**5. Modelling of Kenya's Future Electricity Grid**

This section reviews modelling studies that have been conducted relating to the future of electricity grid in Kenya. Similar studies in selected African countries and outside of Africa are also reviewed.

PV as a technology for electrification of rural Kenyan communities was modelled in [21]. The results showed that interconnecting a number of solar minigrids into one common grid leads to better technical performance. Moreover, the more minigrids connected, the better the performance and the more the power available for the national grid for supply to consumers.

A system-level model of Kenya was presented in [2]. This was used to assess the combination of PV and hydropower to displace diesel-based power generation. The research tested various generation mixes for the years 2012 to, and results obtained showed the value of high penetrations of PV exceeded expected feed-in-tariff payments.

In the work of [62], a spatially explicit supply model was developed to seek for a least-cost PV electrification model to connect consumers that are not served by the national electric grid. Information from individual consumer demand was used to develop this model. It was concluded from the results obtained that PV minigrids can serve up to 17% of the country's population. This includes due consideration for latent demands—i.e., demands currently unmet but which either grid expansion or access to PV electricity may stimulate.

In another study by [63] a rural electrification spatial model for Kenya was developed to identify various approaches to meet electricity needs for various places in the country. The analyses considered both diesel-based generators, various renewable energy sources and expansion of the national grid. Results obtained showed that renewable energy sources can play a key role in meeting energy demand for rural consumers.

Another study conducted in Morocco by [64], an energy management algorithm was modelled and simulated using MATLAB/SIMULINK to serve the load. The system is a grid-connected PV-battery which can manage its energy flows via an optimal management algorithm. Results of this modelling showed that the load was well served in all cases by instant solar production.

In Tunisia, research by [65], a grid connected PV system was modelled using a command approach to function under normal conditions and Symmetrical Grid Voltage Dips (SGVD). In normal operation mode, the command developed increased the low solar voltage to a suitable level corresponding to the Maximum Power Point Tracking. Under the SGVD, the control strategy should ensure stable connection as long as possible and inject more reactive power to support the grid faults. The modelled control scheme in the various operation modes presented high performances in

transient and permanent phased. The system improved safety of the overall system and increased the connection time.

Experience from outside Africa has demonstrated how a decentralisation of power production on cogeneration of heat and power plants [66,67] and wind power [68,69] reduce grid loading thus enabling weaker grids, leave capacity for transit if so needed while at the same time reduces grid losses [70]. These analyses are based on a combination of energy systems analyses [71] using the EnergyPLAN model [72,73] and transmission grid analyses taking into consideration hour-step analyses of the 150 and 400 kV transmission grid.

All in all, experience from the literature demonstrate that a decentral deployment of grid connected PV systems could benefit the energy system—though due concern of ancillary services is still required.
