A key factor when planning energy storage systems (ESS), for example for a microgrid, is to determine the expected cost savings and performance benefits provided by various ESS configurations.
Battery modelling offers a powerful way of predicting the lifetime performance and return on investment that will be provided by each ESS option.
Fuel savings are often a key factor in the choice of energy storage configuration, especially for microgrids which are often located in remote communities and rely on diesel generation, with logistical challenges around fuel delivery. However, cutting fuel consumption is just one of the purposes of battery modelling for microgrids.
Battery modelling techniques continue to evolve to better address the wider context of microgrid and renewable energy deployments. For example, simulations are now key to the project development process, as they deliver insights into renewable and storage applications ahead of deployment and help determine how much power and energy are required overall.
Modelling an entire microgrid at a high level is a valuable exercise in assessing the viability of different deployments of renewable energy schemes with storage. However, when it comes to modelling the detail of these systems – such as bridging between multiple diesel generators in a large microgrid, or optimizing the set-points for operating with diesel generators in a smaller microgrid – more precise modelling is required.
High-frequency data, with granularity of no more than ten-minute intervals, is valuable. Such modelling provides insights into system operation, including diesel synchronization and cool-down times, to minimize diesel starts, maximize fuel savings and optimize battery
High-level modelling is typically based on hourly data, and the granularity of ESS dispatch is correspondingly coarse. This kind of modelling is feasible even with minimal data input.
For example, an initial model of a microgrid can be constructed with minimal inputs, such as the coordinates of an island village off the US Pacific coast having a peak load of 150 kW in January. Based on this information, high-level modelling can be used to construct a typical load profile, and location-specific solar or wind data can be downloaded.
The modelling software can then quickly carry out multiple simulations to discover the optimum renewable energy power rating, along with an appropriate level of energy storage. The results illustrate fuel savings and, if sufficient inputs are provided, ROI.
However, precise modelling requires more detailed inputs and time to optimize the dispatch methodology. Combining high-level and precise modelling leads to a more cohesive, informed insight into ESS requirements – in turn, enabling an accurate evaluation of a project’s viability, as well as the development of a detailed strategy to help ensure project success.
When it comes to modelling microgrids, the data requirements are relatively simple. They typically include load, renewable resource and diesel configuration, along with any information on dispatchable loads, such as electric water heaters.
Outside microgrids, ESS may also be needed to support weak grids – such as on islands – playing a critical role in grid stabilization by addressing both the variability of renewables and other disruptions, such as generator
In this situation, battery modelling is typically based on frequency response, where the energy storage output varies constantly as a function of the network frequency: charging when the frequency is high, and discharging when the frequency is low. Here, the input (grid frequency) is simple, but the choice of parameters is much more complex, as it includes frequency deadband, droop slope and the all-important state-of-charge-management function.
It is important to bear in mind that the simulation input (frequency) will be altered by the output (charge or discharge power). What is more, the currently available data may not be a true reflection of the grid’s planned development – especially when it comes to the future deployment of renewable power
This makes it crucial to perform additional modelling when the system is in operation – and periodically throughout its life – and adjust operating parameters accordingly. Experience has shown that even minor adjustments can greatly extend a battery’s life.
Significantly more accurate modelling – and higher certainty of the value of an ESS investment – can be achieved using data with high sampling rates. Sampling of grid data over one week during each season is recommended, to ensure that seasonal effects are properly considered. Collecting data during a major frequency event is also valuable.
ESSs are often called upon to provide black start capabilities, especially in isolated island grids. In the event of a system-wide blackout, a properly designed ESS can energize a path to a generator and provide control power for starting.
While the transformers, feeders and transmission lines are reenergized, the ESS can respond to frequency shifts to keep the entire system functioning properly. Battery modelling during the ESS planning phase can help ensure batteries always maintain a reserve of energy for such black starting.
For maximum accuracy, battery modelling software should run the same algorithms as battery management systems – mimicking real-world battery behaviour. Efficiency losses and thermal management should be included, to ensure the system’s cooling is adequate. Since battery losses increase with age, it is also vital that modelling software accepts inputs for different stages of battery life, as well as includes aging outputs. This is the only way to ensure a degree of certainty around the lifetime of the system and its end-of-life characteristics.
As grid operators explore ESS opportunities, they should keep battery modelling top of mind – simulations should be at the forefront of their strategy to maximize cost savings throughout the entire process. Taking the time to conduct battery modelling opens the door to quantifying energy storage benefits, including fuel savings in microgrids, and provides an opportunity to ascertain the project’s ROI potential.
Michael Lippert is marketing and business development manager for Saft’s TTG Division