Utilimetrics Quarterly — Utilimetrics Fall 2012
Change Language:
Tackling OMS And AMI Integration
Terry Nielson And John Wambaugh

Understanding the strengths and weaknesses of AMI can lead to a successful OMS-AMI integration.

Since implementation of the earliest Automated Meter Reading (AMR) systems over 15 years ago, utilities have tried to integrate their Outage Management Systems (OMS) with their AMR and newer AMI (Automated Metering Infrastructure) systems.

These projects had varying degrees of success, but most frequently they ended up falling short of their goals because the integration had not accounted for the strengths and weaknesses of AMI.

Outage Management without AMI

The core of an OMS is an outage location prediction engine that accepts customer calls as input and then utilizes some heuristic rules that attempt to predict the location of the open point in the electrical circuit. The open point might be a blown transformer fuse, a downed service wire, a lateral fuse, or a circuit breaker that opened under fault. To make these predictions, the OMS must know which customers reported outages and how the customers are fed from the transformer and connected upstream.

Under most circumstances, customers are good at reporting outages, and only a small percentage need to call in for the system to make reasonable out ageclearing device predictions. Customer calls are valuable in helping the OMS predict which device opened. They can also provide valuable troubleshooting information, especially calls that include descriptions of key information like the locations of sparking wires, smoke and even load bangs.

Outage Management with AMI?

Since most AMI products provide a capability for meters that experience a complete or partial loss of power to report back, it might seem that the obvious opportunity for improvement in outage management would be to attempt to replace the calls from customers with meter outage reports and eliminate the need for customer calls.

Many utilities claim that once the AMI system is in place and connected to the OMS, there is no need for customers to call in. Unfortunately, once the details of the solution are studied closer, this turns out to be a bad idea for several reasons. First, the realities of the AMI infrastructure and integration usually mean outage reports will take at least several minutes to trickle up to the OMS.

Second, many of the outage reports won’t make it to the OMS due to data collisions that occur when a large number of meters lose power at the same time and AMI communication is not 100 percent reliable. Finally, meters don’t report other clues such as smoke, sparks and noises.

It is unrealistic to expect the AMI to OMS integration to eliminate the need for customers to call in, or expect the AMI system to always provide an accurate outage prediction before the first customer calls because there are always a handful of customers who are quick to call when they have an outage.

So if the replacement of outage calls with AMI outage reports is unrealistic, can the AMI to OMS integration supplement the customer calls and provide quicker and better outage predictions? The short answer is: “Yes, but...”

After deploying AMI to OMS integration and running with it for a period of time, early adopter utilities found that a small but significant percentage of the outage reports turned out to be wrong.

To the OMS end user, the outage predictions made based upon AMI outage reports can be less reliable because they are not validated and can be triggered by other field conditions such as maintenance and construction, meter tampering, power sags and dips, and even faulty meters.

These problems make it seem to the users that customer outage calls, by comparison, are a much better way to predict outages. However, some early adopter utilities reported an improvement in reliably identifying power outages of up to 15 minutes using properly filtered and validated AMI outage reports in addition to customer calls.

By digging into their outage management records, utilities will find they have a large number of “OK on arrivals.” These are situations where the crew arrives on site, typically to a single customer outage and find there is no problem associated with the utility service. It is often a problem associated with the customer’s service behind the meter.

Traditional attempts to address this problem include creating a script for customer service representatives that instructs customers to check breakers, confirm that power is out for the entire premise, etc. This may improve the situation, but it doesn’t eliminate all “OK on arrivals” and it increases the per-call cost of answering outage calls.

A successful approach implemented by utilities with AMI is to have the call taker utilize the capability of the AMI system to “ping” the meter and to see if there is response. If there is a response from the meter, then it can be safely assumed that the service is powered and this is a customer premise problem. Several utilities have reported a significant reduction in truck rolls with this process.

During a power outage where there are numerous, widespread outages, it is common to have multiple areas of damage with one area smaller, or nested within the other outage. Typically the restoration crew will address the larger outage and not know about the nested outage. In these situations, the crew, control center and OMS think all customers have been restored.

The customers who are part of the nested outage won’t get restored until the utility learns the area is still without power. Common practices to address this issue include contacting customers to confirm power has been restored and having crews patrol all of the circuit downstream from the location of restoration.

This practice may result in poor customer perception of the restoration performance because it appears the utility does not know the outage situation. This is another area where AMI to OMS integration can help. Immediately upon restoration, the AMI system can be used to “callback” the meters that were just restored by pinging them. Any that indicate that they are not restored can be used to create a new outage within the OMS.

AMI OMS Integration with an Intermediary

Using AMI to improve outage predictions and outage response times is challenging but not impossible. Similar to the need for validation and filtering of meter reads provided by AMI for the purposes of billing and presentation, the power outage and restoration alarms from AMI must also be fully validated and filtered to be effectively used by an OMS.

The approach that has been successful in some newer AMI to OMS integrations has been to add an intermediary that does some processing between the OMS and the AMI, without significantly slowing the “real-time” nature of these AMI alerts. There are several options for the intermediary:

Enhanced Meter Data Management System (MDMS) functionality

Complex Event Processor (CEP) on top of middleware ESB (Enterprise Service Bus)

Enhanced OMS functionality

In all cases, the intermediary performs functions to filter the AMI outage reports to improve the accuracy and reliability of the outage reports. Filtering can look for outage reports that are not sustained outages (such as a momentary outage), outage reports that are stale, and even duplicate messages.

Filtering can be conducted to look for meters likely to be out due to maintenance or construction activity and not unplanned events. The intermediary can also perform functions such as event throttling, translation of AMI specific identifiers, and standardize and consolidate information from multiple AMI regions and technologies.

The successful integration of AMI with OMS includes utilization of the two-way nature of AMI by verifying power status. This request must account for the reality that AMI can only report the positive (such as that a certain service point has power) and not the negative, as the lack of response can be due to lack of power, failed meter, or failed communication. The intermediary provides advanced capability to support this request response character of the AMI using similar translation and consolidation features used to enhance and manage the outage reports.

Terry Nielsen is a senior vice president for UISOL. He has 25 years’ experience related to OMS and distribution management system products and successful implementations of those technologies while at CES International, SPL World Group, Oracle and UISOL. He has been involved in the definition and delivery of OMS and DMS software to over 20 utilities worldwide, ranging from small utilities that serve less than 50,000 customers to those that serve over 2 million customers. Nielsen has extensive experience working with utilities to improve the outage management processes, major event and storm restoration practices, damage assessment processes and both inbound and outbound multi-channel customer communications.

John O. Wambaugh is a senior vice president for UISOL and a member of the Utilimetrics board of directors. He has 25 years’ experience in AMI and MDM leading the development and implementation of these technologies at companies such as eMeter, CellNet, and Schlumberger. He has conducted numerous courses on the implementation and operation of AMI and MDM. He has played key roles in multiple AMI implementation projects, including projects that have involved installation of more than 13 million electric, gas and water meters at rates in excess of 5,000 meters per day.