former operations manager at ConEd for 38 years, described it to me, two major factors may be involved in under-performing systems.
One is sheer scale. When a big storm hits—such as last fall's one-two punch of Hurricane Irene and an early snowstorm in the Northeast—widespread outages may cause tens or hundreds of thousands of "last gasps" from meters to the advanced metering infrastructure (AMI) system, essentially overwhelming the system's ability to make sense of the data.
The other possible factor is that when legacy systems (GIS, e.g.) get converted to newer, more powerful systems, data errors introduced by manual inputs or a lack of granularity in grid schema can misinform the new system, hampering an effective response. Garbage in, garbage out. For instance, automated fault isolation and switching can be ineffective or counter-productive without granular operation details down to, say, the size, rating, age and installation date of the cables. Lack of granularity in an operational model can lead to sub-optimal performance.
In an age when "smart grid" is sold on higher expectations, particularly around outage restoration, these fundamental issues can significantly impact customer satisfaction, Hennelly told me.
"The question is: How did customers feel about the information they got about restoration times?" Hennelly asked, rhetorically.
In an extended outage, customers are making decisions about whether to stay home or move to family or friends' places where the power is on. People are justifiably concerned about their safety, lost income, spoiling food—a range of issues that matter to them emotionally and financially.
While he acknowledges that "I'm an operations guy, not an engineer," Hennelly made a few points about the current landscape around outage detection, management and restoration notification.
First, the history of such systems has just reached the point of being customer-focused, Hennelly said. Originally, outage management systems were designed for engineering and operations departments at a utility, not to produce timely, accurate outage maps and notifications for customers. Transformer load management (TLM) data provided accurate transformer load data from CIS meter reading data. This valuable information provides engineering the loads on the system, which could guide upgrades and thus help prevent unexpected outages from equipment malfunctions. Unexpected outages, of course, historically and still in most locations, relied on customers calling in to notify a utility that an outage had occurred.
However, TLM did not provide the customer with any information. Now ETR or estimated time of restoration is a requirement in many states. Where utilities once had clear goals for restoration alone, now they are attempting to give customers more timely and accurate information on outages and estimated times to restoration. With expensive new outage management systems (OMS), geographical information systems (GIS), customer information systems (CIS) and mash-ups between them and other data sources, interval meters are expected to self-identify outages, distribution automation is expected to isolate faults and restore service where possible in short order and the right field crews can be dispatched to the right location with the right tools and solutions.
Hennelly said he expected the newer technologies to mature and for vendors and utilities to get better at systems conversion and integration.
"This will be solved," Hennelly said. "It's not like going to Mars. The technology will mature. It takes time to understand how poor quality data affects outcomes.
"But if you want a customer-facing system you have to do the legwork," he concluded. "Get on the right path. Once you understand what's needed, you'll want to fix it."
Intelligent Utility Daily