Episode Summary:
Thorium is long been considered a safer, better alternative than uranium for fission based nuclear power projects. Molten salt designs promise all the nonproliferation benefits of thorium, plus simpler, more cost-effective reactor designs combined with intrinsic safety. But the Chinese and the US have started work on molten salt reactor projects the promise to take this promising technology out of the lab and onto the grid.
The autonomous driving space is fluid and dynamic, but a few major players are surfacing as leaders in the race to Level 5. Cruz, Waymo and Tesla are the most quoted, but a new partnership between Ford, Lyft and Argo AI to deploy self driving cabs in American cities promises to shake up the race. The combination promises synergies similar to the Tesla approach: Ford supplies the vehicles and service support, Argo AI the court technology, and Lyft provides scheduling, billing and cloud connected data infrastructure. The result may be more than just self driving, but a new way for automakers to monetize Level 4/5 systems.
Access all episodes of This Week in Engineering on engineering.com TV along with all of our other series.
Transcript of this week's show:
To see any graphs, charts, graphics, images, and/or videos to which the transcript may be referring, watch the above video.
Segment 1: With increasing pressure from climate scientists, the Intergovernmental Panel on Climate Change and concerned governments worldwide, nuclear energy is enjoying a surge in interest not seen in decades. Fusion power of course is the Holy Grail, but most experts agree that large-scale decarbonization of electricity generation will require a bridge technology to span the decades until commercial fusion is ready. Several novel fission reactor designs are under development, as are conventional fission plants scaled down as small modular reactors, but the government of China is about to build the first of a commercial fission technology the promises most of the upsides of nuclear with few of the disadvantages: thorium powered molten salt reactors. The concept has been around since the late 1940s, and experiments have been performed on a small-scale since the 1950s, but thorium fueled fission reactors had an insurmountable problem: they’re useless for producing weapons grade plutonium, which in the 1950s was the primary task of US and Soviet reactor research, along with ship propulsion. Of course for civilian use, a reactor design that can’t be used for atomic weapons is a major plus, and there is an added benefit to thorium: high level radioactive waste with a half-life of 500 years, far more manageable than the wastes produced by current uranium fuelled technology. And thorium is relatively abundant worldwide. The fuel is a major part of this technology, but so is the reactor design itself. Molten salt designs dispense with the traditional reactor core with its fuel rods, control rods and mechanical systems to operate them. While the exact layout of the Chinese reactor has not been disclosed, that nation’s nuclear research program has suggested that a novel single loop molten salt design with the thorium fuel dissolved in the liquid metal primary loop would be low cost, operate at atmospheric pressure, burn up bred plutonium for better energy efficiency, and have negative temperature and void coefficients… making them intrinsically safe with simple mechanical systems. From a commercial perspective, a major factor is the ability to operate without large quantities of cooling water, making small molten salt reactors ideal for remote power or process heat generation. The Chinese government expects that the first commercial reactor will be complete in 2030, and that several will be deployed in the plains of central and western China. Beijing has also suggested that the technology might be deployed in nations participating in China’s Belt and Road Initiative. With one atmosphere operating pressures, a simplified heat exchange system and an intrinsically safe physics package, the technology would seem to be perfect for a scalable, deployable clean energy system. One significant technical challenge remains however: corrosion. The molten salts are hot and reactive and to address this, multiple nations are researching advanced materials. Without water as a moderator, hydrogen embrittlement won’t be a factor, although beyond alloy composition, long-term durability of welded joints, pumps and seals is still under research. The materials problem appears to be well in hand, and the Chinese are reporting promising results with a commercially available super alloy, Hastalloy N. Molten salt reactors are under development worldwide and in the US, the Tennessee Valley Authority and Kairos Power will demonstrate a 50 thermal megawatt molten salt test reactor at the TVA’s Oak Ridge facility with an eye toward commercial deployment of a 140 MW electrical plant. Timelines for completion have not been announced. The road to nuclear fusion may run through molten salt.
Segment 2: The automotive industry is changing faster now than at any time in the last century, and that change is happening on two fronts: electric propulsion, and autonomous driving. Both sectors have been characterized by a large number of small-scale startups, multiple different technologies under development, and then eventual shakeout, leaving fewer, more viable companies. Ride hailing major Lyft, for example, recently sold their self-driving technology division, Level 5, to Toyota, and has inked a deal with Motional, a joint venture that includes Hyundai, to deploy fully autonomous cars on the Lyft network in 2023. But Lyft is not leaving the autonomous driving space, as Ford Motor Company announced a new deal with Argo AI and Lyft to commercialize autonomous ride hailing services at large scale. The deal brings together the three core components of a successful commercial autonomous taxi service: motor vehicles, self driving systems, and a system to order the service and process payments. The group plans to deploy vehicles rapidly, starting with passenger rides in Miami later this year and Austin Texas in 2022. Initially, the vehicles will carry safety drivers, and Lyft users inside the defined service areas will hail vehicles by choosing a self driving car on the app. At this point, the group plans to deploy approximately a thousand autonomous vehicles on the Lyft network over the next five years, in multiple markets. Lyft software is mature and systems are already in place, and of course Ford has over a century of auto production experience, so the question is Argo’s contribution. As part of the deal, Argo will have access to anonymized service and fleet data from Lyft, allowing the company to more rapidly develop code and deal with the major limitation on current AI driven self driving systems: edge cases. For its part, Lyft will receive 2 ½% of the common equity of Argo AI as part of a licensing and data access agreement. Ford is establishing centers in Miami, Austin and Washington DC to support the vehicles, including fueling servicing and cleaning. Collaborations between software companies and automakers are not new, but the critical takeaway from this deal is that Lyft is monetizing the data collected by their large fleet of vehicles. Testing is always a major problem when developing safety critical systems, something that Tesla has addressed by uploading user data from their customers’ electric cars. In theory, this should remove a major obstacle for Argo AI who can combine simulation with large-scale real-world testing. For Ford, it establishes a market for turnkey taxi vehicle services, allowing companies like Lyft to offer services without developing expensive hardware infrastructure. With urban self driving systems now carrying passengers in the US and China, to date there appears to be widespread consumer acceptance of the technology. It seems clear that self driving will begin in geo-fenced, urban and suburban areas before moving nationwide. We’ll report again when the service is up and running.