A new research report, published in science journal Applied Thermal Engineering, shows the potential of saving South Africa’s power utility, Eskom, billions of rands through the maintenance optimisation of coal-fired power stations.

The research, based on the operations of the Kelvin power station, in Johannesburg, shows how optimum maintenance of power stations can unlock cost-effective decision- making within a sector plagued by ageing infrastructure, neglected equipment maintenance and reduced funding, says lead author Ian Mathews.

The paper, ‘A simulation-based prediction model for coal-fired power plant condenser maintenance’, outlines the outcomes of the development of an integrated, semi- empirical thermohydraulic simulation model, or digital twin, for the power station and its integration with economic and maintenance models.

“We have been involved in the simulation optimisation field for the past 30 years and have shown the value of complex modelling for optimum decision-making in deep mines,” says North-West University’s Professor Marius Kleingeld, one of the authors.

“We could not find, internationally, such a similarly complete, fully integrated model for coal-fired power stations.”

Within the context of applying optimisation simulations to another field to enable the production of “more with less”, and safely, the team reviewed the option of approaching Eskom, with many of the parastatal’s coal-fired power stations running beyond their design life.

While getting Eskom on board for the new study proved unsuccessful, privately owned Kelvin power station agreed to do so.

The study, which started in June with the first site visit to the Kelvin power plant, introduced the model for the investigation of the potential optimisation opportunities in the power station’s ageing infrastructure, which was far from operating at its design capacity.

“All the important engineering indicators in the thermohydraulic model were verified to within 5% accuracy,” Mathews says, explaining that this enabled a fully verified computer digital twin of the specific power station as it operates in situ.

The integrated model also accounts for discrete events such as unforeseen breakdowns to ensure the simulation results are as practical as possible.

“As condensers have an important impact on maximum generation output, the simulation model was used to provide new insight into the cost implications associated with delayed or premature maintenance of a condenser,” he adds.

The model was used to determine the most efficient predictive maintenance schedule for the condensers of a 60 MW and a 350 MW unit.

The results showed that, when applying the new predictive maintenance schedule, considerable savings could be realised.

While the simulations were focused on the Kelvin power station, the findings were extrapolated to Eskom’s entire fleet of coal-fired power stations, revealing potential savings in excess of R900-million a year just from the maintenance optimisation of the condensers, says Kleingeld.

Further, with a myriad of elements in a complex power station that can be optimised for further savings, the simulation can be used to investigate the effect on plant performance and economic viability of any change to any component, such as boilers, in any coal-fired power station.

“Most of the proactive work we do at Kelvin power station, such as cleaning condensers, is based on experience. Planning was always reactive. Typically, we could only plan four weeks in advance. However, with this simulation, we can start our maintenance planning much longer before the maintenance is needed,” says Kelvin power station deputy engineering manager Ignus Cronje.

This is an important advantage for the efficient running of the station.

The model can be used in financial planning and yearly maintenance scheduling to enable the power stations to run at their required design capacity.

The research, sponsored by engineering company ETA Operations, was also presented for a postgraduate research qualification at North-West University under Kleingeld.