Skip to main content

Coordinating Optimisation of Complex Industrial Processes.

Periodic Reporting for period 2 - COCOP (Coordinating Optimisation of Complex Industrial Processes.)

Reporting period: 2018-04-01 to 2019-03-31

The goal of COCOP (Coordinating Optimisation of Complex Industrial Processes) is to increase the competitiveness of the European process and automation industry. The objectives are to increase product quality, improve sustainability, reduce operation costs, and to improve working conditions. The approach is to consider all operating conditions of subunits and unit processes of an entire plant, and mutual interactions of the subunits in order to improve the overall plant-wide operation. Innovative development of mathematical process models and application of plant-wide monitoring and control by optimisation is the COCOP approach to reach these objectives. A direct application of optimisation forms such a huge mathematical optimisation problem that it cannot be solved with traditional optimisation algorithms within the required time frame. An alternative approach, benefited in COCOP, is to decompose the entire optimisation problem into sub-problems. The proposed concept relies on the facts that truly large systems typically have some definite structure, since they commonly arise from a linking of independent subunits. This structure and the domain knowledge can be utilized while decomposing the control problem.

The importance of the project is threefold. Primarily, it improves the efficiency and environmental impact of European production processes, and thus, the competitiveness of European industry. The project also gives new tools to operators and other personnel to improve their skills and competence, which improves working conditions and make process operator work more lucrative. The project also provides new method for process automation industry and improves it competitiveness.
In the beginning of the project the requirements and the use cases of the pilot plants are defined also considering the broader potential use of the concept in other industries. The results include defining the use cases for both the copper and steel pilots, and applying the social innovation methodology. In addition, key performance indicators for impact evaluation are defined for pilot processes as well as for the social point of view. To facilitate adoption and improve impact the characteristics of operator work and user acceptance factors are also studied.

A generic architecture for the system is designed, and necessary data pre-processing methods and module-specific analysis methods for the information flows are defined. Methods for evaluating data quality, pre-process data as well as tools and methods for feature extraction are developed. The work is demonstrated with prototype implementations of COCOP system components based on an asynchronous loosely-coupled data-driven and event-driven message bus architecture. For the run-time system implementation a message brokering system between COCOP system components is successfully utilized in offline testing and for a laboratory test processes. This verifies the architecture suitable for upcoming online testing during the last phase of the project.

The work done on process modelling contains the development of simulation models that are needed in pilot cases. The unit processes are modeled both for scheduling and for advisory tool purposes. Many of the models are developed from the beginning but many of the models are based on existing models. The existing models might have been originally developed for other purposes and cannot directly be applied in COCOP, and actions to include, upgrade, simplify or transform legacy models to fit the COCOP concept are identified. The optimisation is implemented in a novel way using well established optimisation model development approaches in combination with deep domain specific knowledge. For copper production scheduling an optimising and scheduling prototype is developed. For the steel case models are developed to maximize castability in secondary metallurgy, to achieve the target temperature range in continuous casting, and to minimise the billet head-tail temperature difference before the rolling mill. The coordination optimisation is used to optimise converter batch schedules in the copper case and a Genetic Algorithms based optimization framework is used for the steel case to define the optimal values of the defect-related key parameters for each sub-process. The COCOP adaptation workflow guideline includes a Digital Maturity Analysis and Human Factors Milestones starting from feasibility evaluation, continuing to implementation, and an action plan showing the status of fulfilling user requirements at different stages of the project.

To ensure the maximum impact of the COCOP project the dissemination activities follow the dissemination strategy including publishing results in scientific articles, blog posts, organisation of joint workshops, participation in special interest group events, tweeting project activities and posting updates to the COCOP LinkedIn group. Exploitation of COCOP solution is started by defining potential business models that can be used by partners or other potential companies. The business model definition is based on Österwalder’s business canvas concept that helps the project to describe, analyze, design and develop business models more systematically.
The decomposition optimisation methods can be used for calculating control commands which improve efficiency and reduce operation costs. A less known fact is that optimisation algorithms produce also calculated results which are related to the sensitivity of the solution and gives information of the active constraints and bottlenecks of the obtained optimal solution. These results of the optimisation calculation can be presented to the process monitoring and operating personnel. Presenting these results appropriately to the operator and other personnel improves the overall understanding of the constraints of the process and bottlenecks of the operation, which affects to process operation and efficiency. Based on this understanding the operators can improve their skills and understanding and be more motivated in their work. To adopt the optimisation results to day-to-day work of the process operators also result in the areas of social innovation and co-creation integrated to the COCOP project.

Typically, non-optimal operation, effects of externals disturbances, and unnecessary rework and internal material circulations are compensated with extra fossil fuels in considered industries. Therefore, adoption of COCOP results are in line with the general goals of reducing environmental impacts (CO2 reduction) and use of fossil fuels.

The COCOP system architecture and its implementation is expected to facilitate integration of distributed processes, and provide the required scalability and standardised interfaces for integration of information and control systems used in the industry. The approach enables using the decomposition optimisation method utilising existing control systems and infrastructure, and allows creating new advanced control solutions and applications increasing the awareness of operators and the impact of their decisions. This development facilitates the competitiveness of the European automation and control industry.
COCOP concept