Small and single batch production, particularly for large (some cubic metres), heavy (several tons), and complex workpieces, is still a challenge for machine tools users (see Figure 1).

To achieve high product quality, timely QC (Quality Control)-production loops are required, which are inefficient and expensive.

The organisation and sequencing is also difficult because the process performance practical knowledge is almost unavailable.

As a consequence, this type of production is still affected by inefficiencies and waste (energy, raw material and time).

SOMMACT approaches these issues by the detection (in-process embedded traceable measurements) and compensation (adaptive control and self-learning) of geometrical effects of varying external and internal quantities, such as temperature gradients and workpiece mass (see Figure 2).

The SOMMACT vision is based on three pillars:

1. A new metrological concept to enhance the measuring capabilities of machine tools, to monitor machine geometrical deformations reliably and to inspect machined parts characteristics traceably.
2. Enhanced sensor systems, measuring the 6 degrees of freedom (dof) of each machine component, and a control system integrating machine and workpiece data with environment and load conditions, and adapting machining accordingly.
3. A self-learning model of the system performance, accumulating knowledge on the machine behaviour, based on calibration and real-time measurement data, and on their relationship with workpiece characteristics (e.g. mass).

The advantages are an improved product quality at competitive costs, and a prediction capability of the system performances based on increasingly reliable model. The machine tool measuring capabilities are enhanced to the point that it can be used as a Coordinate Measuring Machine (CMM). This (i) avoids or reduces QC-production loops, (ii) provides workpiece traceable measurement results and (iii) inputs valuable data into the self-learning core.

SOMMACT will measure the effects of process disturbances on geometric errors of individual machine components, store the corresponding data and associate them with corresponding known disturbances ( e.g. ambient temperature and workpiece mass) and apply quasi real-time adaptation of geometric compensation tables under the supervision of the self-learning core.

Self-optimisation methods will be applied to steadily improve the product quality. Individual stored geometric errors will be combined with (i) on-board workpiece measurement results, (ii) independent CMM measurement results, (iii) timely, swift re-tuning data and (iv) possible full recalibration data.

SOMMACT advancement over the State of the Art are summarised in the following scheme.

Main expected results

• 100% of improvement of product quality;
• 20-30% reduction of total manufacturing time for small and single batch production;
• 15-20% increase of machine availability;
• Strong drive toward “zero defects” target;
• 70% reduction of workpiece moving phases;
• 15-20% increase in acceptable operating temperature conditions.