Fire Risk Prediction Using Cloud-based Weather Data Services
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Dry and cold winter seasons result in very dry indoor conditions and have historically contributed to severe fires in the high and dense representation of wooden homes in Norway. The fire in Lærdalsøyri, January 2014, is a devastating reminder of town fires still posing a threat to a modern society. In order to reduce conflagration probability and consequences, it is necessary to have an accurate estimate of the current and near future fire risk to take proper planning precautions. Cloud computing services providing access to weather data in the form of measurements and forecasts, combined with recent developments in fire risk modelling, may enable smart and fine-grained fire risk prediction services. The main contribution of this study is implementation and experimental validation of a wooden home predictive fire risk indication model, as well as outlining a wooden home fire risk concept. The wooden home fire risk model focuses on the first house catching fire (indoors) in a potential conflagration event. Such a fire would be critical to intervene prior to the fire developing exterior flames and embers post flashover, and thus high likelihood of fire spread. The implemented model exploits cloud-provided weather measurements and forecasts, to predict the current- and near future fire risk at given geographical locations. It computes the indoor wooden fuel moisture content of houses that may catch fire, using measured and forecasted outdoor temperature and relative humidity, and estimates the time to flashover. The latter is found through an empirical relation with the fuel moisture content, and can be used as an indication of the fire risk, beyond the modelled single house. The model implementation was integrated into a micro-service based software system and experimentally validated at selected geographical locations, relying on weather data provided by the RESTful API’s of the Norwegian Meteorological Institute. The validation took place by applying the model to predefined cases, with an outcome known from observations or theory. The first part is a general evaluation of the outputs, considering three historical fires. Then, seasonal changes and natural climate variations were considered. Our evaluation demonstrates the ability to provide trustworthy and accurate fire risk indications using a combination of recorded weather data and forecasts. Further, our cloud- and micro-service based software system implementation is efficient with respect to data storage and computation time. Finally, the novel fire risk concept is demonstrated for a selected city, based on model output. It successfully depicts the implications following reduced indoor humidity by utilizing location specific fire risk contours.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it