Cognitive Solutions
The Cognitive Automation Platform will coordinate a set of specific cognitive solutions at the various levels of functional organisation of the automation (from planning to sensors) based on the analysis of the different use cases involved in CAPRI.
Planning & Control of Asphalt Production[CAP1]
Status:
CAP1 applies knowledge extraction techniques through the data collected. In this case, algorithms are being developed for data cleaning and identification and selection of relevant variables within the physical model of the plant. Besides, intelligent visualizations have been carried out.Description/Objective:
One of the main problems in asphalt production is knowing the correct composition of the raw materials as well as their relevant properties. The extraction of knowledge will allow the development of an expert system that generates the adjustments of the setpoints of the controllers, the amounts of material to be incorporated into the processes such as the mixer, and other data. This expert system aims to minimize the excess asphalt material in hot containers at the end of the day in order to optimize the production process.Outcomes
Cognitive control of Asphalt Drum [CAC1]
Status:
CAC1, has been developed based on a control algorithm where sensors and actuators are used to calculate the optimum values for the different variables that run the drum, actually a dynamic modeling of the rotary drum is being created through model-based identification methods through several experimental tests performed at the asphalt plant taking into account some of the main variables, this identified model will be required for a Model Predictive Control (MPC) to optimize the rotary drum control calculations.Description/Objective:
The asphalt production process begins when the stockpiled aggregates in the cold feeders are metered and conveyed to a dryer drum where they are heated to a specific temperature. The control objectives are to obtain a dry product at an optimum temperature for the next process and fumes (combustion gases) at the possible lowest temperature, on one hand not to damage the bag filter and on the other to minimize energy consumption, thus increasing the efficiency of the drying process.Outcomes
Cognitive sensor for bitumen content [CAS1]
Status:
CAS1 is currently a laboratory prototype, almost in its final form.Description/Objective:
Asphalt is obtained by mixing aggregates, virgin bitumen and additives. One of these additives is Reclaimed Asphalt Pavement (RAP), which is obtained by removing and milling old asphalt from roads, highways, or pavements. When this additive is added to the mix, the virgin bitumen content is recalculated considering the bitumen already contained in the RAP added. Currently, the bitumen content in RAP is roughly measured offline in a laboratory. Aiming to obtain a precise and automated measurement of the bitumen content in RAP in-line and in real-time, CAS1 will be developed at AIMEN’s laboratories and placed in the RAP line in EIFFAGE’s plant. CAS1 is an optical sensor that will make the asphalt manufacturing process more flexible and automated, reducing decision-making times and improving productivity and product quality.Outcomes
Cognitive sensor for tracking of steel billets and bars [CSS1]
Status:
CSS1 is done in two parts: the first connects the billet from the casting machine to the rolling mill by marking each billet with laser and later reading with visual cameras; the second connects the output from those billets, the bars, from the rolling mill to the finishing units, using laser marking and visual cameras too. In both cases, a close connection with the MES systems and Level 2 systems is necessary. The first part of the sensor is fully operative, whereas the second one has been operatively tested but faces some implementation difficulties due to a revamping in the rolling mill. However, a mitigation strategy has been developed based in one concrete type of orders.Description/Objective:
The final product of steel long producers are bars of different composition, size and length. As the process is discontinuous and the product varies in form, size and number during the whole production operation, the detailed tracking of the final bar and link to the detailed data collected at different moments is a challenging issue. In fact, the amount of data is also very variable, from the liquid steel as one unit, to the intermediate product (a billet) in units of tens up to the final bars in units of hundreds.Outcomes
Steel
Cognitive Solutions
Cognitive control concept[CPC1]
Status:
The development of CPP1 has not yet started. Once other work packages are completed, development will commence.Description/Objective:
Before running a continuous manufacturing line, appropriate process settings must be selected. Therefore, to find these settings, an initial set of trials is executed. Based on the generated data and insights, suitable process settings are determined. As the number and scale of experiments affect the overall sustainability (e.g., energy consumption), the Design of the Experiments (DoE) plays a critical role in minimising environmental impact.Outcomes
Cognitive operation solutions for pharma production [CPO1]
Status:
Several scenarios that typically require human intervention have been evaluated, categorised, and defined. The development of algorithms supporting line operators by suggesting appropriate countermeasures is currently ongoing.Description/Objective:
The operation of continuous manufacturing lines demand highly trained and experienced personnel operating the equipment. Employing a cognitive concept (CPO1), to support line operators in selecting the correct process settings and making appropriate decisions about maintenance, can reduce downtimes and streamline processing while guaranteeing maximum quality.
Cognitive planning solutions for pharma production [CPP1]
Status:
Several scenarios that typically require human intervention have been evaluated, categorised, and defined. The development of algorithms supporting line operators by suggesting appropriate countermeasures is currently ongoing.Description/Objective:
The operation of continuous manufacturing lines demand highly trained and experienced personnel operating the equipment. Employing a cognitive concept (CPO1), to support line operators in selecting the correct process settings and making appropriate decisions about maintenance, can reduce downtimes and streamline processing while guaranteeing maximum quality.
Cognitive sensor for blend uniformity [CPS1]
Status:
CPS1 covers the construction and realisation of the mechanical interface needed for integrating the Raman probe in the process. Functionality utilising an electric motor has been developed, to properly present the monitored material to the probe. Furthermore, the algorithm for computing the API concentration was developed, and the programming of the motor actuation was completed. Currently, tests with the prototype on the ConsiGma CTL25 (GEA) manufacturing line at RCPE are ongoing.Description/Objective:
Continuous manufacturing of pharmaceutical oral dosage forms, like tablets, requires the implementation of appropriate quality control concepts. The real-time monitoring of critical quality attributes is an essential component of a quality control concept. The CPS1 solution supports prediction of the concentration of active pharmaceutical ingredients (API) in the material stream during manufacturing, using Raman spectroscopy. The availability of that information enables the development of sophisticated quality control concepts, ultimately reducing waste material and improving product quality.Outcomes
Cognitive sensor for granule quality [CPS2]
Status:
The project team has accomplished the mechanical setup of the Parsum particle probe and developed algorithms for computing characteristic size values from the captured size distribution. Initial tests using the generated data in a process control concept have been successfully completed.Description/Objective:
Continuous wet granulation is a commonly used unit operation in secondary pharmaceutical manufacturing. Its purpose is the creation of granules of well-defined size distribution. As the granule size affects critical quality attributes, real-time monitoring is of the utmost importance, reducing defective end products. The development in CPS2 provides real-time measurements of the size distribution. This data is fed to a process control concept (CPC1), which selects optimal granulation settingsOutcomes
Cognitive sensor for product moisture [CPS3]
Status:
The mathematical model, the core component of CPS3, has been developed. Currently, the parametrisation of the model is ongoing, and the project team is executing trials to generate the necessary initial data set. Once we have completed the preliminary setup, we will generate another data set, to validate the 'theoretical' granule moisture estimates using experimental data.Description/Objective:
In a pharmaceutical “from powder to tablet” manufacturing line that uses wet granulation, drying is crucial to produce intermediate granules with well-defined moisture content. However, typically, there is no direct measurement of the granule moisture. Using available sensor data, such as temperature and humidity, as well as a mathematical model, CPS3 aims to estimate an effective granule moisture measurement.Outcomes
Cognitive prediction of dissolution [CPS4]
Status:
The development of the dissolution prediction model requires the availability of experimental data. Therefore, we executed several trials using RCPE's continuous wet granulation and tableting line. During these trials the specific dissolution profiles were measured. Based on the generated data, we have created prototype models which will be validated against a set of additional experiments.Description/Objective:
The dissolution is essential to the intended performance of a pharmaceutical tablet. It describes how quickly the active pharmaceutical ingredient (API) dissolves in a defined solvent. Therefore, it is one of the most critical quality attributes that must be monitored and controlled in pharmaceutical processes. The dissolution rate is highly dependent on processing conditions. The purpose of CPS4 is to predict the dissolution profile from process settings and available sensor data.Outcomes
Cognitive sensor for fault detection [CPS5]
Status:
The CPS5 algorithm requires large amounts of process data. It requires both data from fault-free operations and also operations showcasing faults, e.g., clogging of filters. We generated this data by running a number of dedicated experiments on RCPE's ConsiGma CTL25 (GEA) manufacturing line. Descriptions of the faults that occurred were manually transcribed into a suitable JSON file format, allowing easy import during algorithm development, which is currently in progress by our colleagues at Nissatech.Description/Objective:
The reliable operation of all involved unit operations, at all times, is essential in continuous pharmaceutical manufacturing lines. Faults and defects must be detected as early as possible, before causing any additional product deterioration. CPS5 aims to detect these potential faults using available sensor data.Outcomes
