How much cement could we do without?

Dr Rick Lupton
Department of Mechanical Engineering, University of Bath

Friday 18 December 2020

Cement is an important material

We need to use less of it – but where?

A few years ago, when I was a postdoc at the carefully-named Use Less Group in the Engineering Department in Cambridge, we wanted to know how much cement could be saved through material efficiency improvements, by using it more efficiently. Cement production is associated with a lot of CO2, and we don't currently have the technology to produce zero-carbon cement substitute at large scale, so making less of it is the only option in the short term. The group had done a lot of work on material efficiency for steel and aluminium, and we wanted to ask the same questions about cement. But material efficiency is achieved through specific technical changes in the design and manufacture of components, buildings and infrastructure – so we first needed to know what all this cement was actually being used for.

The data is not great…

… at least if you want to know more than what kind of container it was sold in.

How we use cement in the UK

Mass of cement (kt, UK, 2014)

How much cement could we do without?

William Shanks, Cyrille Dunant, Michał Drewniok, Richard Lupton, André Serrenho, Julian Allwood
Resources, Conservation and Recycling, Volume 141, February 2019
https://doi.org/10.1016/j.resconrec.2018.11.002

We could reduce CO2 emissions by ~50%

How we use cement in the UK

Mass of cement (kt, UK, 2014)

They would say things like "Shanks et al show that 15% of all cement is used for repairs and maintenance, so that's clearly a critical area for decarbonisation".

Why's that a problem – isn't it good that people are using it? Well, yes, but I know that this is our best estimate, assembled from lots of bits and pieces of data and informed assumptions.

For example, this bit comes from some data on the number of railway sleepers produced, together with an estimate of how large and dense a sleeper is. That's probably about right.

But this bit comes from some slightly vague statistics from the European cement association, without much detail of what this includes or how it was determined.

So when people refer to that bit, I don't like it.

I'm not apologising for having published a result that's a bit rough. As Carveth Read said: "It is better to be vaguely right than exactly wrong" – if what we need to know about is repairs and maintenance, there's no point saying "I don't know about that but I can tell you how much is used for railway sleepers?".

But I really want to be able to properly keep track of how certain we are about different parts of these results.

It's a general problem

We don't have good tools for modelling and tracking uncertainty in material production systems.

Applications:

Process-level
Whole economy
Materials
Energy

Process-level: using control data

Ana Gonzalez Hernandez, Richard C Lupton, Chris Williams, Jonathan M Cullen https://doi.org/10.1016/j.apenergy.2018.02.181

Absolute Zero

Screenshot 2020-12-18 at 11.12.17.png

The UK will reach zero emissions in 2050 with today’s technologies:

We will close anything that cannot be electrified or which has process emissions
We will electrify everything else
Ignoring intermittency, we will have 60% of the electricity we would want if electrifying everything today.

Allwood, J., Azevedo, J., Clare, A., Cleaver, C., Cullen, J., Dunant, C., Fellin, T., et al. (2019). Absolute Zero. https://doi.org/10.17863/CAM.46075

Another example is UK FIRES, an EPSRC Programme Grant focusing on the opportunities for UK industry in decarbonisation and resource efficiency, which I'm a co-investigator on.

We published a report a year ago called Absolute Zero, which is all about meeting the UK's net zero targets using today's technologies: we don't have time to wait for "magic beans" that will solve the problem in future.

That means we can't rely on CCS, and there will be no more cement; no flying or long-distance shipping; all energy will come from renewable electricity, and there will be less of it available. And no beef of lamb.

That's not to say we shouldn't be developing things like CCS, and hope that they will make life better in due course; but for zero carbon by 2050 (and new cuts by 2030) we should have a plan that doesn't rely on that working in time.

The report received quite a lot of attention and was debated in the House of Lords earlier this year. Whether or not you agree with the specifics, it's very useful for framing interesting research questions and opportunities: pushing things to extremes is a class creativity technique. And by talking about Absolute Zero rather than an 80% reduction, you get a clarity that's very useful.

UK production system
under Absolute Zero constraints

Ongoing UK FIRES work: what are the opportunities for UK industry highlighted by these constraints?

Using Bayesian inference and MCMC

Richard Lupton, Julian Allwood. Journal of Industrial Ecology 22(6), 2017.
https://doi.org/10.1111/jiec.12698

Let's go back to that problem of tracking uncertainty in these models. I had fun and some success applying the framework of Bayesian inference, using Markov Chain Monte Carlo sampling, to better keep track of the uncertainty in different parts of the system.

Just briefly, if you're not familiar with what that means, here's a simple example.

On the left we have a very simple system, with one input and two outputs. We know that mass must be conserved, but we're uncertain to start with about the size of the input and distribution of the outputs. So as far as we know, the system could look like any of these little pictures.

If we than have a piece of data related to one of these flows, the inference process lets us narrow down the possibilities to those that are consistent with the data. For example, on the right, the colours show the probability of each possible situation, and initially it's quite uniform, although we think the output flows are more likely to be equal than very different. If we then get some data measuring the size of the input, our updated knowledge becomes more specific.

Results after some data…

…and after more data

Conclusions

Screenshot 2020-12-18 at 11.12.17.png

Thank you for your attention!

Here is a video of the uncertain steel Sankey diagram animating.

More 2

Here is a Sankey diagram illustration of UK carbon flows:

How much cement could we do without?

Cement is an important material

The data is not great…

How we use cement in the UK

How much cement could we do without?

How we use cement in the UK

It's a general problem

Process-level: using control data

Absolute Zero

UK production system under Absolute Zero constraints

Using Bayesian inference and MCMC

Results after some data…

…and after more data

Conclusions

More

More 2

UK production system
under Absolute Zero constraints