CAPRI: Paving the road ahead towards transforming process industries

Cristina Vega, as project coordinator from CARTIF technology centre, summarises the motivation behind the CAPRI project perfectly. “The world needs to change,” she says. “We need to reduce consumption of all the raw materials we use and reduce the energy needed to process these materials into intermediate or end-user products. The digital transformation of our manufacturing and process industries will help us do just that.”

The CAPRI project aims to advance this transformation of industrial processes by developing and testing a novel Cognitive Automation Platform (CAP), enabled by cognitive solutions like sensors, information technologies, and using data analysis and the use of data in real time to aid decision making for planning, operation and control. This will in turn help improve flexibility and performance in processing plants as well as improve the quality of products and intermediate flows.

The architecture of reference for this novel CAP is being defined by FIWARE open-source technology for smart solutions to ensure replicability and scalability of the technology across a range of industries. The CAP architecture is modular, and supports knowledge models, machine learning and different cognitive modules for logistics and planning, operational management and process control. It is also scalable, allowing further advanced applications to be developed and added, making it replicable for all SPIRE sectors.

Plants are connected to the smart modules – smart events processing, smart knowledge modelling and smart decision support – through a smart IoT connection and controlled through the CAPRI user interface. The system is due to be demonstrated at three types of plant – steel, asphalt and pharma – and during this phase, the lessons learned will be applied to enable further development of autonomous systems for the future operation of plants based on the combination of embedded cognitive reasoning and optimised human-driven decision-making.

“At this stage of the project, just over halfway through, we have all the cognitive solutions working,” says Vega. “Some of these cognitive solutions are only working at laboratory scale, but we are pleased that most are already implemented into the pilot sites and we are already getting good results.

“So far, we have 39 results from these 19 cognitive solutions prototypes and these have all been added to CAPRI’s Zenodo channel, to ensure we fulfil the FAIR principles of Open Research Data Pilots, which are findability, accessibility, interoperability and the reuse of digital assets.

“We have also identified eight key exploitable results in this work to date,” she continues. “For example, in the asphalt use case, we have identified the two sensors that are currently unavailable on the market. These measure things like the content of bitumen in the recycled asphalt, or the filler present in the cold aggregates per batch.”

Another key part of CAPRI’s ambition is to adapt the 6P methodology for digital transformation for process industries, initially created for discrete manufacturing plants, allowing those that have identified a need to digitise their processes the opportunity to develop a workable roadmap for that transformation. The 6P methodology is a tool that provides analysis of the six main dimensions of an industrial process that need to be assessed to determine readiness for digitisation in any enterprise. These dimensions are product, process, platform, people, partnership and performance (the six Ps) and these are grouped for analysis into three technical pillars and three socio-business pillars.

In CAPRI, this methodology can be applied at two levels – the enterprise level for the development of a transformation plan along the six dimensions, and at the pilot or experimental level to show expected improvements digital transformation will make after the pilot ends.

The project is using the 6P methodology at its pilot studies and hopes to be able to measure the impact of the experiments and identify the dimensions that are improved by digital transformation, thereby providing measurable success stories and best-practice roadmaps for others to follow.

“Developing common structures like the CAPRI CAP, and its reference architecture, that can be replicable in other sectors, will help everyone involved in process industries and manufacturing achieve our common objectives of reducing energy use and waste, reducing the consumption of raw materials and lowering our CO2 footprint,” concludes Vega.

“Sensors are easy to install in industrial plants and it is easy measure the data they produce. The big questions we are addressing are what can we do with the data, where can this data be used and how can this data be collected and processed? Now, with the CAPRI CAP, we can begin to answer these questions.”