JEMS is developing and selling waste-to-fuel transformer plants. These plants are transforming any hydrocarbon-based waste into a high-quality synthetic diesel. In these plants, JEMS uses a chemical- catalytical de-polymerisation process that runs on low temperature and low-pressure. Due to the low temperature, no harmful gasses (like dioxins or furans) are produced as by-products. More specifically, the process temperature of this technology is a few hundred degrees lower than the threshold to produce carcinogenic gasses. Organic waste that can be used includes wood, paper, waste fuel & oil, plastics, textile, rubber, agricultural residues, weed, yard trimmings, cultivated plants, food leftovers, coal, crude oil, and others. The quality of the synthetic diesel fuel is one of the highest. Due to the high cetane index, flash point, low Sulphur content and low clouding point, the synthetic fuel can be used in any modern diesel engine or electricity generator without any negative technical or mechanical impact. It can be used for any modern or older diesel engine for transportation and/or electricity generation as well as for heating. As a result of a chemical process, the chemical composition is stable and can therefore be also used for long- term storage. Furthermore, such diesel can be used as an additive for low temperature use due to its very low clouding point.
The latest such transformer plant is an industrial rate machine for the chemical transformation of organic waste material into high-quality synthetic fuel. The transformer plant has been designed and built for continuous operation. This plant is already using the latest available software and hardware technology allowing remote control and maintenance of each part of the plant and the process itself. However, it does not include any analytics, anomaly detection, prediction or optimisation features. There is a high need for better understanding, optimisation and decision making given the availability of data.
Within FACTLOG, JEMS wishes to upgrade the existing plant with management, predictive and proactive features that will be deployed at the test machine. For this purpose, JEMS will be using the existing and proven intelligent big data processing platform D2Lab developed by NISSA for the creation and management of the cognitive digital twins and Qminer from Qlector for large scale data analytics. This approach is based on a novel integration of the data- and model-based approaches (analytics and deductive reasoning, respectively) enabling the so-called data-model continuum (data analytics generates models, reasoning produces implicit data) that is the basis for achieving the requested self-improvement. It belongs to the novel trend of data-enabled AI, which connects Big Data, HPC and AI, enabling the realisation of the computation intensive AI methods on high performing computation architectures (including edge resources, like GPUs).
Note that such plants are typically installed in rural and remote areas, for various feedstocks and run under different conditions across the globe. Currently they are being operated with highly qualified personnel and with high cost of personnel training. Introducing automation, remote control, optimisation and interconnectivity between the plants, would significantly ease the operation. JEMS intends to install more than 1500 of such plants across the globe in short time which would be impossible in traditional context.