/RESI-One-click-for-future-building-energy-simulation

An easy plug-in for future weather, extreme weather, power outage data download

Primary LanguagePythonMIT LicenseMIT

RESI: One Click for Future Building Energy Simulation

Welcome to the RESI: RESIlient Community!

Seven years ago, when I was a master student and read my first journal article in our field. Most paper began with the statement “building accounts for 30% of energy consumption in the world”, however, this number has not decreased until now. I was always curious about the reasons behind this and realized that scalability is one of the major barriers, not just in modeling, but also in data acquisition, control optimization, retrofit, and implementation.

I was always wondering if there was anything I could do to simplify these processes. This motivated me to develop RESI, a modular-oriented platform where everything is modularized, allowing users to create a pipeline like Lego plug-and-play easily.

MODULE1: Future Weather Data Module

image

This module provides future weather data for the years 2045-2054 and 2085-2094. It includes:

Typical Meteorological Year (TMY)

Extreme Warm Year (EWY)

Extreme Cold Year (ECY)

Specific Heat Wave Events

The data is available in both EPW format and as a pandas DataFrame.

Here's how to use it:

Step 1: Installation <The latest version has been updated to 2.0.9>

pip install ResiSim==2.0.9

Step 2: Import

import ResiSim

Step 3: Query

epw, df, heat_wave, Spic, Sdeb, Sint = ResiSim.Main.weather_query(city='Syracuse', time_span='Mid-term', sce='TMY')

Step 4: You can save this file locally like this:

with open('test.epw', 'wb') as file:
    file.write(epw)

Step 5: Detection, you can use this function to find heat wave events for your own dataset

heat_wave = ResiSim.Main.find_heat_wave_events(df)

For city:

I integrated Google Geocoding, so ideally, you can type any location in the U.S., and you will get weather data for the nearest city.

For time_span, use:

"Mid-term" for 2045-2054 "Long-term" for 2085-2094

For sce, we have:

"TMY" (Typical Meteorological Year) "EWY" (Extreme Warm Year) "ECY" (Extreme Cold Year) "All" (All raw data)

Citation:

Jiang, Z., & Dong, B. (2024). RESI: A Power Outage Event and Typical Weather File Generator For Future RESIlient Building Design and Operation.

MODULE2: Power Outage Module

image image image

This module provides Power Outage data for U.S. from 2002 to 2023.

Step 1: Query

Outage_df = ResiSim.Main.outage_query()
Outage_mcmc = ResiSim.Main.outage_mcmc(Outage_df, sample_size=1500, time_window=5)

For sample_size:

MCMC sample size

For time_window:

the time range you want to sample

MODULE3: Modularized Neural Network Incorporating Physical Priors

Import and run, you will get a data-driven building energy model, which can be used for dynamic modeling, control optimization, energy calculation, and retrofit. Local test ongoing... image SeqPINN_Dynamic

MODULE4: Disturbance Forecasting Module for temperature, solar, and occupancy prediction

A Bayesian-CNN-LSTM model is developed. CNN can extract features from high-dimension space, while LSTM has a strong capability for temporal feature extraction and Bayesian approximation can estimate the distribution of the model output, allowing our model to predict confidence levels by accounting for uncertainties rather than providing a single-point prediction. image

Temperature prediction (24 timestep ahead, hourly resolution):

image

Solar radiation prediction (24 timestep ahead, hourly resolution):

image

Occupancy prediction (96 timestep ahead, 15min resolution):

image