Design Engineering
Showcase 2021

Optimising Cold Chain Vaccine Delivery via Drones in Rural Nigeria


Project Details

Alex Luo
Design Engineering MEng
Dr Chandramohan George
Masters Project

The COVID-19 pandemic has emphasised the necessity for equitable vaccine distribution whilst minimising waste. To address these logistical challenges, drones have been seen as a potential solution in extending vaccine supply chains to remote areas. However, due to high investment costs, there are concerns about their true value. This study aims to answer how drone routing and scheduling can minimise the costs of drone delivery with cold-chain requirements. Overall, this study has highlighted how drone program costs can be impacted by weather, vaccine type and refrigerator investment which underlines the key factors to consider when developing future drone programs.

Modelling Zipline’s Target Area

It is estimated that around 20% of an LIMC may not have access to vaccination programs due to poor road infrastructure and geographical obstacles. This final leg of the journey to rural clinics is called the last-mile and one company called Zipline has been using medical drones over the past five years to strengthen last-mile deliveries. This study was based on Zipline’s COVID-19 vaccination delivery program in Kaduna state (Central-North Nigeria) for a realistic context. The optimisation therefore mapped out a central distribution centre with drones that can cover an 80 km radius of surrounding healthcare facilities. Local wind, temperature and obstacle maps were also used create a realistic modelled environment.

Map of Birnin Gwari's healthcare facilities with temperature, windspeed and obstacle maps.

Drone Routing

This study focused on two main costs. First individual journey costs were optimised through picking the best route and power usage. The best route was determined based on sub-costs of time, energy, battery degradation and data costs which were also dependent on the weather conditions. Results showed that for one journey, the optimised power at 519W reduced costs by 13% and time by 19% but increased battery degradation and energy use.

Map of Birnin Gwari's healthcare facilities with temperature, windspeed and obstacle maps.

Scheduling and Drone Item Minimisation

Secondly, the least amount of drone items to deliver a monthly supply of vaccines were determined. Here, drone items mean the number of drones, batteries and chargers. To determine the relationship between drone items and delivery times, scheduling was used which was adapted to account for cold chain constraints too. These included ensuring deliveries arrived at times where all vaccines could be administered before the end of the day. Results revealed that the optimal drone item set consisted of 22 drones, 29 batteries and 10 chargers which saved 33% of investment costs compared to Zipline’s benchmark.

Map of Birnin Gwari's healthcare facilities with temperature, windspeed and obstacle maps.

Contextual Parameters

In the context of Nigeria, there are varying levels of healthcare, climate and infrastructure. Thus, for both optimisations, the contextual parameters were varied to analyse the value of cost optimisation in alternate scenarios. Varying these parameters has revealed how areas with higher temperatures and windspeeds would benefit more from journey optimisation. Furthermore, factors such as refrigerator availability and vaccine type in the scheduling optimisation were critical factors that could vary drone item costs. This was due to cold chain constraints, which should therefore be accounted for in further explorations.

Map of Birnin Gwari's healthcare facilities with temperature, windspeed and obstacle maps.