The monitoring system has been designed in close collaboration with Melbourne Water within the Melbourne Waterway Research-Practice Partnership, and more especially the B3 (Wetland) and C3 (Monitoring) projects.

The set-up installed in the Troup Creeks wetland in Hampton Park, Victoria is an illustration of the monitoring system installed on the different sites.

TRCS_syst-installed — System installed in Troups Creek (zoom in the bottom right of the picture)

Monitoring objective

The main objective is to design and build a water level monitoring system for wetlands or equivalent water infrastructures.

The design objectives are:

TRCS_Objectives — Objectives for the design of the system

First prototype

The aim of the first prototype was to test the sensor performance, the communication and the system autonomy. The board used is the amazing Pycom Lopy4.

Final design and installation

TRCS_syst-installed-zoom — Installation of the system on site. The picture on the bottom left show how the system is easily attached to the pole. The main figure shows the system installed close to another water level sensor (for testing)

The system is build using two pipes of different diameter. The main idea is to leave on part of the system (the pipe with the biggest diameter) attached to the pole and to unscrew the rest of the system.

TRCS_set-up-details — Details of the system build with two pipes of different diameter to facilitate any maintenance action

The system use a waterproof ultrasonic sensor (JSN-SRT04), and a DHT22 sensor to measure the air temperature and relative humidity. The temperature and relative humidity are used to correct the distance measures based on the real celerity of sound.

Public platform

All the data is accessible in real-time on the Mind4Stormwater online platform. Anyone can visualise one or several parameters simultaneously and download the data.

TRCS_live-data — Live parameters from the Mind4Stormwater online platform.

Communication & autonomy

The communication system is used low-power wide-area network, that is to say long range, low power and low cost technology. In order to improve the coverage the system is using both LoRaWAN and Sigfox networks.

The system is powered by 3 AA rechargeable batteries connected to a 0.5W solar panel. Previous have shown that this setup is sufficient to function for several months.

TRCS_bat — Battery voltage for two set-ups (US v2: without solar panel, and USv3: with a solar panel). The fluctuation are due to sun,light during the day and change of temperature

Performance of the system

The figure below present the evolution of the water level and the daily cumulative rainfall (from the closest Melbourne Water weather station). The result shows a quick response of the catchment to the rain.

TRCS_water-level_rain — Water level (mm) and daily cumulative rainfall (mm) from a Melbourne Water weather station

Laboratory tests have shown that the expected accuracy of the system is 2 cm when the distance is corrected with the air temperature, and when the distance to measure is between 0.2m and 1.9m

The two following figures present the comparison of the air temperature and relative humidity between the system and the closest weather station from the Bureau of Meteorology.

TRCS_air-temp — Comparison of the air temperature measure by the system (DHT22) and the measure from the BOM (Bureau of Meteorology)

TRCS_rel-humid — Comparison of the air relative humidity measure by the system (DHT22) and the measure from the BOM (Bureau of Meteorology)

The results show that the DHT22 sensor is giving results sufficiently accurate to correct the celerity of sound. The difference are more important with the relative humidity because the DHT22 sensor is located above the water (in the wetland).

The last chart presents the evolution of the battery voltage for a month. The voltage is maintain over time even after days with less sun (end of January).

TRCS_voltage — Evolution of the battery voltage over time

MicroPython code

The code is accessible on Github:
https://github.com/fcherqui/Mind4stormwater/tree/master/US_water_level
Using a Pycom Lopy4