Materials and Dematerialization, by Vaclav Smil: A review

I wanted to like this book. I really did, and not just because Bill Gates thinks the author is great: Our civilization depends on material flows in ways most of us don't consider or aren't aware of. Just the amount of earth and rock that must be moved and processed in mining metals is huge. These material flows in turn place huge impacts on energy use, well before we get to the mineral smelting and purification steps we all think of when we think of the climate impact of metals. And with energy the main driver of climate change, it is worth asking whether we can support modern levels of comfort, not just for the industrialised West but for all humanity, while reducing the current energy and material intensities required to do so.

But. I've run into a snag. Early chapters covered fields with which I am not familiar. Smil's apparent command of the intricacies of different industrial sectors and ready supply of well-footnoted numbers is very impressive, even if the text could have benefited from a good editor. (English is obviously not his first language). Then I got to Section 3.1.2 on wood, about which I know a few things, and 3.1.3, pulp and paper, about which I know quite a lot, at least in the Canadian context. 

For starters, he ignores the massive shifts in the paper industry due to the Internet: demand for newsprint and other wood-containing grades cratered initially as personal and classified ads moved to the internet. This trend was accelerated by the disappearance of weekend advertising inserts and flyers, catalogs and directories. (When was the last time you saw, let alone used, a phone book?) He also ignores the massive growth of packaging grades (there's no app for a box) due to the online e-commerce phenomenon heralded by Amazon and Ebay. None of this is mentioned. Instead he focuses on the disappearance of photographic grades of paper. While undoubtedly tragic for the employees of the mills affected, this had a very minor impact overall, leading to the closure (if memory serves) of two small hardwood pulp mills in Canada out of something like 100+ mills in the country at the beginning of the century. 

And when he does mention declines in newsprint demand, it is only to attribute the decline to the move to online news sources, not the much larger loss of that portion of the daily paper that was once dedicated to advertising.

Finally Chinese imports of wastepaper are not their only source of fibre; virgin fibre from South American or Indonesian pulp mills is a huge source. The Chinese paper mills which once consumed recycled materials imported from North America are now increasingly supplied by the internal Chinese domestic market, which has grown to the point of supporting such an industry; and the packaging material produced is used for exporting goods to North America and elsewhere. 

None of what he writes is demonstrably false, although to be fair I haven't verified his numbers; it's more that many statements are met with a strong reaction that, yes, that's true, but the reality is far more complex and some key drivers that could lead to changes in material flows are not being considered. essentially it looks like he pulled numbers from a wide range of sources, but didn't bother talking to anyone in the industry. 

So for a book intending to cover material flows in some detail, it is unfortunate that the first impression is of playing fast and loose with underlying trends and drivers, even if the numbers are correct. One is led to wonder how well other industries and the drivers affecting them are understood and described. 

Sadly I will be reading the rest of it with a somewhat more cynical eye. I'll post updates as I continue. 

High-speed rail versus short haul flights: personal experience in Europe

Recently I had to get from Hannover to Brussels. This looked like a good opportunity to compare rail and air options. In order to be sure to capture all the costs (carbon emissions, time and money), I decided to calculate hotel to hotel rather than gate to gate or platform to platform. 

In both cases I included walking from my hotel to Hannover central train station (20 minutes). The train ride to the airport is 26 minutes with departures every 30 minutes, so I allowed 40 minutes for that trip. I also allowed 20 minutes in the station for locating the right platform, whether the inter-city train or the one to the airport. Additionally I assumed one should get to the airport 90 minutes early. At Brussels airport I allowed 20 minutes to get to the train station in the basement of the airport, 20 minutes to find the right platform and 20 minutes for the train to Bruxelles-Noord. From there, it was a 10 minute walk from the platform to the street and another 10 minutes to my hotel. 

Lufthansa offers an early morning option through Munich, total time 7h30 of which 2h30 in the air and with a connection time of 50 minutes. (While the point of this note is not scheduling convenience, note that this option also meant leaving the hotel at about 3:30 am.) The alternative, an evening flight through Frankfurt, involved only 2h00 in the air, but 1h45 on the ground, for a total time of 7h55. (In this scenario I'd arrive at the hotel at midnight).

So by air: two inconvenient flight times and a total travel time between 7h30 and 8h00; arriving at the airport an hour early instead of 90 minutes would reduce this by 30 minutes. Prices were about $300 Canadian before seat selection or a checked bag (no carry-on bags allowed on these internal flights, apparently, even in business). $330 gets a seat and a checked bag, but the timing shown here does not include waiting around the baggage carousel in Brussels.

(The photo is of me on the red-eye on my way to Europe).

The train leaves Hannover at a civilised hour of 10:31 and gets to Bruxelles-Noord directly, with 33 minutes to change trains in Köln, at 15:26. So this is just under 5 hours in transit (at speeds up to 250 kph), versus two or two and a half hours in the air. But the total travel time by train was just under 6 hours door to door. Cost in first class (which basically gets you an assigned seat with a bit more legroom) was just over 100 euros.

In all fairness, the 33 minute time to change trains in Köln was whittled down to 6 minutes by delays en route. Deutsche Bahn arranged for the two trains to be facing each other across a platform so no need to run to another platform, but still this was unsettling. Like you, I was under the impression the Germans could get the trains to run on time... and the train continuing to Brussels was delayed by about 15 minutes. Nonetheless the train is the big winner on cost and time in this case, and probably on carbon emissions as well as both trains were electric and thus dependent on average grid emissions.

(The photo is the train station at Frankfurt airport, taken on my way to Hannover. This train was also delayed! Almost 30 minutes in fact. Shocking.)

So the moral is two-fold. First, trains can be cheaper, faster and (presumably) better from a carbon perspective than short hauls, especially if you need to change planes -- availability of a direct flight would have changed this calculation significantly. Second, don't book trains with excessively short connecting times as Deutsche Bahn doesn't know if you've actually gotten on the next train (unlike the plane where your boarding pass is scanned at the gate) and can't be counted on to hold it, even for a minute or two. 

So where's our high speed rail from Quebec City to Windsor? Tied up in politics and (possibly) in not wanting to cripple Air Canada's cash cow (the Montreal-Toronto run that I know so well). Sad. 

In the meantime, safe travels! 

Canadian Academy of Engineering: A personal saga

This week the Canadian Academy of Engineering announced its list of new Fellows, and yours truly was included. 

Needless to say I am humbled and honoured. It has been a long and tortuous path to this point, and nothing along the way pointed to this level of recognition. 

I obtained got my first real job, as an apprentice mechanic at the local Mercedes Benz dealership, in the spring of 1975. (I passed on a lovely 250 SE sedan while there, still regretting that decision). By about 1978 I was employed by a small independent shop focused on British cars, and by extension Swedish cars (they used British carburettors). I got to know the local oddball scene: among the customer's cars were a Ginetta G15, a TVR Grantura Mark 2A, an early Dellow, and a raft of Jags, Healeys, MGs and Triumphs. (While there I picked up a Rover 2000 TC in British Racing Green with the Rostyle wheels, no regrets there except for the rust that eventually killed it). The shop also serviced German cars (as the Swedish cars had German electrical and fuel injection systems) and basically anything else that rolled in the door. I finally got around to getting my Auto Mechanic's certificate (2nd class) in 1982.

By 1984, I was married and a father. The mechanic's work was fun but the pay poor; and after 10 years it was time to move on. As well we were starting to see onboard diagnostic systems, and I was worried the computer would make the fun part of the job (figuring out what was causing that odd set of symptoms) obsolete, leaving only the wrenching. So with my wife finishing up her degree in social work, it was my turn to go back to school. 

McGill Engineering was not impressed by my lack of scholastic credentials and refused to even entertain my application, so in 1984 I went across town to Concordia for a year of pre-engineering math and science courses. Straight A's here encouraged McGill to reconsider my application, and I started in Mechanical Engineering in the fall of 1985, at age 30. 

Summer jobs in the next few years included a return to the garage and a stint in a garment factory; but getting a summer job over in Chemical Engineering assisting J.-F. Bond, a graduate student with Murray Douglas and Ron Crotogino, proved to be a smart career move. After spending two summers helping J.-F. get his equipment built and running, in 1989 I got the B.Eng. diploma and went to work as a graduate student myself, with Ron as my main technical supervisor. Along the way this Master's project got turned into a PhD project. My second child was born along in there as well. I managed to defend the thesis in late 1994 and start work at Paprican as a research engineer before hitting 40. 

At Paprican I worked on a variety of energy efficiency projects, some successful (the Energy Monograph and subsequent efficiency studies), some less so (I was instrumental in killing off the Szego mill). The thought that pulp mills also make heat and power led to the thought that maybe they could make other things; and while byproducts such as lignin, methanol or turpentine were well known, there was room to expand this, especially where the end product could be used to displace a petroleum-based competitor. The Biorefinery program was the result, in about 2005; this led to the LignoForce system installed at West Fraser's Hinton pulp mill and a few other success stories.

I had the pleasure in this period of working with a great group of people, from lab technicians and pilot plant operators to research scientists and engineers, and on to mill operators, process engineers, equipment designers and manufacturers. I learned that innovation requires a huge range of skills and viewpoints, not to mention cash, to be successful. So cheers and thanks to the supporting cast of characters, without whom I would not be writing this; you know who you are.

Throughout that time, I remained convinced that fuels should only be made from the residues for which no better use could be found; products too close to the barrel of crude are too dependent on the price of that crude, and can't be economically displaced by wood-based materials unless the barrel is consistently well above $100. Products further down the value chain from the raw barrel, such as phenolic resins, are less sensitive to the price of oil; and it is possible for lignin to compete in plywood resins as long as oil is above about $40. Some products are potentially profitable even if the oil were essentially free, although care must be taken to understand how total market volume tends to shrink as you move into specialty chemicals. 

But the new climate change imperative and government reaction to it means that the playing field for fuels may well be levelled by a combination of policy moves which serve to generate an "effective oil price" against which wood-based fuel products can indeed compete. Such policy actions include carbon taxes or financial support for early-stage plants, and serve to recognise the fact that the costs of burning fossil fuels, in terms of fires, floods and heat waves, are not included in the price at the pumps. Of course this requires governments to actually act on their newfound understanding... 

So the Academy has now inducted a former British car mechanic... it will be interesting to see how that plays out, for me as well as for the Academy. Stay tuned!

The forest biorefinery: an update in an era of Net Zero

Recently the Finnish forest products giant UPM announced the building of a forest biorefinery at the chemical park in Leuna, near Leipzig in the former East Germany (1). The plant, with a capital cost of €550 million, will produce 220 kT per year of bioproducts from beechwood through an enzymatic hydrolysis process: mono-ethylene glycol (MEG) and mono-propylene glycol (MPG) will come from the cellulose component, and a carbon black substitute for use in rubber will be produced from the lignin. This follows on the development and commercialization of the BioVerno renewable diesel process at UPM's Kaukas mill in Lappeenranta, Finland (2). There are some lessons to be learned here.

Research is key. And expensive. And time-consuming.

Yes, research costs money. As a newly minted research manager back in 2001, I arranged a tour of research labs around the world to meet my new peers. This trip included the UPM-Kymmene labs at the Kaukas mill in Lappeenranta. These labs were very well staffed, with about 100 researchers.

Labs are expensive to operate. Wet labs are very expensive to operate. And pilot plants, with high headroom space you can back a truck into, cranes for moving heavy bits around, and high load-bearing floors, are extraordinarily expensive. Add on conventional overhead (IT, HR, payroll, etc.), $100k/y to support university work, travel to mills and conferences, and laptops and cell phones for everyone, and a typical research team, consisting of a senior scientist, a junior scientist and two or maybe three lab technicians or pilot plant operators, costs at least $1M/y before building anything in the pilot plant or paying for maintenance and operation of gas chromatographs or scanning electron microscopes. Capital equipment in the lab or pilot plant is additional.

And it's not enough to set it all up -- you have to persevere. In the mid-2000's, UPM got serious about the BioVerno renewable diesel process. They expanded the labs at Kaukas, hiring new staff and buying new equipment. (They also stopped letting Canadian research managers drop in for a chat and a look-see). The result, after about a decade of research investment at the $25M/y scale and some significant capital expenditures, was a commercial-scale biofuels plant. Call it a round billion dollars, more or less, over a decade, more or less, before the first commercial litre was pumped into a tanker. Another decade at the same investment level has led to the Leuna plant.

Location matters

Following the 2009 Nordic Wood Biorefinery Conference in Helsinki, I visited Leuna (3). Liepzig is an hour from Frankfurt by air, with Leuna another 30 minutes by car, so I made a quick one-day detour on my way home. Commissioned in 1917, and located so as to be beyond the reach of French bombers (!), it was thoroughly rebuilt after German reunification, and is today one of the most modern chemical parks in the world. Infra-Leuna manages the site, and rents space along with access to pipe racks carrying water, steam, naphtha, ethylene and so on. At the time, attracted by the beechwood common in the area, Fraunhofer (Germany's equivalent to Canada's National Research Council) was preparing to build the Chemical and Biotechnology Research Centre onsite, with a focus on enzymatic hydrolysis of hardwoods and other processes.The initial investment was €25M in capital, with another €25M for the operating costs in the first few years. Now UPM is building a full-scale plant there. 

So why build a full-scale plant in Leuna, and not Lappeenranta? First is the availability of beechwood in reasonable quantities and within reasonable haul distances; second is the existence of a well-funded research laboratory onsite. But that is not the whole story, because Lappeenranta also features both hardwood supplies and R&D labs.

The key is that Leuna has a range of tenants. At the time I visited, this included Total (which operated a large refinery) as well as several major chemical companies attracted by easy access to the naphtha left over from Total's gasoline and diesel production. So it is likely that all the products proposed by UPM will find buyers onsite. But if any buyers are further afield, Leuna, located in the centre of Europe, is well served by rail if not canals. 

Implications for the Canadian forest sector

I don't want to minimise the success of Canadian efforts in lignin, in cellulose nano-crystals, in cellulose filaments and in other areas such as tall wood buildings. But the world is going to need large-scale pathways to non-fossil products in a big hurry, and existing successes are, arguably, in niche markets. Profitable niches, yes, but their impact on global challenges is small.

The common theme here is the requirement for a world-scale, consistent, reliable, long-term commitment to innovation. Innovation, which starts with university research but which must progress through the pilot plant stage before commercialization, is time-consuming and expensive, and cannot be judged quarterly on benefits delivered. Canadian forests represent a world-class resource, in terms of both volumes and quality; furthermore our forests are largely harvested according to sustainability metrics set out by a range of arm's length organisations such as the Forest Stewardship Council. 

At a very rough scale, UPM has been spending of the order of FPInnovations' entire annual budget, on bio-products alone, for close to two decades now. They have also invested in full-scale commercial plants once the research results have warranted it. This corporate investment is above and beyond publicly funded research at VTT or in various Finnish universities. Finally UPM is not alone in this; Stora Enso and Metsä have followed similar paths. 

The Canadian forest sector and governments at all levels (recalling that forestry is a provincial responsibility) can't rely on a few million dollars here and there, with budget cuts (and associated loss of highly qualified personnel) every few years when things are a bit lean, if there is any hope of taking advantage of our world-class, sustainably-harvested, non-food, forest-based biomass to compete in the coming world bio-economy. Arguably, we should have been investing at this level a decade or more ago, because, as the IPCC has stated recently, it's getting a bit late to start.

Notes

1. https://www.upmbiochemicals.com/about-upm-biochemicals/biorefinery-leuna/ 

2. https://www.upmbiofuels.com/traffic-fuels/upm-bioverno-diesel-for-fuels/

3. https://en.wikipedia.org/wiki/Leuna_works

 

Distributed power generation in urban areas

An interesting article in today's Washington Post newsletter talked about the issues around getting power from wind farms in rural areas to electric car charging stations in urban areas over existing grids. (Click here to see the article). The article mentions a study showing that the US will need to invest $125 billion into the grid by 2030 to meet the need for electric car recharging. 

This got me to thinking about distributed power generation in urban areas. After all, this is where the majority of cars are located today, and where the power will be needed tomorrow. I live in a dense urban neighbourhood characterised by 100-year old row houses with flat roofs, all fronting on narrow streets. The image shows the area as seen by Google Maps.

My house has a surface area of 1100 sq. ft, on a lot 25' by 100'. If I include the total distance from the middle of my street to the middle of the alley in back, I get a total depth of 125'. (Street centrelines are 250' apart.) So 1100 sq. ft., divided by the total urban area attributable to my property (25' x 125'), gives 35% of the urban area taken up by flat roofs when extrapolated to the larger neighbourhood. 

A quick check of the Interweb gives a year-round average solar power generation capacity of about 4 kWh per square metre, per day, for solar panels located along the border between New York state and Quebec. (Click here for the maps.) Assuming this is representative of Montreal, about 100 km north of the border, that works out to an average of 165 Watts per square metre, again on an annual basis -- it will be much lower in winter months, of course. Still, 1100 sq. ft. equals 100 square metres, so this is a year round potential average power generating capacity of 16.5 kW using solar panels with tracking systems to keep them aimed directly at the sun. Regular readers will know I've been tracking my own power use for Hydro Quebec over the last two winters, and I know that my peak power use in a cold snap on turning up the heat immediately after a Hydro-mandated pricing event was never more than 8 kW. Winter average power use this year was closer to 4 or 5 kW, even with a fairly cold January, helped by the fact that as the resident of a ground floor flat surrounded by other buildings, I have very few exposed outside walls. So in theory 100 square metres of solar panels on the roof would provide me with excess capacity, especially in those really cold but bright sunny winter days. 

I extended the analysis to the entire borough, which has a total area of 8.1 square kilometres. I suspect potential usable area of 35% of this is a large number; to start with, there are wider streets as well as parks and other areas where solar panels could not be mounted. But assuming 15% of the borough were covered with panels, the neighbourhood could make, on average, 200 MW. While this is only 0.5% of the 40 GW peaks Hydro hit last winter, it is still requires only 8 square kilometers of urban area, and no need for extended transmission lines. As well it would go a long way towards recharging a local electric vehicle fleet, especially if augmented by storage systems of some kind such as masses of parked electric car batteries all plugged into the grid. (Most cars sit 90% of the time).

I haven't taken the time to go through the economics, but one online source suggested costs of $3 per installed Watt in Ontario. Covering my roof would thus cost close to $50,000, which I wouldn't want to do on my own without serious subsidies; but the neighbourhood could make 200 MW for $600 million. And $3 per W is about 5 times cheaper than Site C in BC ($16B for 1.1 GW), without the need for long-distance transmission lines. 

In Quebec the main issue will be Hydro Quebec's mandate, which does not encourage small-scale or distributed generation unless it is off-grid, but I assume that could be changed by an act of Parliament. But I've met people in Alberta who have put solar panels on the roof of the garage to charge the Tesla; the open-access grid there allows excess power (when the owner of the car is at work on a sunny day, for instance)  to be trickled back into the grid for a credit.

Food for thought.

Sensibo heat pump controller: Update

You may recall my dismay in my last post to discover that the Sensibo device does its thing via a cloud-based system. Turns out this was justified: the other morning, my modem decided to pack it in, and I no longer had control over the heat pump.

The pump remained at its previous setting, just as it would if the batteries in the remote control wore out; but I couldn't change the settings via the app. 

I assume that pre-programmed settings would also not work. However, I don't know whether programmed changes to settings would be reinstated once the modem was repaired, as I didn't have any programs set at the time the modem went offline. 

Of course the remote control still worked, as it is entirely local. 

So overall I remain convinced that something that works entirely locally, or that defaults to local when needed, would be preferable.

Peak power pricing: Sensibo heat pump controller

Last winter I was part of a pilot project with Hydro Quebec to manage my power consumption according to demand. (Details can be found by clicking here, and early results here.) Net result? About a $400 saving over the winter. 

So far so good. And last winter, much of it spent in some form of lockdown, meant I was always home at 05:55 to get up and pad around turning thermostats down. Next step is to automate the process so I can sleep in, or travel. Turns out Hydro is working on this, with their HiLo subsidiary offering smart thermostats for baseboard heaters. But there are a few problems with their approach so far. 

First, these are controlled through Hydro's servers talking to the thermostats via my router. Google Home will also do this; but I recall a server problem at Google a while back that left people unable to even turn on a light. So keeping local control is a better idea in my view.

Second, at this point their equipment will only control baseboards. I've got one baseboard that needs controlling; otherwise, I've got a heat pump, two heated floors and two convection units with built-in thermostats. So that doesn't do me much good even if I were willing to accept Hydro reaching into my home network and fiddling with stuff. 

Separately I discovered a type of gizmo for controlling any heat pump or air conditioner that has a remote control. This is a bit like the universal remote you used to be able to buy to replace the remotes for the TV, set-top box, PCR, VCR, tape deck and DVD player (ask your Dad). Two looked interesting: the Sensibo and the Cielo (for reviews, click here). On the face of it the Cielo has more features (including a display), but ultimately I ordered the Sensibo from Home Depot, on the basis that I can just take it back to the store if there is a problem. At $129.99 Canadian before sales tax, it was a bit more than the US price, which did not include shipping to Canada. (It is interesting to note that the black one was $179.99. Not sure what that is about.)

 

Photos show what is in the box: the device, a 110V adapter and a USB cable. There is a second proprietary port on the underside of the device, but with no indication of what this might be for. 

The instructions in the box are terse, verging on non-existent. The manual is not easily found on the Sensibo website, but I found it on Lowe's site (click here). The main installation step, which was minor, was installing the app (available for iOS and Android, but not PC) and connecting it to my WiFi system. Then you point your existing heat pump remote at it and press On a couple of times. After that it is entirely run from my phone. 

The scheduling function has a seven-day calendar that repeats. I discovered that if you try to schedule stuff with the heat pump off, all you can do is turn it on or off; the heat pump will start up with the last settings active. But if the heat pump is on, you can program all the functions: mode, temperature, fan speed and vane orientation, etc. So far all very neat.

A couple of downsides. Unlike the remote, which takes a couple of AA batteries, it needs the 110V connection. This sort of limits where you can put it, which is further limited by the fact that, like the remote, it needs line of sight to the heat pump to work. Photos show a jury-rigged set up, with the gizmo plugged into the 110 outlet on the stove; I have since plugged it into the outlet for the microwave, which involved drilling some holes in the cabinetry.

 

Powering it off and back on seems to bring it back to where you set it, so that is fine. 

Finally I didn't think I could control it from beyond the range of my WiFi system. But initial tests made by walking up the street until my phone was no longer connected to my WiFi imply that I can control it at a distance; this means that my phone talks to my router over cell networks. (I turned the phone off before coming back home, to be sure it really was talking over the cell network). It turns out that the app and the device talk to one another through a cloud-based server of some sort, and do not communicate directly. The Cielo system is apparently similar. Oh well. Hopefully this system will still be live in a year or a decade. I still have the remote and a stash of AA batteries, so there is a backup plan. But the lack of clarity around this in the Sensibo advertising material and installation manuals is disappointing.

So I am getting ready for next winter, hoping that I will be able to go away for a while and just leave a program running that will dial back twice a day, even if Hydro isn't asking for it. The only real issue is a I have a single convection heater with integrated thermostat and a second baseboard heater that really need to be adjusted manually, especially if the temperatures are low enough that the heat pump is no longer effective. (The coldest it got at the beginning of Hydro events last winter was -17C, which, as we all know, is relatively balmy for this part of the world.) I can set these to a low temperature before going away, but will still have to get up at the crack of dawn to adjust them if I am home. So while not a complete solution, progress is being made. Stay tuned for updates.