MODELLING THERMAL DYNAMIC RESPONSE TO A
3-HOUR TOTAL POWER SHUTDOWN EVENT
Marc Dupuis
1
, Ingo Eick
2
, Friedhelm Waldmann
3
1: GeniSim Inc., 3111 Alger, Jonquière, Canada, G7S 2M9
2: Hydro Aluminium Deutschland GmbH,41468 Neuss, Koblenzerstrasse 122, Germany
3: Hamburger Aluminium-Werk GmbH, 21129 Hamburg, Dradenauer Hauptdeich 15, Germany
Email: Ingo.Eick@hydro.com
Abstract
The aluminium electrolysis process is an energy-intensive alumina reduction process
which is more and more suffering under the high energy cost. Under these conditions,
power could become a trading asset for the aluminium industry. Partly, actual power
contracts contain clauses for power supply reduction and buyback, which have to be
carefully considered with regard to the compatibility of the pot technology, i.e. not
harming pot life in the long term compared to making gains on the power market in
the short term.
To better understand the thermal and chemical impact in a pot during power cut back
or modulation a dynamic lump model was used as cell simulator to predict the pot
behaviour based on a 3-hour shutdown measurement.
After calibration of the thermal and chemical response of the pot in the model, the
model shows good agreement with the measurement data and allows to investigate
temperature development, sludge formation, alumina solubility and changes of acidity
during power shortage. With this tool an optimised pot setting and preparation can be
established to avoid sludging during power shortening and an anode effect after
power up.
Keywords: Aluminium electrolysis, energy cost, power modulation, dynamic lump
model, Dyna/Marc, sludge formation, anode effect
Introduction
In the past, energy costs accounted for about 30% of total production costs. With the changes
in the power market in Europe and elsewhere over the last 10 years, this share has risen in
some cases to more than 50%. As described by Richard et al [1], more and more smelters now
have a huge financial incentive by reducing their power consumption on request during peak
demand periods.
Power modulation events, however, affect the smelting process [2,3]. This is why studying
strategies to minimize process perturbations using mathematical models [4] prior to their
actual implementation in smelters is extremely useful.
The essential prerequisite is to have a reliable and validated dynamic cell model, which
contains the chemical, heat and voltage balance. An electrolysis cell is a highly non-linear,
multi-physics process with solubility and sedimentation effects, the interactions of which are
Page 1 of 7