University of Washington Tacoma CES Projects

Stream Discharge Measurement System

The Puget Creek Restoration Society (PCRS) first partnered with the UWT seeking an inexpensive measurement device for measuring stream discharge for local research.
Visit: The Puget Creek Restoration Society

Problem Statement

The Puget Creek Discharge Measurement device provides an automated way to measure flow, velocity, and total discharge (volume per unit time) for use in potentially shallow creeks and streams. The initial users of this product will be the Puget Creek Restoration Society. Eventually anyone connected to watershed management may use it. Current methods of measuring creek discharge are either too cumbersome, such as measuring by hand, or are optimized for rivers and thus do not work well in creeks. The product will features two sensor groups, one to measure flow and one to measure depth. The collected information in relevant units is displayed to the user while collecting data.

The (PCRS) conducts studies of ground water run-off to identify and quantify levels of toxins introduced into local streams and eventually into the Puget Sound. By identifying levels of toxins at a given time for a single stream and its average water discharge, estimates can be made as to total amount introduced into the Puget Sound over time.

Background of the System

Original Contributors:

The design began as a Senior project (2013) at UWT by Anthony Claudio and Cristopher Claeys. There were a number of challenges to overcome for them, and time didn't permit the project to be completed by the deadline established.

The New Team:

Bob Landowski, Abdiel Cabrera, and I were given the opportunity to work on the project over the summer of 2013. The first task was to understand the fundamental goals of the project. Designs like this involve understanding more than just the electronics needed to accomplish some task. In this case, we had to learn about how water discharge measuring is accomplished using existing systems and identify what limitations those systems might have for this application. Why re-invent the wheel if not necessary?

What is Stream Discharge?

"Discharge is the volume of water moving down a stream or river per unit of time, commonly expressed in cubic feet per second or gallons per day. In general, river discharge is computed by multiplying the area of water in a channel cross section by the average velocity of the water in that cross section:

discharge = area x velocity.

Current-meter discharge measurements are made by determining the discharge in each subsection of a channel cross section and summing the subsection discharges to obtain a total discharge." [1]

Commercially available systems were evaluated such as the Swoffer Fiber-optic sensor. Two major limitations were identified with this particular product. First, the unit costs thousands of dollars. Not all organizations, like the Puget Creek Restoration Society, have the funds to purchase a device like this. The goal our design was to provide an economical alternative to existing designs. Secondly, the device has been designed for water flow measurements in larger bodies of water like rivers. The streams that are the target of the PCRS studies often have a depth of less than 6" While the fiber-optic sensor does operate in relatively shallow waters, it requires around 2-3" of water for the propeller to operate correctly. Our design, ideally, would not rely on a propeller or at least be able to measure water velocity in 2" of water with a smaller propeller.

In researching existing tools used in the industry, it was also found that most provide discrete water velocity measurements. The user would physically measure and record the water depth and then measure the water velocity at that depth. Following the algorithm briefly described in the section above "What is Stream Discharge?", the various recorded values would then be manipulated to provide the discharge at that transect. Our design would automate that process to record the various depths and velocities to ultimately provide the discharge at that transect without the user needing to record and average a number of values at each step.

System Design

The system needed two main sensors for measuring depth of water and water velocity.

The work done over the summer quarter was based partially on the design on the previous team. The "black box" design remained the same because the problem being solved hadn't changed. However, for our own benefit, we essentially started from the beginning in designing the hardware and software. The state machine used in our design was similar to the one developed by Anthony Claudio and Cristopher Claeys.

Sensor Design

Due to time constraints, the propeller implementation for measuring water velocity was preserved. Alternative approaches were identified such as using strain gauges, pressure transducers, and acoustic technologies. The final design for this prototype used a hall-effect sensor to measure the number of pulses generated as the propeller spun a shaft with an embedded magnet by the sensor. The number of pulses per unit time was recorded and a function was adapted to relate that value to a given velocity using a Swoffer Fiber-optic sensor [2] for calibration.

A pressure transducer was used to measure depth. The propeller and sensor array was built into a sliding rail on device. The main supporting rail rests on the floor of the creek, and a control rod is used to lift and lower the propeller housing to the desired depth. At each cross section of the transect, the user would drop the propeller housing to the floor of the creek bed and the depth could be recorded.

Using the differential pressure between the atmospheric pressure above water and the pressure generated by submerging 1/4" airline tubing to various depths that was attached to the opposing nozzle, a differential voltage could be attained. The differential voltages were run through a chopper stabilized amplifier which provided a larger, more usable reference voltage that could be used to map the voltages generated to a specific depth ±1/4".


While there a number of optimizations that could be made to the system, the velocity measured by the device had a margin of error <%2 when compared against a Swoffer fiber-optic sensor. The depth measurements were accurate to within 1/4". The price for materials in the system was under $200. Given that we only had 2 months of classroom time to design, develop, and test this project, we consider it to have been a success.

The improvements needed in the system include some functionality that ensures the user is holding the device perpendicular to the flow of water. Incorrect readings are produced when the propeller is not oriented directly into the flow of water. This could be corrected by incorporating sensors that would prevent measurements from being recorded if the device was not being held in a stable vertical position. The system also lacked persistent memory storage. Once a final measurement was provided, that data was lost when the next transect measurements were initiated.


     (2)  Swoffer fiber-optic sensor

Team Members for Stream Discharge Measurement System

Faculty Sponsors:

George E. Mobus, Ph.D.
Robert Gutmann, Ph.D.