The story resumes around 8 pm on day 6, Saturday. After a challenging third run was completed without a score improvement, the team debated the next move. New tyres should have made a noticeable improvement, but didn’t. Were the tyres cancelled out just by the current-guzzling hill starts in the lap 2 traffic jam, or did the higher gear turn out to be counterproductive? It was easy to predict the effects of a higher gear on the hill performance, and separately on the downhill and relatively flat sections. But it was very hard to reason about the net effect on a whole lap.
Meanwhile, briefly and reluctantly, the team were drawn from the paddock by the smell of beer and pizza at the SEM Saturday night party. On their return, blood sugar and morale restored, they began an assault on the car’s weight. The inaccurately named “seat,” actually a hard horizontal panel, was removed with the blessing of our long-suffering drivers. Minus 0.8 kg. One of the former seat’s three supporting aluminium crossbeams was cut out. Another 1.5 kg? Data acquisition electronics were removed – they are essential for analysis and learning after a test, but of little use while driving. And so on. Take care of the g and the kg will take care of themselves, as they allegedly say in Formula 1. One team member voiced grave concerns that the car might float away if it wasn’t tied down. We didn’t have the means or inclination to weigh the jettisoned material, but all told we may have shed 5% to 8% of the car’s 68 kg. Hugh and Shane processed data from both complete runs until they were thrown out of the paddocks at 11 pm, looking for clues to guide the choice of gearing for the next run, even considering a change of motor. Back at the campsite, they went on into the small hours.
Sunday morning was still and almost cloudless. The final prototype session was set for 9.00-12.00. The decision was to revert to Friday’s drivetrain setup. From that baseline, there should be a clear step forward from the new tyres, combined with an evolved strategy to manage momentum, and increasingly tactical and precise driving. We wouldn’t risk the uncertainty of a gear change in either direction, much less a change of motor. Niamh suited up and rolled out.
By the time she reached the first corner, she knew there was a problem. The car’s acceleration and speed were low. It crept up the hill, but the speed was far too low to even make the necessary 25 km/h lap average. Our fourth and final attempt looked like it was dead. The trackside team conferred with the driver over the phone link. They knew from the morning race briefing that there could be a reprieve – fifth attempts might be allowed, time permitting. Time wasn’t co-operating: there was about half an hour to the latest permitted start. Even if run 4 could somehow be completed in the regulation 43 minutes, there would be no time for another.
They made the call to abort the fourth run and try to debug the car in the hope of a fifth attempt. Niamh peeled off into the exit lane before completing lap 1 and was immediately pushed around towards the starting queue, bypassing the paddocks. Barry and Sorcha brought tools, cracked open the power electronics, and began to probe. The fault was traced to the accelerator pedal. It should deliver a signal ranging from 0 volts with the foot off the pedal to 5 V when fully down, but the maximum signal was now less than 2 V. When Niamh floored the pedal to command 100% voltage, the controller thought she was commanding 30 or 40%. The electronics team spoke urgently of noise and isolation, as they often do, but there was no time to investigate the cause of the problem. Sorcha jacked into the controller’s Arduino microprocessor, and edited a line of code to rescale the accelerator input. We lifted the back wheel off the ground to check that the power system could spin the rear wheel, and spin it against a lightly applied brake. Then a quick brake check on the test ramp, because brakes. Then we were back into the line-up. The whole operation took less than 10 minutes and Niamh hadn’t even got out of the car.
At last, a drama-free run. The somftware hotfix worked and the unexpected fifth attempt was flawless. It seemed routine now to manoeuvre around traffic on the hill and play dynamic chess for the best lines on corners, all in the most graceful way, marshalling the car’s precious energy. Niamh came home on time and the score was reported as 236 km/kWh – an improvement of 50%. With minutes left in the session, we waited to see if any late runners would overtake us. Nobody did. Our overnight score would have left us at 27th, but we’d jumped to 21st. Without wasting valuable time, we moved into the celebration phase. That, however, is beyond the scope of this report. Then it was time for the inevitable packing up. This seems like the moment to thank the formidable Kieran Flannery of Quicktec IT, who provided and drove the van that carried the car and tools, along with much of the paraphernalia that made our campsite civilised.
Geec 2.0 had completed its journey. Our final ranking, 21st of 50 competitors in battery-electric prototype, was up slightly from 23rd of 51 in Rotterdam last year. Our score was a little lower, but so were the top scores on the tough London course. Shell Eco-marathon is engineering as a full-contact spectator sport, even more so than F1. When the trophies are presented at SEM, the whole driving, engineering and business team come on stage as equals. It’s not surprising that grown men and women cry on the trackside, for a whole range of reasons. As well as metal and plastic, the cars on track are made of hours and hours of mental and physical labour.
In the spring of 2014, we exhibited the skeleton of the Geec 1.0 with a frisbee to indicate where the steering wheel would one day be positioned. That car went on to become the most efficient car ever built in Ireland and made a proud debut for Ireland at SEM. In autumn of 2015 we began to design the Geec 2.0 from a blank sheet. It will go further than its predecessor. It’s the platform from which we can move into the top international division. Its bodywork is as good as almost anything on the London track, in aerodynamics and structure as well as looks. The chassis is light and strong. The shape and weight distribution give the car good handling at speed, and the power electronics have been proven robust under the toughest conditions.
We could not have done this without our many supporters and partners. Thank you to Shell E&P Ireland, ÉireComposites, ARD Precision, Caulfields, Bahco, RMS, Blackstone Launchpad, CADFEM UK & Ireland, ANSYS, GE, Pat Rynn Engineering, Mondello Park, Easy Composites, Enform Plastics, and Quicktec. Our home, NUI Galway, has put its resources behind this project from day 1, and we thank the many people there who have helped in diverse ways. Galway University Foundation has provided strategic funding to help the Geec to become sustainable, through the Ryan Award for Innovation.
The core of our team will scatter soon to start professional careers or further study, or just take a well-earned break. In September, a new design team will assemble. Some have been contributing to the team for a while and there will be some we haven’t met yet. Geec 3.0 will probably look like Geec 2.0 on the outside, but under the lid will be a better machine. By midnight on Sunday the campsite conversation was already turning to plans for next year’s car. Watch this space.