Metrolinx has released a long study about the feasibility of using electricity generated from hydrogen fuel cells as an alternative to conventional railway electrification with overhead wires. The “Hydrail” project page contains links to both a quicky “fact sheet” and to a 353-page report. The report itself contains a 13-page Executive Summary giving a high level view of the proposals and recommendations without much of the technical detail.

It is impractical here for me to review the entire document, and indeed this is not really needed because a great deal of the content is a tutorial on hydrogen technology. The report is clearly written by people with more of a background in hydrogen technology and marketing than in railway planning and operations.

Fascinating though this is, the report does not address the most crucial issue of all – what are the implementation scenarios for hydrogen propulsion depending both on technical maturity and on policy decisions still to be made about the evolution of the GO Regional Express Rail (RER) system.

A great deal of confusion lies in the process Metrolinx is following to provision RER. Their intent is to farm the entire thing out to a private consortium:

Design-Build-Finance-Operate-Maintain (DBFOM) Procurement Process

Metrolinx is intending to engage a contractor to upgrade the GO network using a Design-Build-Finance-Operate-Maintain (DBFOM) model. As part of the tender process, bidders will be able to propose both hydrail and overhead wire technology to electrify the GO network. The benefit of this DBFOM approach is it allows one single party to manage all the interrelated decisions necessary and oversee each phase of the process from design to maintenance. This ensures optimal performance is achieved for the entire system, which can create efficiencies. [Website]

However, as the industry now stands, the information needed to allow an informed assessment of technical maturity, feasibility and risk for hydrogen trains at the scale of a GO/RER implementation does not exist. There is a lot of speculation, but it is based on much, much smaller and simpler implementations of various aspects of the technology.

The intent of the proposed study is to acquire as much information and experience as possible so that bidders can bid intelligently. The real challenge will be for this to happen before the Request for Proposals is issued at the end of 2018.

There is a subtle change in the text above to statements by Metrolinx CEO Phil Verster in 2017 when he said that it would be up to bidders to decide which technology they would choose to offer. Instead, the description above states that bidders can propose either technology and it would be up to Metrolinx and the Government of Ontario to decide which version to implement. It is quite likely that for the riskier new technology, bidders will be less willing to accept broad technical risk, and they will charge a premium for this. Whether the government of the day will see any extra costs as worth the investment remains to be seen.

Indeed, although the report states that the Cost:Benefit ratios for conventional and hydrogen options are similar, there is no mention of the risk premium a bidder might place on one option over the other. Moreover, the actual calculation of the ratio is not explained, nor are the total costs given. This raises the question of whether a higher cost is offset by a higher assumed benefit so that the ratios come out similarly, even if the magnitudes of investment differ.

At a recent Board of Trade appearance, Verster was asked about electrification, and replied with praise for Ontario’s “hydrogen economy”. It is quite clear that he drank the Kool-Aid and the government’s usual fascination with technology is getting in the way of his proper role as CEO. Immediately afterward, he reverted to the position that it is up to the would-be builders/operators of the RER network to propose technologies and the risk they are willing to assume.

Later the same day, when asked at a Metrolinx Town Hall about the possibility that hydrogen efforts would delay electrification, Verster replied with the standard response that the vendors will decide. However, the timelines for investigation of hydrogen and the contract award date suggest that a lot of work will be jammed into a very short period, and that Metrolinx’ own technical investigations will overlap the bid process.

A fundamental problem with Metrolinx “benefit cases analysis” (also misleadingly termed “business case analysis”) lies in the calculation of presumed benefits which are built up from a variety of factors. These include not just direct spending, but also the imputed value of effects such as reduced travel times, reduction of congestion and the value of environmental improvements. This side of the analysis is not present in the report, and so it is difficult to ascertain the “benefits” against which each scheme is measured. As for costs, so many elements of the hydrogen train proposal are little more than assumptions about the scalability of existing technology, it is hard to believe that the cost estimate is much beyond the back-of-an-envelope stage.

The capital and operating cost estimates presume a level of certainty about the hydrogen option which simply cannot exist at this point. Indeed, a major purpose of the planned work is to provide the technical basis on which a bidder might construct a proposal. Some capital costs included for conventional electrification are not included in the hydrogen scenario, and there is a wide variation in the range of projected operating costs.

With a planned launch of RER by 2025, the timelines are quite tight because major decisions on the infrastucture needed for either alternative must be made soon so that RER is “ready to roll” when planned.

Notable by their absence are key pieces of information:

• What is the relationship between the timelines of the proposed hydrogen investigations and prototyping, and the timespan of the DBFOM procurement through all of its phases from initial tender up to revenue service? Can the research phase be completed in time to inform bids from potential builders/operators of the GO/RER network?
• If the DBFOM bidders depend on investigative work done by Metrolinx or others on its behalf, what liability will Metrolinx have for non-performance if their work turns out to be incomplete or faulty, and therefore prevents the successful execution of the contract?
• What is or will be the position of the railways, CN and CP, to the presence of hydrogen trains on their systems? Their dislike of electrical distribution and overhead structure in their territory is cited as a benefit of the hydrogen alternative, but one must ask how the railways will view the risks of a new propulsion technology co-existing with their operations.

This brings us to a fundamental question about RER and electrification, regardless of the technology. At the risk of being accused of environmental insensitivity, it must be said that electrification is not a prerequisite for RER implementation at the service levels now planned. Indeed, electrification makes the system design more complex especially where GO services operate over other railways’ territory. The tradeoffs are between many issues including the increased intrusion of more frequent GO service in corridors now hemmed in by residential development rather than by industry. This brings noise and pollution from frequent service with diesel locomotives. Even electric trains are not silent.

Reading between the lines, one might well think that full electrification is now contemplated as something for the future, in the mid 2040s, not in the 2020s. This is fundamentally tied up with questions of implementation and roll out, none of which is addressed in the report because it assumes this is a matter for future study.

Although much discussion reads as if RER will appear overnight in January 2025, Metrolinx plans to begin building up service levels from current to the RER proposal on an incremental basis as infrastructure improvements are completed. This means that a substantial portion of “RER” based on existing technology would exist before electrification, by whatever scheme, actually is “turned on”.

An important part of any implementation plan will include the mechanism by which a DBFOM bidder will take over existing assets, and this necessarily must be spelled out as part of the tender process. This will lead to two huge transitions occurring in parallel: the move from direct Metrolinx capital and operating responsibility for the GO system to a separate provider, and the technology transition from diesel to electric on some or all of the network. Whether Metrolinx has the capability to manage something on this scale, or will simply dump the responsibility in the provider’s lap and hope for the best, remains to be seen.

There is also the fantasy that the “risk” will be transferred from the government to the provider, but that risk comes at a price, and what is effectively “risk insurance” usually has a cap. Examples of capped liabilities, or even of providers walking away from their responsibilities, are not hard to find. Of course there could be problems with conventional electrification too, but they are less likely with a mature technology.

In this article, I will review the recommendations so that readers who want the “short version” can get my opinion without reading all the way to the end. In a separate future article, I will turn to specifics in the detailed report.

Recommended Next Steps

The report includes recommendations for a great deal of further study. Despite the optimism about the future of hydrogen trains, the breadth and depth of this list shows that a great deal more information is needed including operational experience from a prototype of a locomotive and of systems needed for hydrogen production and fuelling. All of this would have to occur before a bidder would be able to incorporate this experience into an RER DBFOM proposal.

• Design:
  • Complete the projects to create conceptual designs for an HFC bi-level EMU and an HFC locomotive.
  • Refine the Hydrail System configuration and size in the Operational Simulation model, including development of concept designs for hydrogen production, storage and fuelling subsystems.
• Prototyping: Commission the production of a prototype HFC locomotive that can enter revenue service, including the development and prototyping of the refuelling and hydrogen production subsystems that can work with the prototype HFC locomotive so that Metrolinx can learn valuable lessons concerning the Hydrail System’s operations, performance, and reliability.
• Railway operations: Further investigate the operational areas of the Hydrail System such as maintenance and refuelling.
• Cost: Recognizing that the Hydrail System would represent a new approach to delivering RER, collaborate with industry vendors to further investigate infrastructure and vehicle delivery and operational costs.
• Implementation: Further define the development and build phases and the transition plan required to initiate a Hydrail System on the GO network.
• Hydrogen production: Identify location options for the hydrogen production facilities.
• Regulations: Work with the safety regulators at the federal and provincial levels to reach clarity on the regulatory environment that will apply to Hydrail.
• Electricity price policy: Work with the provincial government to develop an electricity price policy that could be applied to the Hydrail System.
• Hydrogen economy: Work with the provincial government to develop a cross-government business case for hydrogen (including the Hydrail System).
• Align with the RER procurement process: Integrate the outputs from these steps into a Hydrail System Reference Concept Design that can be used in the RER procurement process. [p. 21]

Design & Prototyping

There are two basic types of trains using electric propulsion. One of these is similar to existing GO operations with a locomotive hauling a train of unpowered coaches. The other is a multiple-unit operation where each car is self-propelled. The best known example of this in Toronto is the subway system.

All locomotives use electric motors for actual propulsion, but the source of the power varies with the technology:

• An onboard diesel generator (existing trains)
• Power drawn from overhead wiring (conventional railway electrification)
• Power generated onboard using fuel cells

In the case of multiple-unit cars, the options are similar:

• An onboard diesel that directly drives the train (such as the RDC cars once seen on the VIA rail passenger system)
• Power drawn from overhead wiring (or third rail, typically on subways)
• Power generated onboard

These are referred to with the acronyms DMU, EMU and HMU (for diesel, electric and hydrogen) respectively.

The basic problem for any onboard power generation is that the train or cars must have space to house the generation system including fuel for a reasonable period of operation (there is no point in having trains that must be out of service often for refuelling). Diesel locomotives already are designed to carry their own fuel and generation equipment, and transition to hydrogen electric operation depends only on whether the new technology will fit in a comparable space. Coaches are a different matter because the unpowered version has no need for such gear, and conventional electric MUs, like subway cars, draw their power from an external source. Integration of onboard power generation is a net new requirement for space, and it will detract from passenger capacity.

Power does not simply flow out of the fuel cells to the motors on an electric train the way it does from an overhead wire. Hydrogen (which must be carried as fuel) is converted by the fuel cell to power, but the nature of this process does not allow large variations in the power draw, unlike an overhead-based system. To support the demands of fast acceleration, the fuel cell’s power is stored in batteries (which themselves have limitations on current draw) and supercapacitors which can deliver the needed burst of power. The capacitors are also used as storage for power from regenerative braking. (Battery buses use a similar scheme, but they have no on-board power generation relying instead on periodic recharges along their route.) On board generation adds weight that must be carried with the train.

The Alstom Concordia iLint is often cited as the beginning of hydrogen-based rail operations. It should be noted that this evolved from Alstom’s Lint product which is a DMU. Unlike the GO Transit bilevel coaches, the Lint already includes a propulsion system.

Another issue for HMUs vs locomotive hauled sets is that each HMU must have the capability of independent operation. This results in duplication of subsystems, not just for energy generation but also for train control, on each car rather than sharing capacity for the trainset. EMUs have the advantage that power supply is not on board, but they still require duplicate systems for each unit if they are intended to run on their own.

An analogy to urban transit vehicles shows different implementations:

• The newest subway trains in Toronto (on Line 1 YUS) come as a set of six cars, and these cannot operate individually. The tradeoff benefit is that some subsystems and components are shared among cars in the train reducing its overall cost.
• The older cars (on Line 2 BD) are built in pairs which share equipment and can operate individually. This allows them to be made up in trains of 2, 4 or 6 cars.
• Streetcars/LRVs are intended to operate on their own, although train operation is possible (Line 5 Crosstown is designed for trains of up to 3 units). Every car must have a full set of gear, and this makes the unit cost of the vehicle higher than would be the case for permanently linked sets. The cost offset lies in the simpler infrastructure an LRV would need and its design for operation on surface rights-of-way.

In the case of railway operations, the fourth option familiar to everyone is to have a completely unpowered train with almost all major subsystems consolidated in a locomotive.

A hydrogen-based system does not require the construction of a power supply infrastructure (overhead, feeders, substations) along a route right to the last mile of service. The attraction lies in the avoidance of this infrastructure and the ability to electrify in places where this might not be economically feasible with conventional technology or because of constraints by the owning railways.

Metrolinx proposes to complete work on design both for a locomotive and for an HMU, and to commission a prototype hydrogen-based locomotive.

HFC Locomotive Pilot

Currently, GO rail vehicles are locomotives that are powered by diesel engines. An HFC locomotive would be an electrically powered locomotive where the electricity is generated using hydrogen that is stored on-board the train.

Metrolinx plans to commission concept designs for an HFC locomotive and will then consider building a prototype that could be introduced into service on the GO Transit Network. The prototype would gather valuable feedback on operations, performance, and reliability.

EMU Design Concept

An electric multiple unit or EMU is a multiple unit train consisting of self-propelled carriages, using electricity as the power that drives the train. An EMU requires no separate locomotive. An HFC EMU is an EMU where the electricity is generated using hydrogen that is stored on-board the train. Ontario is engaging with train manufacturers Alstom and Siemens to produce concept designs that incorporate hydrogen fuel cells into bi-level trains similar to those currently used by GO Transit. [Website]

GO Transit bi-levels are larger than the Lint rolling stock and will require more power to operate. This begs a tricky design problem in that passenger space could be lost to the propulsion and fuelling systems partly undoing the benefit of the large GO coaches. One could operate powered and unpowered coaches in a multi-car consist, but it does not take many cars to reach the point where one has, in effect, reinvented the locomotive with a single powered car hauling the others as trailers.

The question then would be why bother developing a specialized powered coach when one could simply use a locomotive. One tradeoff is that depending on the demands placed on a locomotive, it might provide poorer performance because the tractive effort can be delivered only on its wheels, not on every car in an MU train. This is an important consideration for quick starts from closely spaced stations.

Although not explicitly stated in the report, this debate seems to have already been answered in early thinking about hydrogen trains on GO Transit with the conclusion that locomotives are the preferred implementation hauling a consist of six cars. Twelve car sets would be created simply by coupling two trainsets back-to-back. This would allow GO to electrify system-wide while continuing to use their existing passenger cars rather than transitioning to a completely new fleet.

Railway Operations

An important factor in designing a commuter rail operation is to determine the cycle of refuelling and the locations where this will be done. On the existing GO network, according to the report, there are three fuelling locations: Mimico Yard, Barrie and Lincolnville. The actual fuelling cycle varies depending on the service pattern. For routes where most trains operate only in peak periods, they can be refuelled either overnight or during the midday break, provided that facilities to do so exist where they spend their midday layovers. However, for routes where most trains are out all day long, the period available for fuelling is much shorter.

Based on observed operations at Mimico Yard, the time needed to fuel an empty diesel locomotive is about 23 minutes, and any hydrogen alternative must meet or better this rate. The problem becomes more complex for HMUs where each unit must be filled individually (or several coupled units in parallel). Moreover, as the GO fleet becomes larger and service levels rise, it will not be practical to fuel all trains in only a handful of locations.

This is an activity that only exists because the hydrogen trains must carry their own fuel.

This brings us to the problem of getting fuel from hydrogen generating plants to the trains. In the simplest configuration, there would be a relatively small conversion unit on each corridor at yards or online storage locations. However, the report discusses other options such as a large-scale central generation unit linked to fuelling sites by pipeline or with tanker trucks carrying either gaseous or liquified hydrogen.

One important factor in the hydrogen proposal is the ability to use cheap off-peak electricity to drive electrolytic conversion. The amount of electricity required is not trivial, and the desire that most generation occur overnight concentrates the draw on the electricity network. Ideally, hydrogen production should occur as close as possible to the major sources of power (transmission lines) rather than further away, but this does not always fit with the topology of the rail network.

Much of the technology for generation already exists, but it is the scale of potential GO requirements that presents challenges.

In the case of conventional electrification, the power requirement is distributed over the day and coincides with the period of peak demand. This is an added cost, but the tradeoff is that the power supply can be granular tapping into the Hydro grid at various locations along corridors, and the conversion losses for direct use of the power by the trains are much lower than for the hydrogen alternative.

Maintenance facilities for hydrogen trains would have their own special requirements for safe operation, and this would vary depending on whether only the locomotives carried fuel, or if GO operated a fleet of HMUs each with its own onboard system. The new facility in Oshawa was supposed to have been designed with future conventional electrification in mind. The degree to which it can be easily adapted for hydrogen locos or HMUs is unknown.

The report discusses possible train consists, and each brings its own operational requirements and challenges:

• For locomotives, the report prefers a configuration where each loco hauls a six-car set of coaches. This reduces the power requirements for individual locos, and a typical train configuration would be two sets coupled back-to-back.
• For HMUs, trains could operate in sets of varying length because there are no unpowered cars (although one could also design shorter trainsets sharing the fuel storage and power generation facilities).

From time to time, one reads of “savings” from operating shorter trains off peak. This is not as straightforward as it seems because units that are taken out of service after the AM peak must be stored somewhere through the day and then be reunited for operation as longer trains for the PM peak. This creates a need for extra labour, careful scheduling, and many additional movements of units to and from storage areas that have to fit into the overall operational plan.

By analogy, the TTC stopped cutting subway train lengths in the evenings years ago because the overhead cost of this activity greatly exceeded the benefit in power and maintenance savings.

The report speaks of the need to optimize train operations in light of fuel consumption and capacity. Unlike a diesel engine with a large fuel tank or a conventional electric train with easy access to power, the hydrogen trains must be designed around a duty cycle where excess fuel consumption is minimized. This could impose operational requirements that do not exist with conventional technology.

Implementation

A key issue for implementation of a hydrogen train system is the short timeframe between the research period, the contract award to a provider/operator of the GO/RER network, and the risk that the technology will not be sufficiently advanced.

HFC locomotives and EMUs of the type needed for RER do not yet exist; therefore, real-world experience relating to reliability is limited to the operation of light rail vehicles and buses. This means that even though the vehicles will go through an extensive development and testing process, there is still a risk that when they enter revenue service, they will experience unexpected issues with reliability that cause in-service failures and consequential impacts to customers.

This risk also applies to the hydrogen production, storage, refuelling, and dispensing facilities; and it is probable that the greatest risk relates to the integration of these systems. However, this risk can be mitigated by implementing and operating a small-scale prototype of the end-to-end system so that lessons can be learned about the system’s operation, performance, and reliability; and so these lessons can be fed back into the design, build, and operation of the full system. [pp 17-18]

Prototyping is a wonderful idea, and proving the entire system from creation of hydrogen through transportation and storage, fuelling and train operations on a modest scale is quite a reasonable approach. However, this runs headlong into the 2025 implementation date for GO/RER services.

There is a related problem in system design in that if fuel generation is to take place on a distributed basis with modest sized installations at key locations around the network, this is at a “prototype” scale. If the approach were to be a very small number of large generation plants plus a distribution network, that is quite another scale, one which does not lend itself to prototyping.

GO Transit’s fleet plans have, until quite recently, been based on the assumption that electrification would come in on a few major lines, and that diesel operations would remain on a good-sized chunk of the network, trains serving the territory GO does not own. This would allow existing equipment, notably locomotives, to remain in service to haul diesel trains while service grew into the RER model with electric operations where this was feasible. If the desire now is for a “big bang” implementation with electric operation over the entire network as soon as possible, this would leave GO with equipment they could not use. If the new trains have a lower capacity than existing equipment, this will affect the number of trains GO must operate to serve current and future demand.

If one considers hydrogen as a viable technology choice, the up side is that areas now considered unsuitable for electrification can be converted much sooner as several issues with overhead-based systems are avoided:

• Hydrogen trains can run over existing trackage with no need to modify it as the electric ground circuit, nor to modify signal systems to co-exist with electric operation.
• Track that is subject to flooding (the Richmond Hill line) no longer presents electrical problems if the track is underwater.
• Power distribution no longer requires construction of overhead systems along the rights of way, and the power conversion process exists only at discrete hydrogen conversion plants.

Moreover, there is an assumption that the freight railways would be more amenable to the presence of hydrogen-powered trains on their lines than they would of conventional electrification.

In an ideal world for hydrogen advocates, the entire system would switch over quickly, but it is hard to see this happening given the technology hurdles and GO’s long-standing tendency to change incrementally.

Cost and Power Availability

The report claims that hydrogen trains have a benefit:cost ratio that is comparable to conventional overhead power distribution. Considering the number of as-yet uncertain factors, this requires a leap of faith that the estimates used to build up the cost base are valid. There is also the long-standing problem of determining just what a “benefit” might be and how to value this although, to be fair, the same issue exists for conventional electrification too.

An important factor in the overall calculation is the question of power cost and the timing of consumption. Hydrogen generation would use off-peak power, mainly overnight, that is thought of as sitting there for the taking. With the rise of other energy storage systems, notably battery-based vehicles, that surplus might not be as big as is thought, and this could affect future pricing.

The timing of demand also affects the rate of draw on the electrical grid because all of a day’s power consumption would occur in a shorter period, and there is the added penalty of a lower energy conversion efficiency through the chain of electrolysis, storage and transportation of gas, fuelling and on-board power generation. The claim is made that the projected requirement is a few percent of Ontario’s power consumption.

To put this in context, the estimated power requirement is 2.2 to 2.5 gigawatt hours (gWh) per day, and scaling up on the basis of a 300-day year (counting weekends as one “day”) gives a total draw of 750 gWh. The entire TTC system, primarily the subway, consumes 450-500 gWh per year, and this will grow both with increased service on the electrified modes (subway and streetcar) and with the TTC’s stated desire to move to an “emission free” fleet with battery buses.

(The TTC figure is based on a report of bulk power purchase contracts which included the cost per megawatt hour, the total contract prices and the percentage of total demand represented by each contract. Although the report is a decade old, TTC power requirements have not changed much over the intervening period.)

One section of the report, there is a discussion of the amount of surplus power expected to be available over coming decades from the mismatch between generating capacity that cannot be throttled and changes in demand [p 33]. However, the situation is not as straightforward as plugging a hydrogen-based system in to the grid to collect all of this excess power.

The report contains a long section detailing the evolution of Ontario’s power generation infrastructure and the degree to which there is available off peak power, also known as Surplus Baseload Generation (SBG). A key section notes:

The assessment shows that, while the level of SBG is expected to decline, electricity production is expected to exceed Ontario net demand by 1-4 percent from 2024 to 2035. There are however seasonal variations in the level of SBG and, particularly given the increasing reliance on renewable energy in the province’s energy supply mix, weather related factors can further impact electricity generation, and therefore SBG. SBG is often available over short periods of time and to fully capture it would require an infrastructure build-out that would allow for the production and storage of hydrogen over short periods of time, which would be idle outside of these periods.

The ability to harness SBG to produce hydrogen is a function of the scale of hydrogen production (current density) and storage (volume) infrastructure, relative to when the SBG is available. Using hydrogen production and storage assumptions from the Hydrail simulation model, along with demand projections as presented in the 2016 OPO, modelling work shows that only approximately 14-35 percent of the annual amount of hydrogen required to support Hydrail can be produced fromelectricity supplied during periods of SBG over the 2025-2035 time frame. Furthermore, the same analysis shows that only 6-23 percent of operating days over the forecast period could be fully supplied by SBG.

A further consideration is the province’s plans to incorporate energy storage technologies in managing the electricity system. The use of such technologies will lead to a further reduction in excess supply, as surpluses would be at least partially captured and stored by the province, thereby reducing the amount supplied to the market. The use of energy storage technologies would limit the ability of market participants to capture surplus baseload generation.

Therefore, it is unlikely that the Hydrail System will be able to operate solely on electricity produced during periods of SBG. Accordingly, while the system would be able to avoid Ontario peak demand hours, therefore producing during off-peak hours, the system as currently designed, will not be able to consistently minimize the price of electricity, as it will be limited by the density of the hydrogen production facilities and the capacity of the hydrogen storage facilities. [p 145]

There is a further discussion of the pricing of power and the degree to which a Hydrail system would be subject to the “Global Adjustment” (GA) which is a charge on all power users for the surplus electricity that is generated but not used.

[…] the Ontario grid experiences very substantial surpluses of green electricity every day. The cost of this unwanted electricity is recovered by the Global Adjustment (GA). The cost of electricity for either Hydrail or track electrification is crucially important to the choice between the two technologies. The entire premise of Hydrail is that hydrogen will be generated by accessing this unwanted electricity. In storing electricity as hydrogen, it will time-shift energy demand by tapping into the unwanted surplus. Consequently, it is our understanding that electricity for Hydrail would not be subject to the GA. [p 110]

However, this scheme requires that the hydrogen conversion is capable of operating from time ot time through the day whenever there is surplus power to be absorbed, and that this power can actually be delivered to the generating sites. This also affects any decision on the location of the conversion sites and storage facilities for hydrogen.

[…] in determining whether to take a centralized or decentralized approach to hydrogen generation, careful consideration should be given to load constraints within different zones of the electricity system, and the implication for the ability to operate electrolyzers without any destabilizing effects on the electricity grid. [p 148]

Due to constraints in the power distribution grid, a decentralized scheme is preferred in the report. The following text describes capacity challenges within some parts of the power grid.

At present, decentralized hydrogen production to support Hydrail is being considered. Relative to centralized production, decentralized production would allow for smaller loads being drawn across different zones of the grid, from production facilities located along each line of the GO network. There are pros and cons to decentralized production, depending on the capacity constraints within the zones being contemplated and the location of potential hydrogen generation facilities relative to the GO network. Such matters would be resolved through detailed planning and procurement activities.

The GO network extends across the East, Essa, Toronto and Southwest zones. In the East and Southwest zones, generation either exceeds or meets peak demand, with generation capacity expected to increase by 2020. Both zones are generally uncongested, and should therefore allow for additional loads. The Lakeshore East line extends into the East zone, while the Lakeshore West and Kitchener lines extends into the Southwest zone.

In the Toronto and Essa zones, resources are less than peak demand, and both zones are load congested. The situation within the Toronto zone is expected to be magnified by the anticipated decommissioning of the Pickering nuclear generating plant by 2024.

All GO lines run through the Toronto zone, however there may be locational challenges within the zone (space, property zoning, environmental, price), which may be compounded by limited public acceptance, particularly given the urban setting, as discussed in Section 4.8. The Barrie, Stouffville and Richmond Hill lines all run through the Essa zone.

A decentralized approach to hydrogen production (described in Section 4.1) would involve the establishment of generation facilities along each line of GO network, crossing each of the four zones discussed. The significant advantage of decentralized generation is the cost advantage of smaller facilities along each line of the GO network. Based on the foregoing, there may however, be other challenges associated with this approach. One significant challenge is the limited capacity within Toronto, a situation that is expected to persist for the foreseeable future.

A similar challenge would also be faced in the Essa zone. Further to this, consideration should also be given to the competing demands of electric vehicles, the continued adoption of which the 2017 LTEP acknowledges could have a significant impact on the province’s distribution network. Should hydrogen generation facilities be established within these zones, careful planning would be required to ensure hydrogen generation does not inadvertently stress local grids. [p 146]

It is quite clear that dropping a major new demand onto the power grid will have its effects especially in a market where some generation capacity is to be shut down, and where there are other potential customers for any available electricity. The grid capacity problems are particularly important because the power requirements for Hydrail (and other schemes) are often cited as being a few percent of provincial annual power generation. However, if that power cannot be delivered when and where it is needed, generation capacity is of little value. (A similar problem exists with a shift to imported hydro from Quebec where the amount Ontario could used is constrained by the grid’s capacity and orientation to existing generating plants.)

A discussion of the quality of planning and investment that have gone into Ontario’s energy network is beyond the scope of this article, but this is an important consideration for any future shift to electrical power. The additional “cost” in power due to the much lower efficiency of hydrogen power storage and conversion as compared with direct power supply by overhead must be taken into account. Although the power distribution infrastructure from GO’s point of view is simplified, this could be at the cost of requiring more power from the overall grid.

The recommended next steps are more optimistic than the discussion of power availability would suggest, but they include the possibility that a special pricing arrangement would be created to insulate GO/RER from future price shocks (see excerpt from the Executive Summary below). This might make GO’s economics (or those of its DBFOM provider) look good, but this would shift costs onto other power network customers while preserving an advantage for the hydrogen option that might not otherwise exist.

Hydrogen Technology

The basic challenge for any electrically powered system is to deliver energy that may start out as a waterfall, a nuclear reaction, wind or sunlight into a form that can be available when and where needed. Storage of this energy can be challenging on a large scale. Everyone knows about batteries, but this technology has only recently scaled up to the level of powering large vehicles, and even then there is a problem of capacity and “refuelling”.

The combination of hydrogen gas and fuel cells offers the ability to convert electricity to a form that can be stored and transported, and then used to drive a reaction that produces electricity.

In theory, this uncouples the production of power from its consumption. A simple example is that if you wake up and make coffee and toast for breakfast, the electricity needed to do this is produced more or less “now”. The toaster and coffee pot did not store this energy overnight for later use, and it is impractical and more expensive to have an energy storage system in your kitchen for this purpose. The economics of this and of much larger energy storage problems is changing as technology develops, but the basic question for a large railway implementation is whether it is better to simply “plug in” the train to the power grid through direct transmission rather than through the intermediate step of storing energy in hydrogen.

There is a cost in power to this process because energy that might otherwise go directly to a train is consumed to produce hydrogen, store it, fuel the train, and then conversion back to electricity. The process has an overall efficiency of 30-42% based on a 60-70% rate in electrolysis of water into hydrogen gas, and a 50-60% efficiency of re-conversion in fuel cells back to electricity [p 29]. There is an additional loss of about 8% for the power needed to compress the gas so that trains can be fuelled at a reasonable rate [p 38].

To be fair, there are some losses in a transmission system carrying power from a substation to overhead supply lines along a rail corridor, but they are nowhere near the level imposed by the hydrogen system.

The supposed lower cost of taking power when it is cheapest for electrolysis may well be offset by the higher overall need for power to produce the same result in train service kilometres.

There are other ways to produce hydrogen than electrolysis and only 4% of today’s global production comes from this source. The remainder comes from fossil fuels [p 32].

A related issue is that conversion to hydrogen is supposed to be a response to government targets on reduction of greenhouse gas production. However, some power generation comes from “dirty” sources and this could reduce the net benefit of a hydrogen system.

Track electrification has no flexibility in the use of electricity, and its demand will peak at times of existing peaks of electricity demand on the Ontario grid. While it uses only about one-third of the electricity consumed by Hydrail, track electrification’s use of electricity at peak periods will likely be derived in substantial proportion from CO2-emitting gas-fired generation.

With Hydrail, electricity will overwhelmingly be consumed at times of off-peak demand when gas-fired generation levels are low to nonexistent. So, despite using three times as much electricity as track electrification, Hydrail would draw less than 7 percent of its electricity input from gas-fired sources and could be comparably effective as electrification for RER decarbonization. Either approach will produce less than 20 percent of emissions from diesel traction. [p 34]

Note the statement that a conventional electrification would use about one third of the electricity of a Hydrail system, and that on balance, the emissions related to power generation would be roughly the same for each technology. Obviously the situation would very from one place to another in the world depending on the proportion of power generated from green sources, but one cannot help noting that Hydrail would throw away a substantial amount of power on energy conversion and storage.

From a technical point of view, production of hydrogen by electrolysis of water is a well-established technology, and it can be scaled up simply by adding more units to a “farm” producing gas. The limitation, as discussed earlier, is in the delivery of power to the site where this production occurs.

The next problem lies in delivery of hydrogen to the point(s) where trains will be refuelled. This brings questions of whether it is practical to generate and store the hydrogen close to the fuelling locations, or whether pipeline or even truck haulage would be used. Each of these has its issues with power availability, the amount of infrastructure required and safety concerns for gas handling. Storage and transportation of hydrogen does occur already in some industrial contexts, but the scale for GO RER would be substantial.

Fuelling itself has challenges both for time (how many fuelling stations, how much time per locomotive or car) and operational constraints at fuelling locations such as yards. This also creates a labour requirement for moving trains and handling the fuelling that has no equivalent for a conventional electrification.

Finally there is the onboard equipment of fuel cells, batteries and capacitors that collectively would provide power to the electric motors. Carrying all that around on every train is am additional weight which will consume electricity adding yet another loss in the process.

When one adds in the possibility that hydrogen production and usage would extend beyond GO RER as an “in house” market, this begs the question of whether we are creating a commuter rail system, or a power utility company. If there really is a market for this scale of production and delivery, why should it be an integral part of the railway system any more than we would expect GO Transit to start building its own power generation system?

There is a real sense through the description of hydrogen technology that, yes, many parts of this system already exist albeit at a smaller scale than GO would use, but that this is a technology looking for applications especially when the overhead energy consumption inherent in conversion, transportation and fuelling are taken into account.

Safety will obviously be a big public concern, although the technology to produce, transport and use hydrogen is well-established although not in a railway context.

The much more basic question is whether this proposal would make sense even with the most benign technology imaginable.

Concluding Thoughts

The case, such as it is, for hydrogen trains is set out at the opening of the Executive Summary:

The study concludes that it should be technically feasible to build and operate a Hydrail System for the GO network, and the system’s overall lifetime costs are equivalent to the alternative of a conventional overhead electrification system. It is also acknowledged that such a system is complex, and at this scale, would be a world first. This means that significant challenges would need to be successfully managed to achieve the objective of starting electrified services on the GO network by 2025.

Two of the most significant of these challenges are:

• Fleet implementation – Designing and building a fleet of HFC rail vehicles for RER services would carry a risk of delay due to the design challenges of integrating the fuel cell system into the rail vehicle platform and the production challenges of building a full fleet of new vehicles.
• Electricity price – Due to the large amount of electricity that will be needed for hydrogen production (1% of average daily generated supply in Ontario) by the Hydrail System, the economic viability of Hydrail will depend on how the electricity price variability risk is apportioned with the private sector. There may need to be a provincial government commitment on this within the RER procurement process.

However, there are also many significant risks in taking forward the conventional overhead electrification system that could also impact the 2025 milestone. One key differentiator between the two options is that the Hydrail System also creates the opportunity for broader benefits to Ontario in terms of economic development in the technology sector and as a catalyst for the adoption of hydrogen in other areas of society.

This risk also applies to the hydrogen production, storage, refuelling, and dispensing facilities; and it is probable that the greatest risk relates to the integration of these systems. However, this risk can be mitigated by implementing and operating a small-scale prototype of the end-to-end system so that lessons can be learned about the system’s operation, performance, and reliability; and so these lessons can be fed back into the design, build, and operation of the full system. [pp 8-9]

This cannot be stated too strongly: neither the vehicles – hydrogen-based locomotives or “HMUs” – nor a hydrogen infrastructure on the scale GO Transit will require exist today anywhere. Work in Europe is a start, but on a much smaller scale and intended for a simpler application than an entire commuter rail network with frequent service.

The most troubling part of the whole report is its concept of the “hydrogen economy”, the idea that “Hydrail” would be part of something much larger, an industrial shift by Ontario to a new form of energy storage and delivery. This has all the earmarks of past “industrial development strategies” in Ontario going back decades where the promise of being a leader in some technology overshadows the actual purpose for which it is developed and implemented. In the worst case, this can be a technology (or a product) looking for a market through the back door when it might not achieve success through the front. Ministerial reputations come into play, and clear-headed assessment of proposals is replaced by avoiding embarrassment on announced government policy.

It is entirely possible that there is a place for hydrogen as a storage medium and fuel cells plus batteries for power delivery. What is not clear is the ability to scale this technology to the requirements of GO RER especially for a “go live” date in 2025. The whole idea of a prototype implementation suggest that real progress on electrification will drift off into the indefinite future, and once again technological “leadership” will pre-empt actual improvement of our transportation network.