Test Day

In the previous couple of posts I introduced you to micro-hydro and ELCs (Electronic Load Controllers), and the Junior Fellows and their Arduino-controlled ELC design. Now, after lots of preparation, the day had finally arrived when they would be testing their design on a real generator. It was an ominous start to the day as the heavens opened up but this didn't set us back - we made sure that the electronics were not being dripped on and began to set up for the test.

Hujan kuat

The plan was to demonstrate that the ELC design could regulate the generator frequency and then tune the system control parameters.  By adjusting these parameters, the system should be able to vary the amount of power dissipated in the ballast load automatically to bring the generator output frequency to 50 Hz.

The ELC was set up on the Jetty, connected through the window to the generator in the workshop.  In the workshop test-rig, the generator is connected to a turbine, driven by water flow from the pump.

Step one was to use the system's frequency measurement module to read the output frequency of the generator with no load connected. When it was stable at ~50-60 Hz we would connect the system and full ballast load to the generator.  Connecting the load would cause the frequency of the generator to fall but the control system should adjust the amount of load until it is brought back to 50 Hz. Tuning the control system in this case meant varying a constant ('P') which determines the amount of proportional feedback in the control algorithm. We want to adjust this parameter so that it is able to bring the system close to a stable 50 Hz in reasonable time.

This is the generator frequency response that we hope to achieve with the tuned P control

Unfortunately things didn't start smoothly.  The generator was running but wasn't producing the voltage we expected, even with full water flow. After a bit of puzzling, someone finally worked out that some genius had managed to install the turbine back to front!

You can see that the cups on the turbine are facing in the opposite direction to the pipe where the water flows out of - the water was mainly flowing around the outside of the cups rather than pushing them around.

Luckily one of the interested TONIBUNG spectators had arrived with a big bag of kueh pisang (deep fried banana fritters) to keep energy (and cholesterol) levels high.  We made pretty light work of them...

The dogs didn't even get a look-in on these snacks

With the turbine now correctly installed we were able to try testing again. Louis started up the pump and Luno slowly increased the flow to the generator until it was running at full speed. The frequency measurement was reading about 55 Hz so we pressed the contactor to connect the ballast load and system to the generator. As we expected, the frequency fell dramatically before recovering slightly to about 20 Hz. We adjusted the 'P' value and the system was able to bring the frequency up to 30 Hz but no higher - it looked like the 'P' value we were looking for wasn't in our range for this test set-up. This was a quick and easy fix - we just needed to reprogram the Arduino to have a larger 'P' range so we could find that sweet spot.  At this point, lunch time interrupted play.

This is the frequency response which we saw when we tested the ELC.  You can see that the frequency didn't rise to 50 Hz - we need a bigger 'P' value to get there. 

After lunch things really began to get confusing - we weren't able to get the generator to produce the voltage level or frequency which we needed for the testing even though nothing had been changed from the morning set-up.  After much puzzling and testing of hypotheses, we concluded that the generator was damaged.

Lots of cogs whirred but we weren't able to get the generator back running fast enough

Without a spare generator, it was time to end the testing for the day and celebrate our successes. The Junior Fellows had managed to demonstrate that their ELC design is able to regulate the generator output frequency. Although we didn't manage to finish off the test plan to calibrate the control system this time around, that can be done once we manage to sort out the generator.

The testing was a good achievement... but there's more work planned!