Improved design and testing in an urban environment
By Stijn de Jong
In 2009 the department of water management started with the development of a cheap rain gauge that is suitable for use in remote areas and tropical areas. The first prototype of this rain gauge was based on an acoustic disdrometer and an audio recorder was used for data acquisition. This disdrometer could be produced for only a fraction of the market price for a rain gauge with the same specifications. Because of the low price this disdrometer also lends itself for large scale application in urban area, where a lot of measurements are desired, because of the spatial distribution of rainfall.
The main objectives of this thesis are the improvement of the design of the sensor and data acquisition, testing the new prototype and develop a method to validate the measurements.
The design of the sensor has been improved in terms of sensitivity and the method of production. From the measurements it is observed that the smallest drop measured by the current prototype is 0.6 mm, where the smallest drop measured by the first prototype was 1.5 mm. In the current prototype two negative effects are observed, the radius effect and the puddle effect, however these effects do not have a visible effect on the measurements.
Data acquisition with an audio recorder, which was used for the first prototype, requires a lot of data storage. This is mainly because the signal is measured continuously. Also the amount of energy needed is a disadvantage of the audio recorder. A new data logger was developed for the sensor to reduce the amount of data that has to be stored and to increase the time that the disdrometer is able to measure. The logger can be used in two configurations, as a standalone unit and in a setup with multiple disdrometers (Rainscan).
For the calibration of the low cost disdrometer an expermental calibration setup is used. With this setup the outputs of the logger are linked to drops of known size. Calibration of the disdrometer showed a clear relation between the signal energy of a drop and the size of the drop. The uncertainty in the calibration curve can be decreased by calibrating the disdrometer with the help of an optical disdrometer.
The disdrometer gives an array of drop energies as output, from which a drop size distribution can be derived. A method is developed to validate the outcome of the low cost disdrometer, in terms of drop size distribution, if there are no other disdrometers available to compare to. With this method it is possible to get insight in the expected drop size distribution based on the data of a tipping bucket.
To test the low cost disdrometer in an urban environment two case studies were conducted. For each case study, one configuration of the low cost disdrometer was used. For the EWI case study the Rainscan configuration was used, where for the Singapore case study the standalone version of the disdrometer was used.
The results from the measurements in the case studies showed a large underestimation of the total amount of rain, which indicates an error in the design. Therefore the results are not used for any investigation of rainfall in the urban environment. Analyses show that this underestimation only takes place during rain events with high rain intensity. The maximum intensity measured with the disdrometer is 26 mm/hr where the tipping bucket measured 150 mm/hr. Several tests showed that this underestimation is caused by an error in the logger of the disdrometer. So far it is unknown what the exact error is.
The main conclusion drawn from the conducted research is that the sensor of the current prototype performs better than the first prototype and that in general the principle works. However the logger designed for the disdrometer is not capable of giving a right representation of the reality under all circumstances. Before the low cost disdrometer can operate in the field for longer periods, first the problems with the data logger have to be solved.