AMPED-Residential
Description
A generalized, agent-based Python model for analyzing minute-level electricity demand - including EV charging, HWC, space heating/cooling, appliances, and lighting- while incorporating photovoltaic generation and battery storage.
Files
Steps to reproduce
AMPED-Residential: Agent-based Model for Predicting Electricity Demand Authors: Ali Kazemian1, Baxter Kamana-Williams1,2*, J. Geoffrey Chase1 1Department of Mechanical Engineering, University of Canterbury, Christchurch, New Zealand 2Sustainable Energy Research Group (SERG), Department of Civil and Environmental Engineering, University of Canterbury Corresponding author: baxter.williams@pg.canterbury.ac.nz Overview AMPED-Residential is an open-source, generalisable agent-based model (ABM) of residential electricity demand. It generates minute-level household load profiles for existing and emerging residential loads, including electric vehicles, solar PV, and battery storage. The model supports demand-side management (DSM) research, policy analysis, and energy system planning across diverse socioeconomic and climatic contexts. Repository Structure AMPED-Residential/ │ ├── AMPED-Residential Agent-based Model for Predicting Electricity Demand.ipynb │ └─ Main simulation file containing the full agent-based model │ ├── Appliance profiles.xlsx │ └─ Average appliance use behaviour and power ratings │ ├── dm_050.csv ├── dm_100.csv ├── dm_150.csv ├── dm_200.csv ├── dm_250.csv │ └─ Domestic hot water (DHW) demand profiles generated using DHWcalc │ (approx. 50–250 L/day, representing households with 1–5+ people) │ ├── tmy_Auckland.csv ├── tmy_Wellington.csv ├── tmy_Christchurch.csv │ └─ Typical Meteorological Year (TMY) climate data from NIWA │ including ambient temperature and solar irradiance for each city └── Prerequisites Python ≥ 3.9 Recommended packages: numpy, pandas, matplotlib, openpyxl, tqdm Running the Model Clone this repository Open the main notebook: "AMPED-Residential Agent-based Model for Predicting Electricity Demand.ipynb" Edit the input parameters and scenario settings as required (see the Pre-processing section of the notebook). Outputs Minute-resolution electricity demand (per household and aggregate) Load breakdowns by appliances, water heating, space heating/cooling, lighting, EVs, PV, and batteries Summary statistics, peak demand analysis, and CSV exports for post-processing Citation Kazemian, Ali; Kamana-Williams, Baxter; chase, geoff (2025), “AMPED-Residential”, Mendeley Data, V1, doi: 10.17632/h29jy9xn4x.1
Institutions
- University of Canterbury Department of Mechanical Engineering