h593286/dynamic-frcm

DYNAMIC Fire risk indicator implementation

This repository contains the implementation of the dynamic fire risk indicator described in the submitted paper:

R.D: Strand and L.M. Kristensen. An implementation, evaluation and validation of a dynamic fire and conflagration risk indicator for wooden homes. Submitted for review.

which uses forecast and weather data observation for computing fire risk indication in the form of time-to-flash-over (ttf) for wooden houses.

Installation

The project is based on using the Poetry package manager.

Start by installing Poetry for your platform using the installation instructions

Create a new virtual Python environment for the project and activate the virtual environment.

To install the required Python packages use:

poetry install

Overview

The implementation is organised into the following main folders:

• datamodel - contains an implementation of the data model used for weather data and fire risk indications
• weatherdata contains an client implementations and interfaces for fetching weather data from cloud services.
• fireriskmodel contains an implementation of the underlying fire risk model

The main API for the implementation is in the file frcapi.py

Weather Data Sources

The implementation has been designed to be independent of any particular cloud-based weather data service.

This library contains an implementation that use the weather data services provided by the Norwegian Meteorological Institute (MET)

Specifically, the files src/frcm/weatherdata/client_met and src/frcm/weatherdata/extractor_met.py shows how to implement client and extractors using

• MET Frost API for weather data observations: https://frost.met.no/index.html
• MET Forecasting API for weather data forecasts: https://api.met.no/weatherapi/locationforecast/2.0/documentation

To use these pre-implemented clients a file name .env must be put into the src/frcm/weatherdata folder having the following content:

MET_CLIENT_ID = '<INSERT CLIENT ID HERE>'
MET_CLIENT_SECRET = '<INSERT CLIENT SECRET HERE>'

Credentials for using the MET APIs can be obtained via: https://frost.met.no/auth/requestCredentials.html

Please make sure that you conform to the terms of service which includes restrictions on the number of API calls.

Testing

The folder tests contains a number of unit tests that can be used to test the implementation.

To execute the tests enter the tests folder and execute

pytest

You may also execute a specific test-file in the test-suite using, e.g.,

pytest test_datamodel.py

or running a specific test within a test-file using, e.g.,

pytest test_datamodel.py::TestDataModel::test_validate

Application integration

The file src/main.py provides sample code on how to compute fire risk indications and how to integration the implementation of fire risk indications into applications.

Running the command

python3 main.py

should result in similar output as below

...

FireRiskPrediction[latitude=60.383 longitude=5.3327]
FireRisks[2024-01-13 00:00:00+00:00 6.072481167177002]
FireRisks[2024-01-13 01:00:00+00:00 6.084901639685655]
FireRisks[2024-01-13 02:00:00+00:00 6.115931368998181]
FireRisks[2024-01-13 03:00:00+00:00 6.146731614820376]
FireRisks[2024-01-13 04:00:00+00:00 6.157474811242355]
FireRisks[2024-01-13 05:00:00+00:00 6.148031627082924]
FireRisks[2024-01-13 06:00:00+00:00 6.135871365180064]

...

FireRisks[2024-01-24 23:00:00+00:00 5.782145706635683]
FireRisks[2024-01-25 00:00:00+00:00 5.789519566277565]
FireRisks[2024-01-25 01:00:00+00:00 5.796865122491866]
FireRisks[2024-01-25 02:00:00+00:00 5.80463250152655]
FireRisks[2024-01-25 03:00:00+00:00 5.812626103446193]
FireRisks[2024-01-25 04:00:00+00:00 5.820770098316547]
FireRisks[2024-01-25 05:00:00+00:00 5.829024065716487]
FireRisks[2024-01-25 06:00:00+00:00 5.837363568856687]

showing hourly computed fire risks for the given location.