At a concert, a band member may break off into a solo to captivate the audience. However, it is the other side of the solo that sets the best bands apart: orchestrating the return of the soloist back to playing as a full band in perfect harmony. The same can be said about the two sides of peak load management. Discussions around the benefits and challenges of “turning things off” dominate, but the real art is to “turn things back on” in an orchestrated manner.
In fact, “turning things back on” after a load management event is much harder than turning things off because of a lack of load diversity. Without load management, load is naturally diversified, given that people’s daily routines differ and are not synchronized. For example, some need power straight after work for water heating or electric vehicle (EV) charging, while others could take much longer to get home after a networking event, a meet-up with friends, a gym visit, or just being stuck in traffic. This is best witnessed by observing activity in an apartment block, as shown in Figure 1. When aggregated, the diversity of behavioral patterns presents the smooth shape in high-level daily electricity load profiles.
Figure 1. Behavioral Diversity Exemplified by Evening Activity in an Apartment Complex
Source: Adobe Stock / zyabich
Composition of a Load Management Event
During load management, demand drops as the load of some distributed energy resources (DERs) like electric water heating or EV charging, which are in a naturally diversified state, are being turned off (or throttled). However, if those DERs are brought back together after the load management event, their demand is synchronized (i.e., all hot water cylinders or EV chargers start drawing power together). This leads to a load pickup which is bigger than the load reduction achieved by the load management event. Depending on timing, the pickup could be bigger than the natural load without any load management, creating a secondary peak. Figure 2 exemplifies this, with electric hot water load management across 100 homes on a single network feeder. It highlights that the load pickup after the two control periods (e.g., when hot water has been remotely turned off) is higher than the load drop at the beginning the two control periods.
Figure 2. Load Drop and Pickup on a 11 kV Distribution Feeder from Hot Water Load Management
Source: Vector
This inconvenient reality needs to be more widely understood, so that the demand side can truly unleash its benefits to consumers and the whole system. The electricity distribution network is currently evolving towards the distribution system operator (DSO) concept thanks to ongoing efforts in many jurisdictions. As a DSO, the two new tasks of network companies can be summarized as follows: (1) use load management of third-party DERs as a non-wires alternative to optimize the need for traditional network capacity, and (2) ensure that DER owners or aggregators can operate load management (so-called virtual power plants) to optimize their market position or provide ancillary services without negatively affecting the distribution network assets. At the same time, as the electrification and decarbonization of the economy continue to ramp up and resilience is at the forefront of public concern, the legacy role of the distribution network utility to deliver value to its ratepayers via safe, reliable, and cost-efficient operations and asset management is more important than ever. Bringing managed load back online in an orchestrated way to avoid unintended network overload and/or outages is therefore a priority for electricity distribution networks.
Load management needs to avoid the potential negative impacts of load pickup after a load management event, to ensure that distribution companies can bank the avoided network build enabled by the demand reduction (DSO task 1 above). This also ensures that third parties can optimize their DER management without negatively affecting distribution network customers or assets (DSO task 2 above).
Of most concern for distribution systems are low-voltage networks, given that diversity is lower (e.g., a smaller group of people is more likely to behave the same way) and that visibility of those networks hinges on advanced metering infrastructure (AMI) data (e.g., due to limited dedicated monitoring equipment on the low-voltage network). Poorly coordinated load pickup such as hot water or EVs, or retail models such as free hours of power, will create peaks that stress the local network or transformer with potential reliability implications for customers. By coordinating load management and harnessing actual network information to do so, the DSO will: (1) ensure that load is only managed when the network is constrained (as opposed to a fixed period over a whole season), (2) optimize the DER schedule across different network levels and in accordance with the actual network configuration (e.g., switching position and outage management), and (3) bring back load smoothly to avoid secondary peaks (including after an outage).
Orchestrating the Response
The importance of managing load pickup is not new. Figure 3 illustrates peak shaving of an electricity distribution asset by deferring the residential demand to later in the evening when commercial demand (not shown in Figure 3) has considerably dropped off. Distribution network operators with experience in load management orchestrate load pickup for both peak shaving (Figure 3) and outage management (e.g., to avoid high load pickup after an outage, which would damage equipment or trip off load). However, the broader DSO or virtual power plant discussions often do not consider practical processes (such as data exchanges) and market arrangements on how load can be brought back in an orchestrated manner.
Figure 3: Shifting Residential Load by Managing Hot Water Usage in Orchestrated Manner on the Vector Network
Source: Vector
In the case of EV charging, Vector’s EV Smart Charging Trial demonstrated that smart charging can seamlessly manage EV charging load while maintaining customer satisfaction. More resources on the results from the trial are available on our website: https://www.vector.co.nz/articles/ev-smart-charging-trial
Time-of-use tariffs or even dynamic prices are unlikely to be a panacea, as they don’t actually consider the physics and loading of the network. Under such pricing, customers will increasingly use timer-based control or localized algorithmic control, which will create a synchronization of the load and secondary peak when bringing load back.
We recognize the importance of several research projects that demonstrate how one could also create network prices that capture the full physics of the network across many network voltage levels. However, these solutions are not yet available across a real physical network, and, even if they were, they would require enormous time and investment in data, systems, and people from the outset, which, in the face of global climate emergency, is not a credible pathway for most utilities.
Three Pillars for DSO Strategies
Affordable decarbonization will hinge on realizing the benefits of load management across the electricity system. When establishing DSO frameworks, more focus needs to be given to the forgotten side of load management — “turning things back on” — so that DER/load management can act as a non-wires alternative, and DER optimization for other markets does not create new constraints on the distribution network, and costly reinforcements that will impact affordability.
Three key pillars for DSO strategies are critical to achieve these outcomes:
- Most DERs will last 10+ years, so any DER connected today needs to be able to be onboarded. To achieve interoperability of customer DERs, utilities cannot pick a single winner, but need to promote several strong protocols (e.g., IEEE 2030.5, OpenADR, Open Smart Charging Protocol (OSCP)) by making them part of the local standard.
- Low-voltage AMI data is available and can provide full low-voltage visibility if network companies can access the data. While this is a given in some jurisdictions, in others, network companies face regulatory or commercial barriers to even access the data.
- Distributed energy resource management systems (DERMS). A DERMS will combine network asset information (ratings, live topology, and switch status) with granular load forecasts and DER information to orchestrate fleets of DERs and their aggregators.
These three pillars will set load management up for success, ensure coordination across all network levels and deliver true customer and societal outcomes.
Steve Heinen, Future Network & Planning Manager
Andre Botha, Chief Engineer
Vector Ltd.
Auckland, New Zealand
What if we introduce a proactive session oriented edge resource transfer control layer 4 protocol as a new added option for a grid forming inter-DER energy control plane? A layer 4 resource transfer session oriented signaling suite that’s already globally deployed at every telco worldwide today? Multimedia and Energy is a resource type that shares same dimensions of: edge generation , storage, adaptation, and consumption. In 1998 after a 5 year funnel of 2 and 3 rd gen session oriented media transfer control protocols refinements by many telephony media engineers they honed in on the Session Initiation Protocol rfc3261. Can we discuss and show how Ieee2030.5 smart Energy profile description over a SIP session oriented energy transfer between 1 or more legs of generate, store, adapt, consume DER edges can enable a proactive session oriented transfer of energy between the DERs. Moreover it could be demonstrated within a few months POC effort among a group of updated DERs using SIP control plane opensource how DER resources capabilities can be dynamically discovered , with energy transfer sessions negotiated, directly between DERS, while also proactively energy transfer policy mediated. SIP scales already today for the many 1000’s of multimedia cellular RAN and IPTV mediated resource transfer sessions that occur today. This could especially be enabling to emerging grid forming Inter-DER proactive mediated energy transfer sessions over a shared smart distribution grid. It requires the smart distribution grid to be an active and dynamic control plane resource member capable of advertising dynamically the feeders current usage level and capacity and capability. If interested to explore this further please reply here. A similar recent use case leveraging SIP is captured in h.771 IPTV use of SIP discovery and negotiation and renegotiation all within the same resource transfer session between generation and consumption and adaptation edges. “Energy is the new Media” for SIP.
@Glenn Algie The SIP protocol could serve the “DSO” orchestrator providing distribution at the edge, to traffic local dispatch signals – horizontally to demand or vertically as an on-ramp/ to the wholesale markets. AMI data access can be managed at the edge. This eliminates the requirements for public funding of AMI 2.0 or 3.0 schemes (lessons learned)
A very important potential benefit of your taxonomy is to migrate grid M&V and power controls into the 21st century. This evolution will require more real time AMI utilization, which will foster innovation from sectors like telecom and AC/DC power electronics.
Networks have their work cut out convincing people they can manage “orchestration” well.
It would be a poor outcome for consumers if “orchestration” turned out to be marketing speak by networks for adding other customer devices to hot water cylinders in their demand management portfolio. The customer experience in hot water management (eg turn them all off at peak times during the day and switch them back on at night) is an overly simplified harm management exercise. It doesn’t exactly demonstrate networks are well suited for managing the complexity of decisions required to maximise the service experienced by each customer and deciding when and which customers should have their devices switched on and off.
Also, any decision to manage export or import congestion without altering electricity prices in the affected area is brave as it makes “orchestration” decisions appear arbitrary and manifestly unfair. For example, how do you convince customers their inverters for rooftop solar PV must be turned off because there is too much local generation while charging them for consuming electricity from the network?