Response surface modeling to predict biodiesel yield in a multi-feedstock biodiesel production plant.

For economic reasons, multi-feedstock plants are preferred to produce biodiesel. However, the optimal conditions of the transesterification reaction depend on the raw material, thus making difficult the achievement of a high yield of biodiesel when different types of feedstock are used under the same operational settings. In the present work, a response surface methodology is proposed to both predict biodiesel yield when different raw materials are used to produce biodiesel and to determine the optimal operational conditions of a multi-feedstock plant. The optimization of the transesterification reaction of five vegetable oils consisting in a wide range of fatty acid profiles has been carried out. Results provided a compromise zone where all the experimental responses satisfied the imposed specifications to achieve the goals, where the best optimal combination of parameters was selected. According to this model, the tested properties of the produced biodiesel are within the limits of the EN 14214 standard. It can be concluded that this methodology provides the most suitable operational conditions to achieve the highest biodiesel yield in a multi-feedstock biodiesel plant, also considering the economics of the process.