by Bradley Williams
Around the world, technology is opening up new opportunities for utilities to provide higher levels of customer service and grid reliability-including greater generation diversity and individualized customer service.
Empowered by technology innovation and supported by policy, consumers are choosing to adopt distributed energy resources (DERs) in record numbers and far faster than anticipated. By managing their energy consumption behavior with little or no input from utilities, consumers are changing how the modern distribution grid works.
Historically, supervisory control and data acquisition (SCADA) systems have been foundational for utility real-time operations in providing the ability to control the movement of energy around the generation-to-distribution grid. But typically, that monitoring and control stopped at the substations or, in some cases, some of the distribution feeders just beyond the substation. So, the overwhelming growth of DERs-generation outside of the utility’s direct control-is becoming an issue, because to control and optimize the edge of the grid requires utilities to have that visibility or model all the way down to the consumer level. Swings in distributed generation and variations in consumption patterns create information that is vital to the efficient control of the network.
As a result, it is becoming increasingly important to ensure a utility’s advanced distribution management system (ADMS) or network management system has this end-to-end visibility.
Why Grid-edge Visibility is Important
As consumers connect more DERs into the grid, their intermittency creates real risk to reliability and grid health. While utilities can’t control how or when these resources are connected, they can use their ADMS and analytics to reduce the risks they pose.
Here is how advanced DER modeling-across the distribution grid, at every transformer or load point, rather than at the substation level-works: A true load model is derived from aggregated net-metered customer advanced metering infrastructure (AMI) interval reads minus the sum of the customer’s beyond-the-meter DER expected output. Each type of DER has its own generation model or expected output schedule, derived from customer DER records and specific class models or from statistical sampling. Modeling photovoltaic (PV) generation is particularly challenging, because the generation values of a PV system are dependent on many changing factors including latitude, time of day, day of the year, direction and pitch of the solar panels, and current weather conditions.
One example of how this increased, down-to-the-DER visibility is particularly helpful is in grid restoration after a prolonged outage, particularly with regard to cold load pickup.
When a utility re-energizes its distribution system, or a part of it, that has been off-line during an extended outage, it is picking up a “cold load”-one that, when energized, draws more current than it would during normal operation to energize transformers, charge capacitors, start up cold motors, heat up light bulb filaments and the like. This current draw, during a cold-load pickup, can be so extensive that it can, if not carefully managed, trip a new outage.
Many factors-such as outage duration, types of connected load, weather, restoration mode, outage causes, the presence of distributed generation or automatic transfer schemes, time of day and load level-will have a bearing on the extent and duration of cold load pickup.
The advent of DERs has rendered ineffective the traditional, substation-based cold-load pickup algorithms where calculations of inrush are substation-centric and don’t account for the impact of directly connected generation on the system. What this means is that the load inrush current from DER isn’t accounted for, and so the power flow can be significantly higher than anticipated. Without proactive management of DERs for cold load pickup, utilities risk having damaged and degraded equipment, compromises in staged processes, system losses and network design flaws.
A New Approach
As previously mentioned, this unaccounted-for inrush from DERs means traditional inrush calculations can create “built-in” network inaccuracies-if what goes into the network model is flawed, what comes out will be flawed, as well. In a severe example, this poses the potential for an enormous blackout as a result of re-energizing an incompletely modeled network after an outage.
With true grid-edge visibility for the network operator, it is possible to model and manage this more effectively by creating granular DER models down to the customer level. The workflow for our DER models is as follows:
The utility’s customer service agent receives the customer’s request to connect their DER and logs this request, typically within the utility’s customer information system (CIS) per their established workflows.
The utility will record the attributes of the customer’s DER in an asset register within the CIS or DERMS to update the customer record. These attributes should include the DER model’s technology (e.g., rooftopo PV), capacity (e.g., 5 kW), and other attributes (e.g., flat roof, panels angled south, etc.).
Automated information management process integration would incrementally add the new DER record as part of the customer DER attributes as part of the ADMS load and DER models that are often aggregated up to a transformer or load-point on the distribution feeder model.
With that information, we create a blue-sky-day output curve for that PV DER unit (i.e., the “granular DER model”) with a real-time weather overlay to reduce the output and forecast ahead to corresponding cloud density as a factor of solar irradiance. This then determines the actual output of the PV. The same can be done for other DER technologies, including wind, battery storage, combined heat and power systems, and microturbines.
Once the individual DER is modeled, it can be separated from the overall load by subtracting the DER from the net-metered load, as previously noted.
Finally, the behavior of DER interconnections is modeled to IEEE 1547 interconnection standards.
Per this standard, when there is an outage, the DER trips off, so the load the utility is picking up on restoration of the outage is much more than it was serving prior to the outage. When power is restored, all the air conditioners-or heaters in winter-on the lines affected by the outage try to restart at the same time, drawing up to six times their normal running load when starting up again from a “cold” state. This initial inrush current sometimes overloads the feeder and trips the protection, creating another outage.
By being able to model the load profile of every distributed energy resource in a utility’s coverage area, the utility can then create risk models by analyzing sample data from customer DER records and specific class models. These risk models can then be applied to the utility’s DER population to reduce asset failure by identifying and managing negative performance patterns, and can be used to better incorporate DER, community and economic data into load forecasting.
About the author: Bradley Williams is vice president of industry strategy, Oracle Utilities. Williams is responsible for Oracle’s smart grid strategy as well as utility solutions for outage management, advanced distribution management, mobile workforce management, work and asset management, and OT analytics. Williams has spent the past 30 years driving innovation in the utility industry in roles, including T&D power system engineering, technology development, asset management and industry analyst.