Rain Validation

For the rain validation, the 3D Stereo disdrometer from Thies Clima, a Weather Station equipment and a square meter method are utilized. The disdrometer quantify the DSDs of the rain, which describes a statistical distribution of the falling drop sizes, with the diameters and velocities of the respective drops. Despite its ability to provide rain intensities, it is unsuitable for application in artificial rain. Hence, to analyse and address this limitation, the concept of water volume derived from the DSDs is introduced and it is predicated upon the assumption that raindrops assume a spherical form. The WeatherHat weather station is equipped with mechanical sensors that operate through physical interaction with wind and rain. It utilizes magnets, reed switches, and moving parts to generate signals, which are then processed by a Raspberry Pi 4 Model B. These sensors provide measurements for wind speed, wind direction, and rain intensity. The square meter method consists of recording the amount of collected water within a predefined time and then calculating its volume using weight and density of water. Subsequently, utilizing the recipient's surface area A,the rain intensities can be computed. Therefore, the collected data is subjected to a comprehensive assessment, involving four distinct analyses that draw comparisons across three distinct environments: the outdoor rain, an indoor synthetic rain, and real-world rain conditions. These comparisons serve the purpose of evaluating the similarity between artificially emulated and authentic rain patterns, while also examining the efficiency of the generated rain under outdoor circumstances as compared to controlled indoor settings. Detailed accounts of each analysis are elaborated upon in the subsequent subsections.

Preliminary analysis of the sprinklers

With the purpose to understand the capabilities of the outdoor rain facility and identify the intensity distribution generated by the different systems, a preliminary test is conducted evaluating the performance of two different garden sprinklers. The equipment evaluated consists of a rotating sprinkler, with a flow of 6.88 L/min, and a spray sprinkler, with a flow of 8.90 L/min. Each of these sprinklers covers a semicircular area spanning 180 degrees, with an effective radius of 9m. The main difference is that the rotating sprinkler has a rotating water jet that scans the actuation area, while the spray sprinkler generates a widespread waterjet in its actuation area.

For the evaluation, the volumetric quantity of water produced by each sprinkler in 30 min is measured in five distinct positions. These positions are linearly aligned with the radius of the sprinkler and spaced at 1.8 m intervals. Additionally, with the value of the volumetric quantity of water in the five positions, the square meter method is employed to compute the average of rain intensity for the sprinklers.

Synthetic outdoor rain

The disposition of the sprinkles can be seen in Figure 4, represented by triangular markers represent the positions of rotating sprinklers, serving as supplementary elements depending on the rain intensity. The yellow markers are employed throughout the entire 112 m generating 10 mm/h rain intensity. However, for the 25 mm/h intensity, rain is only generated in the first 56 m, so it uses yellow and red rotating sprinklers for that distance. Lastly, in the case of rain intensity of 50 mm/h, the green markers are introduced, while the yellow and the red markers starting from 28 m of distance are removed.

Figure 4 includes circular markers indicating the designated positions for the disdrometer measurements, while the cross markers are exclusive measurement positions for the 50 mm/h rain. Furthermore, the square markers delineate the disdrometer zones for each intensity — yellow for 10 mm/h, red for 25 mm/h, and green for 50 mm/h. At the same location as the disdrometer, data is also collected by the WeatherHat and the square meter method. The records are collected over one minute at each point for the disdrometer and for the WeatherHat. For the square meter method, data is collected over 30 minutes, as shown in Figure 5.

Indoor Rain Validation


Outdoor validation with real rain data

To validate if the synthetic rain exhibits similar characteristics to real rain, the disdrometer and the WeatherHat were set outdoor to collect continuous data of the environment over a period of 5 days. Subsequently, is required an initial processing of the data to identify the rainy periods. From the total amount of over 6.700 minutes of recorded data, 370 minutes are detected as rainy conditions, by observing whether the number of drops detected by the disdrometer exceeded 100. For those files, the DSDs curves are compared with the synthetic outdoor rain data and as evaluation method, the Root Mean Square Error (RMSE) is applied. RMSE can be conceptualized as a representation of the standard deviation characterizing the disparities between measurements and a reference curve [6]