GridForge research group
Research methods for on the low-voltage grid.
A standing research group at Monash, building cumulative tools and methods for the low-voltage distribution grid. Each cohort picks up where the last left off, so the work accumulates on a shared platform instead of restarting every year.
The methods.
Five problems on the LV grid, approached using smart-meter data alone: no extra hardware, no extra site visits. Each method is built into the shared platform, so the next cohort starts from what already exists.
01 / Topology estimation
Who is connected to whom?
Inferring the connectivity structure of an LV network from smart-meter data alone: which customers share conductors, where branches join, and where the boundaries lie.
02 / Phase grouping
Which customers share a phase?
Clustering customers by inferred phase membership from voltage time-series correlation, without measurements at the substation or prior knowledge of the wiring.
03 / Impedance estimation
How does current flow?
Recovering conductor resistance and reactance from observed voltages and powers. Underpins load flow, loss accounting, and cross-validation of topology.
04 / Transformer monitoring
What is the transformer doing?
Estimating transformer loading and stress from downstream smart-meter readings, with no instrumentation at the transformer itself.
05 / Situation analysis
What if things change?
Stress-testing networks against rooftop solar uptake, EV charging patterns, and customer-mix shifts. Voltages and loadings under hypothetical futures, and where battery storage should sit.
Each year picks up where the last left off.
GridForge runs on a rolling cohort. The 2025 founding cohort built the platform from scratch: a Streamlit GUI wired to a Python analysis core and an OpenDSS simulation backend. The 2026 cohort is extending it with topology evaluation, refinements to published estimation methods, and new metrics for comparing topology estimates. A second-semester group has also started on optimal battery placement, which is the first work under situation analysis.
A shared workbench means new students see what's already been built, and outputs from one method become inputs to the next.
Meet the cohorts- 2027 Roadmap · impedance estimation, real-feeder pilots Planned
- 2026 Digital twin · AI topology evaluation · topology metrics · battery placement Active · 4 groups
- 2025 Platform pilot · Streamlit + OpenDSS · phase grouping Founding
The part of the grid with the least data.
The medium- and high-voltage grid is well-instrumented. The low-voltage grid, where customers actually live, is largely unmeasured, and its topology is often unknown even to the utility. That is the part we work on.