H2 production through water electrolysis for power-to-X applications is being investigated by comparing different bidding strategies on the electricity spot market in Sweden. For that, a price independent order (PIO) strategy was developed assisted by forecasting electricity prices with neural networks (NN). For comparison, a price dependent order (PDO) with a fixed bid price was used. The optimization of the NN showed that increasing the number of neurons in the hidden layer did not reduce error in forecasting due to possible overlapping of data making the model unnecessarily complex. By using different combinations of data for in-sample training and data from 2016-2018 for out-of-sample testing, preliminary results showed similar trends for PIO and PDO when bid prices are increased. However, the PIO marginally reduced the average cost of electricity when compared to PDO in all scenarios, but this was at the expense of increased non-operating hours (cold and warm mode). Further investigations with a mathematical optimization approach will reveal ideal conditions to run the system with low H2 production costs and increased profitability.
Keywords Variable renewable energy, Nord Pool, day-ahead market, water electrolysis, hydrogen production, process optimization