by Alex Padron, VI Form

Last year, Pete Hopkinson and I discovered a shared interest in engineering. We heard about the Class of 1968 Fellowship Grant and decided to do a project, albeit not having any specific idea in mind. Over the weeks after hearing about the grant, we bandied about many different ideas. These projects included everything from working with holograms to generating electricity with tidal power. We came across the idea of a rotary engine in the end because Pete’s dad was working on a rotary engine project for the Navy, and we thought the technology was extremely interesting. We conjectured that we would be able to run this type of engine on hydrogen and spent a week proving that this was the case. Then, we entered the three-week long process of writing the formal grant proposal.

In our proposal, we specified starting with a model airplane engine, with 1.1 HP, and modifying it to run on hydrogen. Our grant included $500 for the engine, $1500 for machining, and $200 for small parts. The plan was to switch the gas inlet with a hydrogen gas feed and machine the inside of the engine to be compatible with hydrogen combustion.  However, without having the engine, we were unable to make a more detailed plan at the time. The process of writing the grant was not done completely by ourselves. We consulted with two professional engineers who have had years of practice in grant writing in order to make our grant clear and effective. We won the grant money for our project, but were unable to work during the school year due to our busy schedules.

During the summer, we met to work on our project in two week-long periods. All of the following work was done at Pete’s house. Our first order of business was to simply run the engine without any modifications. However, this was more complex than we had expected, as there were few instructions guiding how to do so. First, we had to purchase around $200 of accessories: fuel tank, fuel lines, carburetor, muffler, electric starter, glow plug, and glow plug igniter. We also had to build a mount for the engine to keep it still. During this time, we learned the use of every piece of the engine.

At this point, we were ready to modify the engine. First, we discussed with Jonathan Louter, a leading expert in rotary engines, on a specific design we could use. We came up with a list of prerequisites to make the engine run. First, we needed a controlled flow of hydrogen. This entailed a tank and a regulator, and a way to attach them to our engine. We also needed a source of air for combustion. While on hydrogen, the engine runs at a significantly higher RPM than its comparable on gasoline. The reason for this is that hydrogen has an earlier combustion point than gasoline does. Thus, the starter was relatively slow, so the carburetor did not intake enough air. This meant we needed an additional air source to make up for this discrepancy. In addition, we needed oil in the engine. Previously, the model airplane fuel had been at 17% oil level,  and now we were forced to use oil from an outside source. Finally, we needed a way to protect from the effects of hydrogen corrosion. This manifests itself commonly as hydrogen blistering, when H+  ions enter the steel due to intense heat and pressure and then recombine in the steel matrix to form H2.

The design we came up with started with the question on hydrogen flow. In order to calculate the intake of hydrogen, we needed two values, the PSI of the hydrogen and the cross-sectional area the hydrogen was being pushed through. The engine had a needle valve that was used to control the intake of gas when the engine was running normally. We decided that this could be used to control cross-sectional area. To control PSI, we purchased an 80 cubic foot 2000 psi canister of hydrogen, which was around $150, from Airgas. In order to drop the PSI, we had to purchase a regulator made especially for hydrogen, as normal regulators will leak and corrode if used. This cost us over $200, which was a large unexpected payment for us. In order to attach our purchased tank and regulator to the engine, we were forced to use neoprene tubing, as it was the only type that matched with the engine intake. However, the pores of neoprene were much larger that hydrogen molecules. Thus, we used superglue to solidify the tubes and stop hydrogen from leaking.

Our additional air source was comparatively simple. We held a compressed air nozzle to the air intake while starting the engine, and removed it once sufficient RPM was achieved.

To add oil, first we had to calculate the amount of oil the engine used. We contacted the maker, but there were no statistics done on this respect, so we were forced to test the data ourselves. Once we had this value, we set up an oil drip into the engine. This led to difficulty, as oil went through the air intake. Thus, it caused issues when we were starting the engine with the air compressor. To solve this problem, we coated the inside of the engine with oil beforehand and put in the drip immediately after the engine started.

The task of protecting our engine from hydrogen was the most difficult. We spent days researching this topic. Only after reading five studies on hydrogen corrosion did we realize that, in our case, hydrogen corrosion was inevitable. The reason for this was that hydrogen molecules are extremely small. Commercial coatings do exist, but they are all solid, so they would not have worked for us. Without being able to protect our engine, we were forced to find a way to mitigate the effects of hydrogen corrosion. We came up with two solutions. The most expedient way to accomplish this was to run the engine on standard gas. The heat turned H2 into H+ and allowed it to diffuse out of the metal. The other method was simply to remove the carburetor and muffler from the engine. The latter would be a preferred method for most people using this form of engine, as it could be left to sit out of flying season. Hydrogen will diffuse naturally given time. This time varies, but the research we read suggested that there should be at least a 10:1 ratio of time diffusing to time running.

Thus, we had our detailed design. We had all of the components we needed, and none conflicted with the others in any major way. This took us about a week, giving us four days to set up and run our experiment.  We planned on having two days to run our engine, in order to work out problems we hadn’t accounted for. However, we had purchased a faulty regulator, and it set us back a day to get a new one.

On the last day, we set up our experiment, only to encounter disappointment. The reason was that the engine was “knocking”; the hydrogen was exploding before the engine had passed the apex, causing a knocking sound. We hadn’t anticipated knocking because hydrogen has a very high constitution of octane, which means it is not likely to combust prematurely. This octane value, however, is generated by a combination of ignition temperature and compressibility, since both of these affect how likely a substance is to combust. Hydrogen’s octane is extremely high, mostly because it is extremely compressible. However, Hydrogen has a low combustion temperature, which was causing the issue in our experiment. The heat from the glow plug dissipated through the shell of the engine and heated up the intake area to a point that ignited the hydrogen. Then, because hydrogen burns extremely quickly, it combusted entirely before it passed the apex of the engine.

After hours of consideration, we came up with the following plan. We cut off almost the entire supply of hydrogen and pumped as much air into the engine as possible. Thus, we accomplished two things. First, we increased the specific heat of the fuel-air mixture. This meant that it took longer for the mixture to heat up, so the starter would have time to turn the engine  and the mixture was able to burn in the combustion chamber. Secondly, with fewer fuel particles, the mixture burned more slowly. Thus, even if the fuel ignited prematurely, it would not combust completely in the intake. Calculating the exact results of our method was almost impossible and was certainly beyond our scope. Instead, we kept the needle valve at one click open, at the smallest setting, and kept the air nozzle at full power. Then, we tested different PSI of hydrogen to find one that  could yield a result. Eventually, we hit on 15 PSI, and we got the engine started. We then immediately repeated our results to make sure our data was accurate.

Once the engine was running, we were able to play around with different amounts of hydrogen and air. We did not have time to obtain exact results, but we hypothesized that any amount of hydrogen would cause suction that could take in enough air, up to the point that the engine could take the stress.

This project represents one of the many interesting projects available at St. Mark’s. Unfortunately, this grant is only available to juniors, but there are other options for people with interests. The robotics lab is open all year and is stocked with materials from building season. Pete and I had no prior knowledge of anything about our subject besides basic physics and chemistry. Anyone that is interested in this kind of project can work on it, as long as they put in enough effort. It is also important to note that ideas can and will change. In our original proposal, we allocated $1500 for machining, and we ended up not machining anything. However, this was balanced out by around $1000 of costs we hadn’t expected.