Indoor Fire Dataset of Multi Sensor Data (Small Scale Test Chamber)

The dataset contains 4 incipient fire scenarios (wood, candles, cable, lunts; indicated by the column "scenario_label") carried out in a (2 x 0.8 x 0.6)m^3 test chamber without ventilation. Each scenario was repeated 4-6 times (indicated by he column "number_label") in random order with background sequences in between to reduce the influence of prehistory. Each experiment consists of multiple stages (up to 6 stages, indicated by the column "intensity_label") to simulate the development phase of a real scale fire in the small scale setup. The wood, cable and lunt material was burned using a 12 A heating coil, the candles fire was ignited with a lighter. The dataset consists of 2900 rows and 18 columns and is structures as a continuous multivariate time series. Each row represents the sensor measurements (CO2, CO, H2, humidity, particulate matter of different sizes, air temperature, VOC and UV) from one of four unique sensor nodes placed in the test chamber at a specific timestamp. In addition, a trend value based on the Kendall-Tau coefficient was calculated for each sensor measurement. The "Sensor_ID" column can be utilized to access data from different sensor node positions.