Process Modelling for Industrial Scale Polyhydroxybutyrate Production: A Path to Sustainable Plastic

Process Modelling for Industrial Scale Polyhydroxybutyrate Production: A Path to Sustainable Plastic

Introduction:

The production of polyhydroxybutyrate (PHB), a bioplastic with great potential for replacing traditional plastics, has been a topic of interest in recent years. Researchers have been exploring different carbon sources and their economic viability for large-scale PHB production. In this article, we will delve into the findings of a study that assessed the use of fructose, formic acid, and CO2 as carbon sources for PHB production. Additionally, we will explore the ChemML library, a powerful tool for data preparation, model development, optimization, and visualization in the field of chemical machine learning.

Finding the Optimal Carbon Source:

The study compared the breakeven prices of PHB production when using fructose, formic acid, and CO2 as carbon sources. Surprisingly, the lowest breakeven price of $3.64/kg PHB was obtained when fructose was utilized. This suggests that fructose is a cost-effective carbon source for PHB production. On the other hand, the breakeven prices for formic acid and CO2 were $10.30/kg and $10.24/kg respectively. These higher costs can be attributed to the raw material expenses associated with formic acid and CO2. However, it is important to note that the use of formic acid and CO2 aligns with the emerging sustainable needs for plastic production and contributes to the circular economy through CO2 fixation.

A Step Towards Sustainable Plastic Production:

Despite the higher costs, the study highlights the potential of using formic acid and CO2 as feedstocks for PHB production. With further research and development, these alternatives have the potential to become competitive in the bioplastic market. The use of formic acid and CO2 not only addresses the environmental concerns associated with traditional plastic production but also opens up new opportunities for recycling and carbon capture. By utilizing these carbon sources, the production of PHB can contribute to a more sustainable future.

The Role of ChemML Library:

In the realm of chemical machine learning, the ChemML library plays a crucial role in data preparation, model development, optimization, and visualization. The library offers various techniques to address issues such as one-to-many and many-to-one mappings, as well as feature transformation and selection. These techniques help to remove redundant or irrelevant features, ensuring the accuracy and efficiency of models.

For model development, ChemML leverages popular and efficient libraries like scikit-learn, Tensorflow, and Keras. This allows researchers to utilize state-of-the-art supervised machine learning techniques in the creation of models. Additionally, ChemML offers physics-informed deep learning architectures and pre-tuned machine learning models specifically designed for predicting molecular properties.

Optimization is another key aspect of ChemML. Researchers can optimize models in hyper-parameter space using methods like coarse grid search or evolutionary algorithms. The library also provides ML design methodologies such as active learning and transfer learning to enhance the exploration of compound space. By automating the modeling of specified search spaces, ChemML enables researchers to improve the accuracy and reliability of predictions.

Data visualization is essential in understanding and interpreting modeling results. ChemML offers a separate module that utilizes the Matplotlib and Seaborn libraries for data visualization. This module provides researchers with the necessary tools to visualize key elements of their ML workflow, facilitating a better comprehension of the modeling results.

Conclusion:

The study on PHB production using different carbon sources highlights the economic viability of fructose as the most cost-effective option. However, the use of formic acid and CO2 presents a promising path towards sustainable plastic production. By incorporating these carbon sources, the production of PHB can contribute to the circular economy and address the environmental concerns associated with traditional plastics.

In the field of chemical machine learning, the ChemML library proves to be an invaluable tool for data preparation, model development, optimization, and visualization. Researchers can leverage its capabilities to enhance the accuracy and efficiency of their models, explore compound space more effectively, and gain a deeper understanding of the modeling results.

Actionable Advice:

1. Consider fructose as a cost-effective carbon source for PHB production. Its low breakeven price makes it an attractive option for industrial-scale production.

2. Explore the use of formic acid and CO2 as feedstocks for PHB production. While they may be more expensive, they align with sustainable needs and contribute to the circular economy through CO2 fixation.

3. Utilize the ChemML library for data preparation, model development, optimization, and visualization in the field of chemical machine learning. Its diverse range of techniques and tools can greatly enhance the accuracy and efficiency of your models.

In conclusion, the study on PHB production and the capabilities of the ChemML library demonstrate the potential for sustainable plastic production and the power of machine learning in the chemical field. By incorporating these findings and leveraging advanced tools, we can pave the way for a future where plastic production is environmentally friendly and economically viable.