About Smart HEMS Benchmark
Smart HEMS Benchmark is an open source platform for transparent, reproducible, manufacturer-neutral benchmarking of home and community energy management. It provides shared datasets, standardized scenarios, common metrics, and reference algorithms across the full residential distributed energy resource (DER) lifecycle: siting and sizing decisions with a roughly 15-year horizon, day-to-day home energy management system (HEMS) dispatch, and virtual power plant (VPP) grid support.
Project Lifecycle Stage
Sandbox
Project Special Interest Group
Grid Simulation & Modeling
The Challenge
As households add rooftop solar and battery storage at accelerating rates, with solar PV projected to be the largest renewable source by 2030 and residential storage expanding rapidly alongside it, decisions about DER siting, sizing, HEMS control, and grid support increasingly determine how much value these systems return to the people who own them.
Today those decisions are typically evaluated with proprietary datasets, ad hoc scenarios, and vendor-specific tools, so results from different studies and products are difficult to compare or trust directly against one another. The residential DER ecosystem needs a shared, transparent basis for evaluating siting and sizing choices, HEMS dispatch strategies, and grid-support participation, so that vendors, researchers, utilities, and end users can all work from the same common ground.
Smart HEMS Benchmark addresses this by fixing the datasets, scenarios, and scoring metrics so that any control approach, whether rule-based, model predictive, or AI-driven, can be compared on equal footing across the full DER lifecycle.
Key Features
DER Decision-Making Modeling
Smart HEMS Benchmark models DER decision-making as an optimization problem spanning three areas, each with distinct decision variables, objectives, risks, and time horizons: Location & Sizing (roughly a 15-year horizon), HEMS (single-household daily dispatch), and Grid Support (VPP participation).
Benchmark Topology
A universal description language linking system nodes (PV, ESS, load, EV, grid import/export) and their energy flows to HEMS constraints.
Benchmark Configuration
Standardized configuration covering rated PV capacity, load consumption level and composition, ESS capacity and maximum charge/discharge power, inverter capacity, and import/export tariffs.
Predicted Day-Ahead Data
PV, load, and price curves at configurable resolution, supporting realistic day-ahead planning scenarios.
HEMS Algorithms
Baseline controllers (Grid-only, Grid+PV, Self-consumption) and advanced controllers (AI-driven time-of-use optimization, battery-health-aware control, model predictive control), with support for adding third-party algorithms.
Realization & Ranking
Evaluation against actual PV, load, and price data using common HEMS metrics, sensitivity analysis across PV capacity, ESS capacity, load level, and grid export limit, and multi-metric ranking spanning cost and revenue, comfort, and battery degradation and levelized cost of energy (LCOE).
Key Contributors
- EcoFlow
- CoSES, Technical University of Munich
- Stanford University
Use Cases
The following application areas are the intended use cases for Smart HEMS Benchmark as a neutral, manufacturer-independent research platform.
Vendor Controller Benchmarking
Vendors benchmarking their HEMS controllers against shared datasets, standardized scenarios, and common metrics rather than proprietary, non-comparable evaluations.
Research Reproducibility
Researchers reproducing and comparing HEMS and DER sizing methods on a consistent platform, addressing the current absence of standardized scenarios for evaluation.
Utility and VPP Flexibility Assessment
Utilities and VPP operators assessing the flexibility available from residential DER participation in grid support programs.
End-User HEMS Comparison
End users comparing HEMS offerings, with planned integration into HEMS Finder’s network of more than 100 HEMS service providers.
Collaboration Opportunities
Smart HEMS Benchmark welcomes participation from researchers, vendors, utilities, and platform developers working on:
- Standardized datasets and scenarios for residential DER siting, sizing, and dispatch
- HEMS control algorithm development, from baseline controllers to AI-driven and model predictive approaches
- VPP grid-support evaluation methods for residential battery and PV fleets
- Integration with third-party HEMS and energy management platforms
Research groups seeking a neutral benchmarking platform for residential DER and HEMS evaluation, and organizations building HEMS controllers or datasets, are encouraged to get involved as the project moves toward its late 2026 open source launch.
FAQ
What problem does Smart HEMS Benchmark solve? As households add rooftop solar and battery storage at accelerating rates, decisions about DER siting, sizing, HEMS control, and grid support increasingly determine how much value these systems deliver. Smart HEMS Benchmark provides shared datasets, standardized scenarios, and common metrics so any control approach, whether rule-based, model predictive, or AI-driven, can be compared on equal footing.
What phases of the DER lifecycle does it cover? Smart HEMS Benchmark spans three phases: Location & Sizing (roughly a 15-year horizon), HEMS (single-household daily dispatch), and Grid Support through virtual power plant participation, each with distinct decision variables, objectives, and time horizons.
Who is behind Smart HEMS Benchmark? The project brings together three independent research threads: Dr. Rui Li (former Head of AI, EcoFlow) in China, Dr.-Ing. Anurag Mohapatra (CoSES, Technical University of Munich) in Europe, and Dr. Tao Sun (previously at Stanford University, now at Shanghai Jiao Tong University) in the United States. They presented a shared vision for open benchmarking at the 2025 LF Energy Europe Summit and are developing the platform under neutral, multi-stakeholder governance at LF Energy.