Using big data to select the best catalytic materials to transform fish waste into valuable chemicals.
Due to increased population and heavy industrialization of nations, energy is becoming a very crucial resource. Due to the limitations on non-renewable energy sources, clean and renewable energy production is crucial to maintain the industrialization of the world as well as to reduce its effect on climate and environment.
Currently, the development of new energy carriers is based on, to a large extent, trial and error methodology. This approach has several drawbacks, such as that it has a high cost associated with a lot of failed trials, it has a long-time dependency due to the infinite combinations of possibilities to produce the energy, and it produces a lot of undesired products, like, toxic, inhibitors, etc.. A way to overcome such problems is to do a thorough analysis of previous work, as far back in time as possible, to extract relevant information and use these findings to reduce the current trial and error domain and time into a much smaller domain with higher ratio of success.
Raw material selection.
The raw material selected is based on its high relevance for the Norwegian market and the Norwegian society, within Norway, aquaculture has produced a revenue of over 2000 NOKb with in 2016 . These large sales have a great component of salmon and trout income, reaching a total of 70 NOKb. Fish production has reached a production of 167 million tons of live weight, 55% of this is counted as fish and the remain is accounted for aquaculture. From the fish part, 75% of the fish produced are processed before they are delivered, and they produce up to 80% (of the total) of waste [2-3]. One of the major problems that limit the use of this kind of biological waste is its variable nature. These wastes contain protein (up to 60 %), fat (up to 20 %) and minerals (calcium and hydroxyapatite from bones and scales). Palmitic acid, oleic acid, and monosaturated acids are also abundant in fish waste streams (22 %) .
Selecting a proper catalytic material for a specific reaction, the transformation of different waste into biofuels, could be tedious and complicated due to the broad and vast amount of alternatives. Besides those that are known there are a lot of new possibilities that are unknown. The catalysts that will be taken into consideration are metal oxides such as MgO, among others. These are catalytic materials that can be produced from renewable sources and therefore reducing the cost and environmental effect.
To carry out the desired reactions, we need to know the domain in which the raw materials and catalyst can work together to produce the biofuel. The combinations of possible scenarios are extreme and therefore it needs to be reduced, not only because it is impossible to run such an infinite number of scenarios but also to narrow the operational conditions towards those that are given higher yields of production at the lower energy consumption.
In order to complet this project, we have established a few steps to be carried out. As mentioned the raw material has been selected, fish oil, the catalytic material domain has been reduced and the operational conditions have only some upper and lower boundaries based on activation energies (energy that allows the process to happen) and upper limit energy (based on the process being too energy demanding).
Nevertheless, the domain for this project still remains considerably large. Therefore, within this project we will carry out the following steps
To reduce the working domain, data mining of previously published work will be carried out. For this purpose algorithms will be used to search for those publications’ abstract that have been working with fish oil materials, in the presence of different alcohols (methanol, ethanol, propanol) under different operational conditions (here we will focus on pressure, temperature, molar ratio of reactants, catalyst amount and reaction time). Based on this mining we will be able to determine the different domains that have been used and the outcome of each of these experiments. This information will allow us to prepare a model that could predict the outcome under different conditions and therefore reducing the experimental work and increasing the success towards the desired product.
The chemical transformation of the fish oil waste into the desired biofuels will be evaluated under a smaller domain pre-selected by the previous step. The main effect of the main relevant variables based on their effect from previous works will be evaluated using a Design of Experiment and a response surface methodology. We will focus on the three major variables and experimental work will be carried out in order to verify the predicted values from the models from the previous step. Based on the Response Surface Methodology, an optimization of the process will be done in order to obtain the best yields of the desired product.
When evaluating the new catalytic materials for the new working conditions, the changes that the catalyst will suffer due to the reaction will be evaluated. This will be done using different characterization techniques such as SEM, TEM, microscopy, BET, and other techniques to see and quantify the material properties such as porosity and size.
Perform kinetics modeling over the operational conditions and obtain kinetics parameters for the reaction.
Based on experimental information obtained from the previous objectives a mathematical model for the reaction kinetics will be developed using different methodologies such as the pseudo equilibrium as well as the PSSH. The candidate will proposed different mechanism, establish the reactions pathways and obtained the expression that will be further compared with the experimental data. This comparison will allow us to estimate the kinetics parameters and energies involved that will then be compared with literature.
A PhD position is available for the project.
Publications from this project can be found here
Professor Alejandro Franco
Université de Picardie Jules Verne, Amiens, France
Project leader and main supervisor of the Ph.D candidate
Click above to findspecific contact info.