Oracle is ramping up investment in analytics-based products and services to facilitate smoother utility operations alongside booming EV adoption.
Oracle recently announced that the company is “test driving” EV analytics with several partner utilities. Oracle takes a two-pronged approach, focusing its EV analytics offerings on grid operations and customer behavior. This Oracle investment emphasizes the importance of analytics across the full spectrum of EV adoption.
A Little Background: Analytics and EVs
Many utilities do not know which customers are charging EVs, whether it be in their home or workplace. According to the Transportation Research Board[1], “…in most cases, customers do not report to their utility their EV purchase; as a result, utilities have no information about any additional load.” Some utilities are using detection analytics to identify EV charging loads to help them better predict adoption patterns.
In the early stage of grid planning for EV adoption, utilities need to forecast EV demand and its impact on the grid; develop cost-effective ways to fortify the grid; and make recommendations on how utility investment will be compensated. A distribution model at the feeder level is needed to accurately determine where and when the grid will feel the charging impact. Existing GIS-based connectivity models are not enough in areas where there are extensive non-utility distributed energy resources (DERs) present, including EVs.
When EV penetration has increased, operational planning requires continuously updated forecasts of charging patterns that are overlaid on grid models. That will help operators to predict where EV impact will occur and adjust for it. An ADMS/DMS automatically calculates distribution overloads to provide operators visibility of trouble areas in order of priority.
Utilities can minimize the impact of EV charging on the grid by providing incentives to customers to charge when power resources are available. For example, SDG&E has EV charging rates that vary according to the time of day and include distinct pricing for on-peak, off-peak, and super off-peak hours.
But, all of this is new for customers. EVs charging and TOU rates are not well understood, and utilities need to educate customers and make it easy for them to participate in these types of programs.
What Oracle’s announcement means for utilities:
- Oracle has trained machine learning (ML) algorithms, applied to Opower’s store of AMI data and combined with customer confirmation of EV ownership, to improve the accuracy of EV detection models.
- The EV analytics leverage Oracle’s core analytics, including capabilities from the DataRaker and Opower acquisitions, utility-specific applications such as Network Management System (NMS Oracle’s ADMS/DMS) and C2M and Customer Cloud Service (Oracle’s CIS options).
- Utilities can use the output from the machine learning models to enable two important functions mentioned above:
- Input to planning for grid congestion. Oracle integrates machine learning with NMS.
- Engaging customers to adopt EV charging rate plans. Oracle uses with Opower web energy management to enable this customer experience. ML models detect EV at the household level. This information can be piped into the CIS so that customer service representatives can promote EV charging rate plans during customer calls.
- Utilities that have Oracle’s most recent NMS platform have access to a locational distribution model required for EV planning that accounts for DERs. [Note: It would be overkill if utilities acquire the platform just for a distribution model for the early stage. NMS is built for dynamic operation of the grid.]
- The real test is how utilities operate at later stages of EV adoption. Utilities will want to look to the pilot outcomes to see how the NMS platform with its DER capabilities plus ML analytics perform in pro-actively managing load.
- According to a 2018 study conducted by Opower, analytics displays need to be clear, concise, data drive and consistent. Utilities will want to follow the pilots to see whether Oracle’s solutions, including Opower’s Behavioral Load Shaping, are successful in helping customers make informed trade-offs between convenience and cost when charging their vehicles.
[1] Challenges and Opportunities for Electric Vehicle Charging Detection Using Utility Energy Consumption Data, December 7, 2018.
