Energy Assessment Toolkit

The Problem

The conservation problem we seek to address is deforestation. According to the Food and Agriculture Organization of the United Nations, fuelwood harvesting remains one of the biggest culprits for deforestations in developing countries. Fuelwood harvesting for cooking not only jeopardizes the forestland, which is home to tens of thousands species of flora and fauna, but also contribute to respiratory and other health problems among the fuelwood consuming households.This shows the urgentness of ameliorating deforestation associated with residential fuelwood consumption to protect and maintain the richness of biodiversity, thus preventing the sixth mass extinction.

Our Proposal

To minimize fuelwood-based deforestation, it is critical to provide clean, reliable, and affordable cooking fuel to rural communities. Socio-cultural barriers remain as one of the top reasons preventing adoptions of alternative cooking fuel in rural communities. We are developing an application-based toolkit that can overcome the barrier and promote energy access expansion in developing countries. The toolkit contains a two-phase approach where the initial phase involves data collection and modeling followed by a second phase that monitors the effectiveness of our toolkit for continuous improvements. The toolkit has three interfaces for data input by the community, government, and toolkit users in the initial phase. This includes completing an interactive survey template, which conveys a clean cooking idea, emphasizes potential impacts (e.g. cleaner air for better health) from the application to promote behavior change. The collected data will be fed into three data models: load (e.g. fuelwood consumption, household numbers), resource (e.g. available fuel types, government incentives), and financial (e.g. available funds, allocated capital cost) that will subsequently output a set of optimized results and the machine learning algorithm will be trained based on the efficacy of modeled results. The second phase involves devising a tracking method to determine the progress and effectiveness of our model such as utilizing remote sensing data to monitor forest cover over time.

We Assume that...

The use of firewood as a fuel for cooking will increase in the future considering the growing population if alternative cooking fuel is not provided to cope with increasing demand for cooking fuel

Regulatory frameworks and the need to end the exploitation of forests will be a huge cause of concern for governments all over the world

Government and private players are both actively seeking environmental friendly solutions to tackle socio-cultural barriers of expanding energy access

A single sourced application based toolkit would be more widely accepted than isolated applications built for company specific needs

A large number of organizations are willing to use the toolkit and share their feedbacks to improve the machine learning algorithm

There would be a long-term positive impact to environmental sustainability after users interact with the toolkit

Our tool will help private players estimate the financial breakeven point of their project and could measure the social and environmental impact being caused by the use of the application

The community is willing to learn about the benefits of switching from firewood as a source of fuel to an alternative cooking fuel

Constraints to Overcome

The toolkit intends to create a systematic approach towards replacing firewood with cleaner and more accessible cooking fuels. The modelling knowledge that the application provides uses computational methods, along with the information from datasheets to allow the users to make sure that they have the best possible option to tackle all the project uncertainties they may face. Hence, the biggest issue for us is quality check of the output solutions and the optimization of the machine learning algorithms. Because each solution would require months to years to show its efficacy, we would need to find beta testers willing to provide timely feedback. When the toolkit becomes public, we would also need to develop a communication channel for all users to share their results in an aggregated manner so the machine learning algorithms would have more information to work with.

Current Work

Since this application is made for the organizations that operate directly with the rural communities, we will reach out to them to help us gain a better insight of their missions and challenges while working with individuals to utilize alternative cooking fuels. We also plan to design the community based survey that will collect data from the targeted communities. The primary goal of our team would be to develop a program code that allows us to execute our application based toolkit with all the intended functionalities. Multiple tasks include: 1) Exploring data from external API providers to understand historical forest cover; 2) Developing linear programming and machine learning models in python to automate the results 3) Collaborate with Duke energy modeling advisor for model fine tuning 4) Creating a frontend display portal and reporting tool to show graphical trends and results on the data.

Current Needs

Our project will require funding for R&D on the various features of the toolkit. Upon the completion of the first prototype of our application, we would require customers to run beta tests that give us the loopholes in the application. It would also be helpful to connect with academic researchers who are working on energy access projects in developing countries and identify a potential advisor to guide us in our project implementation. We will also look into the feasibility of the second phase which includes setting up a real time data acquisition from the field with an array of spatially located remote sensors measuring various aspects of trees such as CO2 and presence of the trees over time.