GIM uses identity-protected representative household records to forecast customer-level EV ownership probabilities, charging loads and total 8,760 hourly loads—then aggregates those results to neighborhoods, block groups, ZIP codes and the utility service area.
The term is used differently across the industry. For GIM, it describes a statistically representative customer record that connects household behavior, dwelling characteristics and hourly energy use for scenario analysis.
Each digital twin reflects an actual identity-protected utility customer drawn from the utility service area including information on income, demographics, dwelling, equipment, transportation and hourly-load characteristics. The record acts as an individual decision-making agent whose EV adoption and load response can be modeled.
A GIM digital twin is not a real-time replica of a named customer, a utility billing record or an asset-condition model. GIM forecasts reflect estimates based on a statistically valid sample of actual individual customers within the service area. Available utility data can be used to validate forecast results.
Instead of assigning one average adoption rate and one average charging shape everywhere, GIM evaluates representative households individually and aggregates the results only after customer-level forecasts are complete.
Represent the utility population with identity-protected household and dwelling records.
Match each household with similar observed households using AI-assisted nearest-neighbor analysis.
Translate commuting and charging behavior into customer-level hourly kW profiles.
Add EV charging to baseline end-use loads under weather and growth scenarios.
Produce customer, neighborhood, block-group, ZIP and service-area results.
Consistent variables across the customer, calibration and load databases allow GIM to model EV ownership and charging alongside existing residential demand.
Characteristics associated with technology adoption and energy behavior.
Physical variables that determine baseline and weather-sensitive loads.
Information used to connect vehicle adoption with charging requirements.
Whole-home and component loads provide the baseline onto which EV charging is added.
Observed household and EV information supports the probability model.
Customer results can be summarized at planning-relevant geographic levels.
GIM uses a k-nearest-neighbor process to compare each representative utility household with similar households in an EV ownership calibration database.
Income, education, commuting, age, housing and other characteristics differ across neighborhoods. Applying those differences before aggregation produces concentrated adoption patterns that a uniform forecast cannot reveal.
Both approaches can be made to match the same total number of future EVs. They can nevertheless produce very different local peak-load forecasts because adoption and charging are not uniform.
GIM’s bottom-up approach is designed for the planning questions where customer and geographic variation determine the answer.
The same customer records support multiple views of the forecast, allowing planning, engineering and program teams to work from consistent assumptions.
Baseline, EV charging and combined whole-home profiles for representative customers and scenarios.
Localized EV ownership, peak contribution, hourly thresholds and heat-map results.
Adoption, loads and customer characteristics showing why areas develop differently.
Localized load forecasts used with utility-selected transformer and loading assumptions.
Unmanaged, time-of-use and direct control results with participation and effectiveness assumptions.
Hourly load profiles for CYME, Synergi, WindMil, OpenDSS, GIS and other workflows.
The digital-twin process is not an isolated research exercise. It supplies the forecast detail used by the current GIM planning modules.
Identify likely adoption clusters and forecast charging impacts for current, 2030 and 2035 scenarios.
Locate emerging peak exposure and prioritize transformer, lateral or feeder engineering review.
Combine block-group forecasts with utility-selected transformer size, current loading and planning thresholds.
Connect forecast EV loads with G&T demand charges, co-op margins, program costs, payback and NPV.
Test customer participation, dispatch, duration and rebound against localized and system peaks.
Evaluate EV growth alongside extreme weather, heat pumps, new construction and rebuilding.
Detailed models are useful only when users understand what is observed, what is estimated and how results should be applied.
Yes. Each household record is actually a utility customer within the user's service area with the income, demographics, dwelling unit, commuting and other household information maintained in the customer record. However, each household record is identity protected.
Because adoption varies with income, education, commuting, housing and other characteristics. Uniform allocation can miss the neighborhood clusters that create early distribution impacts.
The k-nearest-neighbor process compares representative utility households with similar households in an EV ownership calibration database and estimates customer-level ownership probabilities.
Commuting distance, return timing, charging requirements and selected charging assumptions are used to build customer-level hourly profiles, which are added to baseline whole-home loads.
Yes. Available AMI, EV registration, GIS, asset and program-performance data can support calibration, validation and more detailed mapping, although they are not required for an initial GIM analysis.
No. It identifies priority locations and supplies scenario-based hourly loads. Detailed voltage, thermal, protection and asset studies remain engineering functions.
Request a guided demonstration of EV adoption, charging loads, geographic results and current GIM applications.