Creating and Testing a 1000 l/s Filling System for Water

  Read in   |   Print Print

"During the development the need to communicate with the simulation grew. We can connect the CompactRIO control system to the Deltares' WANDA simulation on short notice using LabVIEW and CompactDAQ. This reduces both development time and the risk of damaging an expensive pump."

- Martijn Smeulers, Deltares

The Challenge:
Developing a complex and reliable control system to fill and to create tide in a 300 m flume, making it possible to create the world’s biggest human-made free surface waves.

The Solution:
Using LabVIEW software and CompactRIO hardware to create a control system containing multiple PID controllers and testing the control system with LabVIEW and CompactDAQ hardware.

Martijn Smeulers - Deltares
Wim Taal - Deltares

About Deltares

Deltares is an independent institute for applied research in the field of water, subsurface, and infrastructure. Deltares conducts experiments in its own facilities for water and soil research to validate models and to test, for example, the optimal design for infrastructure and biochemical concepts for reinforcing the subsurface.

Delta Flume Experimental Facility

The Delta Flume is a 300 m x 9.5 m x 5 m wave flume to test, for example, dikes. At one side we can create a model of the structure that needs to be tested. At the other side a wave machine generates waves up to 4.5 m, which makes these the highest human-made waves. The flume is filled and emptied by using three pump systems providing each a capacity of 350 l/s. Water is pumped from and to a reservoir, which is alongside the flume. Our biggest challenge is creating tide by continuously adding or releasing water from or to the flume. We need a pipe network with nine valves and a control system with three PID controllers around each pump system to achieve high accuracy while doing this.

Filling the Flume

The pipes around the main pump need to be filled with water before the process can be started. A vacuum pump fills the pipe system before starting the main pump. We monitor the water supply to the pump to prevent it from running without water. We also monitor water temperature because high water temperature will damage the pump.

Three processes control when the main pump is active. We control the pump speed by measuring the flow speed. The pump needs to work under a certain pressure, so we use a second PID controller to achieve a higher pressure at the exit of the pump control valve.

The pump also works in a certain working area, which provides water in the range of 120 l/s to 350 l/s. When pumping water from or to the flume to create a tide, we may want a lower flow speed. Possibly water is fed back to the entrance of the pump by controlling a control valve in a shunt to make set points lower than 120 l/s.

Besides the two control valves (shunt and pressure), valves control the direction from reservoir to flume or the other way around to prevent cavitation and for safety.


We kept safety in mind when implementing the system. Of course all safety inputs including emergency stop, pump failure, and pressure failure are monitored on the FPGA, putting all outputs from the CompactRIO system to a safe state. In addition, the FPGA and the real-time controller are monitoring each other. With software failures, the FPGA puts the outputs to the safe state and reboots the complete system. These errors and reboots are logged on the real-time controller.

The FPGA monitors safety, but we also implemented the PID controllers on the FPGA. The real-time controller provides set points and PID gains. The software on the real-time controller can be modified based on an experiment. However, the PID controllers should be less accessible to change, which is the main reason to implement the PID algorithms on the FPGA.

The real-time side monitors all inputs and scales those to a scientific value. It also monitors if there is a user interface online because an operator needs to monitor the process. If it detects that the user interface was idle for some time, it flags the FPGA to go to idle.

User Interface

The user interface consists of a main VI calling different plug-ins. This increases modularity and makes it easy to modify only one aspect of the application. Besides Preferences, in which the operator can see system states and modify settings, there are the generic Filling (pumping from reservoir to flume), Adjusting (pumping from flume to reservoir), and Tide (doing both over time) functions. Most operators are not interested in all the process parameters, so a detailed process plug-in for debugging is available, which shows the complete process.


Three separated PID controllers control the same process, so we needed a good test environment to make sure the controllers did not control against each other and to avoid damaging expensive and difficult to replace equipment such as the main pump. We needed full extensive testing because the many safety components of the control unit became more complex.

Deltares developed WANDA software, which can simulate hydrodynamic pipe and pump systems. With WANDA, users can easily make a schematic of a hydraulic system and simulate the behavior. At the start of this project, we made a model of the pump and pipe system to test how the system would behave. We simulated different scenarios, such as filling with different water starting levels in the flume. We now use this simulation to test the control system. We use a CompactDAQ system to acquire the outputs of the control system (pump speed and valve control signals). A LabVIEW application sends over the different valve positions and the pump speed to the WANDA model through a .NET API. WANDA calculates the actual valve positions, pressure, and flow speed. The results are offered to the control system, mimicking the sensors, through LabVIEW and the CompactDAQ system outputs.

The simulation interface looks similar to the one from the control application. With extra controls, we can force some valve outputs to unexpected states to see how the controller responds. We can also control some error lines during the simulation process.


We used LabVIEW to create a reliable control system in an environment that was already used within the institute for data acquisition, which made the system easy to maintain. This was one of the main goals since we use the flume for different projects. We may need to modify the control system for some projects. An easy-to-understand control system was a solid requirement.

During the development of the control system, the need to communicate with the simulation grew. We can connect the CompactRIO control system to the WANDA simulation on short notice using LabVIEW and CompactDAQ. This reduces both development time and the risk of damaging an expensive pump.

Author information

Martijn Smeulers, Deltares

Wim Taal, Deltares

Bookmark and Share

Explore the NI Developer Community

Discover and collaborate on the latest example code and tutorials with a worldwide community of engineers and scientists.

‌Check‌ out‌ the‌ NI‌ Community

Who is National Instruments?

National Instruments provides a graphical system design platform for test, control, and embedded design applications that is transforming the way engineers and scientists design, prototype, and deploy systems.

‌Learn‌ more‌ about‌ NI