A resource for MRI institutes / radiologic institutes towards realization of carbon neutral MRI.
The energy demand of “a single MRI machine in clinical use […] corresponds to 25.8 four-person households” [1]. The “daily energy consumption for a clinical MRIs scanner in operation for 13 hours was estimated to be 363 kWh with 1.5 T units and 530 kWh with 3.0 T units”[1].
Thus, yearly (~300 days) this sums up to 108 MWh for 1.5 T units and 159 MWh for 3 T units.
With almost 3000 MRI systems in use alone in Germany, corresponding to the demand of approximately 75k households, this forms a substantial contribution to the national electric energy demand and to the national carbon footprint. In addition, in case of external energy shortage MRI diagnostics cannot be performed or only using emergency generators that again rely on external fuel resources.
Average per examination 20 kWh ± 5 (range, 12–34 kWh).
The dynamic energy demand[1] is given by the following figures:
This shows that
- MRI has an high constant energy demand at both on and off times, due to the cooling
- MRI has peaks of power demand up to 60 kWh for certain sequences (ssfp)
- MRI has non-negligible idle energy demand during on-active times.
which is also nicely summarized here:
MR scanners have a rather high helium demand for the cooling system as upon usage, heating of the system leads to boil-off of helium. On old systems this helium is either gathered for external reliquefying purposes or released to the environment. Newer generation of MRI scanners have a so-called zero boil-off technology where the generated helium gas is locally collected and reliquefied. The additional helium demand is therefore marginal, which is a significant cost reduction, but the energy demand of zero boil-off systems is larger. Mor infor on ZBO can be found here https://mriquestions.com/uploads/3/4/5/7/34572113/advances_in_magnet_design.pdf
In both cases the helium reliquification process has to be included in the energy demand calculation.
The above mentioned energy demand is most probobably from system with zero boil-off techonolgy (Skyra, Verio, Avanto Fit, Espree), but this could not be verified, as avanto system without zero boil-off exist. It is assumed that this is included for the estimations herein.
- (use lower field, but at cost of performance)
- switching off the scanner as long as possible (or measure at night)
- reduce on-active idle times
- optimize MR sequences
- no cooling at night : store helium gas and use solar power during the daytime for condensation.
- use the generated waste heat as heat supply or reservoir for heat pumps. Under the assumption that 20 % of the energy (108 MWh/159 MWh at 1.5/3 T) can be re-used as heat source, this is 21 / 32 MWh unused energy annualy or 6 houshold per MRI system.
- design and usage of more efficient amplifiers
- more efficient cooling / isolation (e.g. cryogen-free systems, but seem to have similar demand)
Active times are in average 13h, ca. 6am to 7 pm mostly at daytime, where local photovoltaics (PV) can contribute to the energy demand. [2]
As depicted below on average we have 50 % PV capacity between 7 am and 4 pm and we are above the 50 % capacity between March and October.
Thus, a strongly reduced external energy demand can be achieved using local photovoltaics supply.
In 8 of 12 months even an autarkic mode might be possible.
With above mentioned savings we might reduce te energy demand of an advanced MRI from 159 MWh down to 80 MWh. In Germany a PV production of 175 to 235 kWh/sqm is possible[3]. Using 200 kWh/sqm as for our calculation leads to 400 sqm of needed photovoltaics to covere the average demand with the average production. To cover the months in spring and autumn a overdimensioning of a factor two might be necessary, leading to 800 sqm. To use the higher PV energy of noon time for evening and morning, a battery storage can be used.
- A storage of 0.5 MWh could cover approximatley 2 days of demand and seems well suited for this.
- Modern Li batteries also easily cover the mentioned peak power demand, thus the battery also serves as a peak power buffer.
- Moreover, the helium liquification can be done at noon time as well to use abundant energy, and switch off cooling at night.
base costs: PV: 210 Euro per sqm [4] , battery: ~500€ per kWh [5]
- 800 sqm PV -------- 168,000 €
- 0.5 MWh battery --- 250,000 €
- setup ----------------- 32,000 €
- Total:-----------------450,000 €
Currently energy prices are at 42 cent per kWh. Thus, outgoing from the non-optimized demand (159 MWh), this autarkic MRI + PV + battery system would amortize itself within 7.7 years. (Without battery below 3 years). And it would contribute to the heating of the building.
And too many roofs are still empty:
In the winter months one needs to buy wind energy to realize a carbon zero MRI.
In summary we have to
- build 800sqm PV
- reduce active-on idle times (optimize work flow) and match active-on with solar cycle
- don't cool at night, but recool during daytime, at weekend: only cooling at daytime
- add buffer battery for mornings, evenings, and as peak-power buffer
- get wind energy in winter
- use generated waste heat for buildings heating ~ 10-20 MWh/year / 4 households
- Optimize MRI sequences
[1] Heye T, Knoerl R, Wehrle T, Mangold D, Cerminara A, Loser M, Plumeyer M, Degen M, Lüthy R, Brodbeck D, Merkle E. The Energy Consumption of Radiology: Energy- and Cost-saving Opportunities for CT and MRI Operation. Radiology. 2020;295(3):593-605. https://doi.org/10.1148/radiol.2020192084
[2] Pfenninger S, Staffell I. Long-term patterns of European PV output using 30 years of validated hourly reanalysis and satellite data. Energy. 2016;114:1251-1265. doi:10.1016/j.energy.2016.08.060 and the Supporting Info: https://ars.els-cdn.com/content/image/1-s2.0-S0360544216311744-mmc1.pdf
[4] https://kostencheck.de/solaranlage-kosten-pro-m%C2%B2
[5] https://energetechsolar.com/1mwh-500v-800v-battery-energy-storage-system
To read:
Turn It Off! A Simple Method to Save Energy and CO2 Emissions in a Hospital Setting with Focus on Radiology by Monitoring Nonproductive Energy-consuming Devices Tobias Heye, Manfred T. Meyer, Elmar M. Merkle, and Jan Vosshenrich Radiology https://pubs.rsna.org/doi/10.1148/radiol.230162 https://github.com/mtmey/ping_tracker