by Rodger Smith, Oracle Utilities
As many utilities complete their advanced metering infrastructure (AMI) deployments and begin to bring more frequent interval data back to the enterprise, that new and increased data flow is feeding transactional applications such as customer information systems, customer care and billing, and operational applications such as network management systems.
An increasing need exists for utilities to build out their business cases for AMI and to demonstrate the initial and continued value and benefits of analyzing and using AMI data—not just collecting it—in customer-facing and operational areas of the business.
Analytics is fundamental to improving and sustaining utilities’ customer connections and business performances. Numerous customer-focused drivers are at play. One of the biggest is the ability to provide more customized, individual service—a more personal and effective relationship with each customer.
Customer Satisfaction, Safety and Personalized Service
Customer data analytics, encompassing structured meter data and more unstructured customer data such as social media and customer relationship data, can provide utilities with a means to better provide customers with information about their usage patterns, target them for new programs that fit their individual needs better, establish new pricing programs based on usage, and implement more effective demand response and demand-side management programs.
Analytics also can provide information to enable utilities to be more proactive with customer service. Using meter consumption data, customer account information and third-party data such as weather, utilities can reduce potential safety risks. Utilities quickly can identify cases of usage spikes and send field crews to investigate, repair and report back.
The following real-life examples show how utilities are using analytics to prevent customer safety hazards:
- Preventing gas leaks by identifying usage spikes. Using analytics, one utility discovered that thieves were entering vacant premises and stealing copper pipes or appliances. In some cases, a pipe would break or the gas would be left on, either of which could have led to massive fires if not addressed quickly. Leveraging sophisticated algorithms, this utility detected these gas leaks while eliminating false positives caused by other causes such as pool heaters, thereby reducing unnecessary truck rolls. Daily monitored tests by the utility detect, on average, 10 cases per year of usage spikes that, according to the utility, would have led to potential public safety hazards within a short time had they not been detected so quickly.
- Proactive bill adjustments after a wildfire. In 2012, an electricity, gas, water and wastewater utility experienced a major wildfire in its service area that necessitated the evacuation of many of its residential customers from their homes. As a fire mitigation tactic, many of these evacuated residents turned on their water hoses and sprinkler systems before they left. Analytics enabled the utility to identify each of these customers proactively and reduce the excess, or fire preventative, water usage from their bills.
Utilities are using analytics to fully operationalize their meter-to-bill processes, redesign their billing exception queues to reduce false positive exceptions, identify new anomalies previously missed, reduce break-to-fix times, and automate many exceptions to bypass manual processes and go directly to work orders.
Also, utilities are using meter data to target demand response and efficiency programs appropriately to customers. Being able to target the right programs to the right customers dramatically increases the expected uptake on each program and the savings results enabled for customers and utilities. Being able to target high-potential customers accurately is the key to a successful demand response or efficiency program. Analytics provides that key by increasing accuracy of the targeting and reducing the outreach costs to potential customers.
As an example, one utility identified its highest residential gas users—the top 10 percent—across more than 150 segments to offer them energy efficiency program options. This resulted in higher program participation rates and lower marketing costs.
Operationalizing Meter Data Analytics
Utilities can use meter data analytics in many operational areas, too. A few examples include better management of misread meters to reduce truck rolls, the use of meter voltage readings to better manage distribution line voltage, and distribution transformer loading analysis to manage the transformers’ effective life expectancy better.
And ties to customers exist in some of this, as well. Take the previous example of targeted customer energy technology programs. An overloaded feeder in a highly residential area is a recipe for an unexpected outage. With analytics, a utility can approach the problem by aggregating meter-level consumption to a subcircuit level—this confirms the overload—then integrate customer information associated with the meters on the at-risk circuit sections, such as individual load characteristics and third-party data such as building size, and target conservation initiatives with the population directly affected by the overload.
After programs are implemented, utilities can measure each program’s impacts and provide participating customers feedback on their usage. These same data results can be used to support regulatory evaluation of the efficiency and demand response programs, as well as information on actual cost recovery.
Revenue protection is another area in which meter data analytics can show immediate financial returns, and many utilities—especially electric utilities—have looked to the area first in operational analytics to prove the business case.
Electricity thieves have become increasingly agile. But with ever-evolving algorithms designed to detect suspect usage patterns’ covering the full spectrum of theft behaviors, an analytical approa0ch to theft has become as agile at detecting new theft behavior.
Here’s how it works: Analytics tests are used to generate a hot list of meters with suspect usage, including theft behaviors such as foreign meter detection, periodic full and partial bypass and usage on inactive accounts. This list is prioritized and passed along to field operators for inspection and action. The reduction in unaccounted-for usage and lost revenue shows near immediate return for the utility.
Nontechnical line loss, or theft, also is being detected and differentiated by using relative voltage methodology. Here’s how this works: Using analytical algorithms, utilities can identify where errors in their connectivity models occur by measuring voltage correlations across meters on the same transformer.
This is particularly useful in detecting such theft issues as grow operations, which can show up as a bypass profile, a “ghosting” meter, or around-the-clock staggered operations (12-hour cycle grow rooms) that show as a constant-use customer profile.
Meter data analytics has aided outage management, too. In particular, outage prediction analytics is aiding utilities in predicting an outage based on call patterns and AMI outage alarms.
Just the Beginning
There is great value in meter data, far beyond what was anticipated when smart meters first were deployed. Using analytics to increase this value can mean increased customer satisfaction with reduced outages, better customer communications, enhanced network management system analytics performance, more efficient monitoring of existing assets, the ability to better meet regulatory reporting on requirements for efficiency and effectiveness and more.
Rodger Smith is senior vice president and general manager of Oracle Utilities.
POWERGRID International Articles Archives
View Power Generation Articles on PennEnergy.com