cropped-logo.png

Use-cases

Use cases

Transforming industrial equipment powered by AI to improve agility, sustainability, and resilience throughout the product lifecycle.

This section identifies the 4 industrial scenarios in which the AIDEAS Solutions will be demonstrated. Being all of them industrial equipment manufacturers, they belong to different sectors, so different problems will be handled.

Rovereto – Italy

AI for Machining Centres​

PAMA SpA is a leading global manufacturer of boring-milling machines and machining centers for sectors such as Energy, Aerospace, and Machinery. It produces and machines the main structures in-house, with quality control at different stages of the process. It currently collects process data and sensor signals from its machines (both in its plants and at customer facilities) and stores them in a centralized data lake. PAMA wants to incorporate AI and other technologies to integrate and manage the entire life cycle of its equipment, from design and manufacturing to customer use and reuse, with the aim of positioning itself as a circularity-oriented manufacturer and differentiating itself from the competition through higher value-added services.

Pilot 1: PAMA SpA – AI for Machining Centres (metal sector)

PAMA SpA is a worldwide manufacturer of boring-milling machines and machining centres and is nowadays leader among large machine tools manufacturers for the Energy, Aerospace and Machinery sectors. All main structures are machined in-house, using PAMA boringmilling machines and machining centres. Quality control is carried out at the various machining stages. PAMA currently collects relevant process data and sensor signals from its machines (located at company shopfloors and at customers sites) and stores them in a centralised data lake. The company is interested in integrating AI and other technologies that could support the management and integration of design, manufacturing, customer use, and reuse of its equipment. Their motivation is to position themselves as a circular-enabled manufacturer and provide added-value with respect to their competitors.

The current practices in the different stages of the product life cycle are described below:

  • DESIGN phase: Performance of precision machine tools and manufacturing equipment are currently evaluated and optimised within the early design stage through advanced multidisciplinary CAE simulation tools. This often requires a very large amount of time for a proper product modelling, and often some machine performance is not correctly evaluated; this in turn will slow down the design loop before converging to an optimal design solution.
  • MANUFACTURING phase: Production of parts is done internally or through external suppliers. Manufacturing errors are recorded by the operators (that inform the design department) and sometimes (not very often) stored in PCs (legacy data). No historical tracking of design/production errors is done and therefore no knowledge extraction from events is performed.
  • USE phase: Geometric and especially thermal errors (due to extended usage of the machine that causes heat generation at the moving elements and this heat causes expansion of the various structural elements of the machine itself), constitute a significant portion of the total error in a machine tool. Machine Tools’ (MT) users of course implement some strategies, more or less effective, to compensate for such errors. However, such strategies are generally tailored to a specific application scenario and suffer the changing of machine configuration, especially in the case of renovation or/retrofitting at end of first life (EO1L).
  • REPAIR/REUSE/RECYCLE phase: Repair is implemented on the basis of run-to failure strategies (and preventive maintenance tasks are scheduled on the basis of best maintaining practice for a specific component / group / equipment). In case of sudden (unexpected) failure collateral damages can occur that cause an increase of maintenance and operative (loss of production) cost. Moreover Reuse of a machine tool is not yet a common practice since there is no deep knowledge regarding the actual ‘health’ state of machine parts/components, at the end of life, and then no rationale behind the sustainable extension of the life of the overall machine

PAMA will exploit the potential of the AIDEAS AI technologies to improve sustainability quality and resilience of its products to foster AI-based digital transformation within the manufacturing sectors. PAMA aims to use the AIDEAS Machine Passport to conform large datasets consisting of timestamped signals sensors and CNC/PLC data collected during the use phase (primarily) as well as other phases in the life cycle. Unified standard service modelling techniques will ensure the aforementioned data compatibility, interoperability, consistency and quality. These data sets will be used: i) to improve the performance of the machine learning models that predict machining process anomalies/deviations with respect to nominal conditions, and ii) to improve designs, fabrication and repair/reuse/recycle of the equipment. Expected contributions of AIDEAS AI Solutions to the different stages of the product life cycle are described below:

  • DESIGN phase: AI-based (data-driven) prediction of products (machine tools) performance within early design stage. Such a prediction will enable smarter and faster optimisation, reduce product development time, and boost digital transformation of industry. The challenge is represented by a robust and confident prediction of new designed machine performance through an AI data-driven approach that parses design space variables (e.g. size, strokes, materials, system architecture, power, etc.). For this purpose, the AIDEAS Machine Design Optimiser solution will be exploited and validated. The AI model is trained with a consistent (labelled) dataset of historical results obtained by correlation of design parameters values and machine performances (e.g. precision, chip removal capacity, cycle time, Kv, Jerk…) actually obtained in real field applications.
  • MANUFACTURING phase: Retrofit the design with the outcomes of the manufacturing process, to optimise the future designs. In particular exploiting AI tools to parse historical data (related to tracking of production problems) to find some insights, in particular to correlate measuring results of single produced part (parallelism, straightness, planarity, perpendicularity, etc.) to the overall accuracy of final assembly. For this purpose, the AIDEAS Fabrication Optimiser solution will be exploited and validated. The AI model is trained with a consistent (labelled) data set obtained by correlation of outcomes of single part measures with part assembly ones.
  • USE phase: Exploiting AI to predict (and then compensate through an adaptive control) the geometric and thermal errors of machine/equipment during machining/production of parts. The prediction model is based on input of process variables (e.g. drives currents, absorbed powers, temperature, etc.) as well as design & performance variables. For this purpose, the AIDEAS Condition Evaluator and the AIDEAS Anomaly Detector solution will be exploited and validated. Predict energy consumptions and correlate to expected errors as well as generate clusters of optimal working conditions (minimisation of errors and energy). Process parameters responsible for any deviation with respect to optimal condition will be then automatically re-tuned through AIDEAS Adaptive Controller solution (via CNC).
  • REPAIR/REUSE/RECYCLE phase: At end of 1st life of the machine, exploit AI-tool to assess how the machine has performed during its life and then identifying which parts (including pieces of SW) could still be reused in other machines, identify smart retrofitting strategies (optimal remanufacturing/renovation approaches that take into consideration energy consumption, safety level, maintainability, productivity, and increased technological level) and estimating operative cost and effective maintenance plan for the second lifetime period. For this purpose the AIDEAS Smart Retrofitter and AIDEAS LCC/LCA/S-LCA and the AIDEAS Disassembler solutions will be exploited and validated.

PAMA KPIs

Oleiros – Portugal

AI for Cutting Machines

D2 Technology is a leading European stone technology company whose main focus is to develop new equipment that helps improve the performance and quality of companies in the sector. The range of equipment is divided into two categories: Production and Environment. The Production Department includes CNC’s, Centre machines, Waterjets, Saw Jets, or Edge Polishing machines, while the Environment department includes systems such as waste recycling, water and dust treatment.

Pilot 2: D2 Technology – AI for Cutting Machines (stone sector)

D2 Technology is a Portuguese SME company, founded at the beginning of this millennium that has been evolving the stone industry throughout its existence. It is a high technology company in the stone sector and today it is especially characterised by the design and production of various innovative CNC equipment for the stone sector. D2TECH has branches in Brazil, Germany and USA, as well as distributors in other countries. D2TECH has defined a set of interventions that will allow achieve a set of strategic objectives in the medium and long term, that coherently contribute to the mission and vision outlined: increase turnover by at least 40%, increase turnover exports up to 55%, diversify export markets by at least three more countries, participate in international events/fairs in order to enhance the corporate image and diversify the strategy of marketing of the company
Handling complexity in the configuration of natural stone patterns is a daily challenge addressed by D2TECH. Due to exclusive patterns and unique finishing, the final aspect of each surface where natural stone is applied is highly dependent on the raw material physical behaviour and the transformation process stages. The current practices in the different stages of the product life cycle are described below:
  • DESIGN phase: Currently the design and conceptualisation of new CNC machines, but also the improvement of current models, is performed by engineers and technicians with the support of CAD/CAM software tools (ex. SolidWorks, DDX – EasyStone, Heglmeier). Such tools don’t provide the knowledge about the behaviour of such machines within its usage, regarding common anomalies or most deteriorated parts. There’s no kind of feedback mechanisms that could support designers in the conceptualisation of new CNC machines.
  • MANUFACTURING phase: A CNC machine is a finely crafted tool that needs correct alignment to function at its best. When a machine is not properly calibrated, the end product may have deformities, or the dimensions may be wrong. Even worse, a machine that is perpetually run out of alignment will have a shorter lifespan. Calibrating a CNC machine is typically performed by humans adopting a trial and test approach, which is a timeconsuming task and can lead sometimes to different types of problems.
  • USE phase: In order to better plan the cutting process, companies need to know on a continuous basis information about the stone in a raw shape (technical data about composition of the stone). Companies aim to start earlier planning and adjust machine settings for cutting, polishing and finishing processes. With respect to the production of the finished product, one of the main barriers is to check the stone properties (quality, dimensions, patterns, among others) since it is a complex natural material. For this reason, it is necessary to monitor the stone, searching for natural cracks, voids, natural defects and patterns. Frequently, during the cutting-phase, a stone can be damaged which requires another stone to be cut, which will negatively affect the pattern initially specified by the final customer. There’s a need to find another stone with similar characteristics, which most of the time is impossible, or readapt the combination of finished stones, so that the pattern can best match the pattern requirements established by the end user. This process is currently performed by human visual inspection.
  • REPAIR/REUSE/RECYCLE phase: With respect to repairment, D2TECH adopts a preventive maintenance scheme which is followed by every customer. However, D2TECH does not have in its software the preventive maintenance methodologies at its full capability, e.g. the right tools to assess if a certain component will reach the end of life and prevent failure. Such preventive measures also include the remote installation of updates to the machines software.

D2TECH will address the AIDEAS Machine Passport as a driver to support the machine life cycle. Expected contributions of AIDEAS AI Solutions to the different stages of the product life cycle are described below:

  • DESIGN phase: By learning from the different patterns of machine usage from their customers, D2TECH would be able to collect such knowledge and use it as an input for the design of new machines. By using an AI-driven approach, common anomalies detected during machine usage could be identified and a root-cause analysis could also be addressed. An AI-driven approach could assist the designer during conceptualisation of the machine by providing suggestions and notifications highlighting how to improve assembly sequencing orders, the most common anomalies and the components which are affected. Within the design phase it is expected to rely on the AIDEAS CAx Addon and AIDEAS Machine Design Optimiser solutions.
  • MANUFACTURING phase: To use an AI-driven approach to improve the calibration process of CNC machines. Regarding the manufacturing phase, it is expected that the AIDEAS Fabrication Optimiser solution will support workers in optimising the machine assembly process, but also to find the best tuning for CNC calibration which is a tedious and error prone process. This solution will be used in the calibration process, by setting the optimal configuration of parameters described in the “before AIDEAS” section. For the sourcing of different suppliers of raw material, this phase will rely on the AIDEAS Procurement Optimiser.
  • USE phase: The main motivation of this phase is twofold: (i) Improve short-time production planning; and (ii) Improve machine efficiency (join several orders for similar types of stone). The production of the finished product can be divided into two different sub-phases (configuration and cutting). The objective would be to learn from operators’ experience and propose a set of configuration parameters according to each type of stone to be processed. Stone cutting process takes into account a final representation of the final product to be produced, which must reflect the requirements specified initially by the final customer. The motivation is to be able to “scan” each stone individually, creating a digital representation of each stone that is unique. An AI-based approach would be able to react in real-time, by a new combination of tiles, and also the best matching stone to replace a damaged stone. For the initial setup and calibration of the machine, which needs to be adapted for the different types of stones, the AIDEAS Machine Calibrator solution will be considered. For assessing the final product quality, in case of readapting to a new combination of tiles to match the initial pattern defined by the end user, three solutions are being considered: AIDEAS Adaptive Controller, the AIDEAS Anomaly Detector and AIDEAS Quality Assurance.
  • REPAIR/REUSE/RECYCLE phase: Regarding recycle and repair phases, D2TECH aims to establish a servitisation approach which enables the collection of usage data from their machines, in order to optimise and readapt their maintenance programmes to be better tailored to the different needs of the customers. In that sense, using an AI-driven approach it would enable the detection if a particular component reached the end of its life or if it can be repaired. The adoption of the AIDEAS Prescriptive Maintenance, would greatly extend the condition maintenance among its clients. Within this phase the pilot will mainly rely on the following solutions: AIDEAS Smart Retrofitter, which will enable to detect, in case of malfunction or outdated machines, can be retrofitted; AIDEAS LCC/LCA/S-LCA and the AIDEAS Disassembler will enable to easily identify which components have reached the end-of-life and the ones that can be repaired.

Langenberg – Germany

AI for Blow Moulding Machines

BBM Maschinenbau is a German mechanical engineering SME specialising in the design and manufacture of extrusion blow moulding (EBM) machines for the plastics industry, used to produce hollow parts such as bottles, jerry cans, drums, water tanks and even small boats. Founded in 1998, the company has stood out for technical innovations such as replacing hydraulic systems with electrical components, and has expanded its portfolio with more advanced and scientifically optimised extruders and extrusion heads. In response to the sector’s growing focus on sustainability, BBM is promoting new technologies geared towards the circular economy, encouraging the intensive use of recycled plastics to reduce waste.

Pilot 3: BBM Maschinenbau – AI for Blow Moulding Machines (plastic sector)

BBM Maschinenbau is a German SME mechanical engineering company, specialised in the designing and manufacturing of extrusion blow moulding machines (EBM) for the plastics industry. EBMs are used for producing all kinds of hollow plastic parts like bottles, canisters, drums, special technical parts, water tanks or even small boats. Since it was founded in 1998, BBM has taken a leading role in technical innovations like replacing hydraulics by electrical components in its machines. Until today, the product portfolio has expanded by the usage of technically advanced and scientifically optimised extruders and extrusion heads. Since the plastics industry is paying more and more attention to ecological and sustainable aspects, BBM is also taking part in finding new and better technologies to support the circular economy by extensive usage of recycled plastics for waste reduction
The current practices in the different stages of the product life cycle are described below.
  • DESIGN phase: Until now, BBM’s core competence lies in the mechanical engineering of its machines. The machine control and HMI (Human-Machine Interface) is purchased as a license and only the sequence programming is realised by BBM. Programming and debugging are exclusively performed at the machine which leads to extended start up phases. At the moment, BBM is working together with a software development company to create its own industrial logic control and visualisation. In this context, BBM wants to implement state-of-the-art technologies and set an industry control benchmark not only for the plastics industry. During the mechanical engineering and design phase, BBM is working together with IANUS. By using complex simulations, flow channels in the extrusion heads are optimised to increase the possible regrind share and general efficiency and to decrease the times for product colours changes.
  • USE phase: Handling of extrusion blow moulding machines requires a good knowledge of process engineering. Starting from first calibration and referencing of drives, ending with evaluation of product quality and manually adapting machine parameters, all actions are done manually by the operator.
BBM will use AI technologies developed in AIDEAS to enhance the agility, sustainability and resilience of EBMs through their entire life cycle. In addition, the AIDEAS Machine Passport provides an important database for the optimisation of all life stages of an EBM. Expected contributions of AIDEAS AI Solutions to the different stages of the product life cycle are described below:
  • DESIGN phase: BBM aims to improve its current extrusion head design by using the AIDEAS CAx Addon and the AIDEAS Machine Synthetic Data Generator. Instead of just simulating and verifying designs created by mechanical engineers, the AI technologies shall be used to set-up and improve the first design, thereby increasing the expected regrind share of the final product and reducing the workload for the mechanical engineer. By increasing the regrind share, the sustainability of the whole plastics sector can be improved. By using the AIDEAS Machine Design Optimiser, the cycle of machine control creation and debugging will be improved and the expected software start up time will be reduced substantially. The AIDEAS tool will help to create a more realistic machine simulation that helps the electrical engineers to test and debug the machine control before switching it on for the first time. Due to a high amount of dynamic and fast moving machine parts, AI based technologies will help to improve the design and resilience of EBMs.
  • USE phase: During the use phase of an EBM, several AIDEAS AI tools can lead to a significant improvement for BBM as well as for the customer. At first start up, the machine and process parameters need to be set to achieve the best product quality. This complex task will be performed automatically by the AIDEAS Machine Calibrator and it will also help to decrease the initial start-up time and downtime of the machine in the production phase. After a power shut down, the referencing of drives will be handled automatically by the Machine Calibrator based on the previously collected machine data. A running machine in production needs constant condition evaluation and adjustment of parameters, for example due to thermal changes. The AIDEAS Condition Evaluator will create a machine score to easily evaluate the productivity during the usage. The early detection of abnormal fault conditions of an EBM with the help of the AIDEAS Anomaly Detector will minimise expensive downtimes and time-consuming production halts. And by the subsequent use of the AIDEAS Adaptive Controller, all necessary machine parameters will be adapted automatically. The final product quality will be assessed by the AIDEAS Quality Assurance. Quality deficits can thus be efficiently detected and the manufacturing process can be readjusted accordingly.

Alcoy – Spain

AI for Inspection Machines

Multiscan Technologies is a company located in Cocentaina (Alicante, Spain) that develops and manufactures artificial vision equipment for the sorting of fresh fruit and vegetables and the inspection of food products. With more than 20 years of experience and a presence throughout the international market, we propose unique solutions for the agri-food market by combining technologies related to machine vision and X-rays. Our solutions have a differential value because they analyze, through different techniques, the entirety of each product. Our team is composed of more than 80 people in three different locations: the headquarters located in Spain, and two subsidiaries in North and South America for sale and service those markets. Our value proposition is based on four key concepts: Customer, Innovation, Global Reach, and Sustainability.

Pilot 4: Multiscan Technologies – AI for Inspection Machines (food sector

Multiscan Technologies is a Spanish SME which provides food tooling manufacturing equipment with state-of-the-art machine vision technologies along with innovative product transport systems to achieve optimum sorting. MULTISCAN provides unique computer vision and X-ray solutions for the fresh fruit and vegetables market, mainly quality inspection machines for grading and sorting processes, as well as for safety and conformity applications. MULTISCAN expertise is focused on tooling for small fruits like olives, cherries or cherry tomatoes, being olives its main relevant business and becoming the specialist in stoned fruits of up to 45 mm in diameter. Their products make an extensive use of machine learning models to detect product features like colour, shape, or size from captured camera images and X-Ray detection. Besides these data sources, the machine controller developed by the company collects large amounts of process and product data from auxiliary manufacturing equipment, like feeding engines, and conveyors. Currently, the company is developing a cloud solution to deliver data-driven added value services to their customers based on this data, through a vertical solution that implements open data services, MES (Manufacturing Execution System) applications and predictive maintenance on top of their manufacturing equipment.
The current practices in the different stages of the product life cycle are described below.
  • DESIGN phase: The company develops the software (including the machine learning models and algorithms used for grading and sorting, the industrial control logic and the HMI) and the hardware design of their inspection machine. MULTISCAN also designs some of the critical electronic components of their systems, specifically the lightning PCB (printed circuit boards).
  • Manufacturing phase: HMULTISCAN subcontracts the manufacturing of the electronic and mechanical components, including the lightning PCB they design. It is important to highlight that there are rather few providers for some of the critical components of the inspection machines, and due to this, the same providers supply to the other competitors. For these reasons, MULTISCAN is not eager to exchange sensible information with them. These components have a specified lifetime which is then part of the specifications of the product. Sometimes the behaviour of this PCB depends on its manufacturing process and the electronic device, for this reason the quality control of this device must be done when the components arrive at the company. MULTISCAN systems are assembled at their job shop in Alcoy (Alicante). The company has a strong continuous improvement culture and applies lean manufacturing methodologies to improve the production process. Assembled machines are shipped to their network of integration partners, which are in charge of calibrating the products and installing them on customer premises. They also integrate and install entire food processing lines (i.e. they act as an integrator in a secondary line of business).
  • USE phase: The calibration of the inspection machine is crucial to adapt to the lightning conditions of the environment and maximise the efficiency of the computer vision algorithms. In the set-up, operators need to configure the sorting process, specifying the features (colour, size, shape) of the product directed to each output. An incorrect configuration may yield high efficiency losses. Besides this, the main downtime causes are due to machine failures due to wearing of components. An embedded line PC implements an OPC UA Server, that provides an interface to other systems, can be used to integrate with the inspection equipment. The controller also collects data from ancillary manufacturing equipment in the line, like feeding engines or conveyors. The OPC UA server publishes these manufacturing data (around 275 variables in total). Currently, MULTISCAN is developing a cloud platform to provide data-driven added value services on top of this data, related to manufacturing operations management (e.g. production KPI monitoring) and predictive maintenance.
  • REPAIR/REUSE/RECYCLE phase: When a customer purchases an upgrade, the decommissioned hardware is assembled into units that are sold to other customers, typically in regions like Africa or South America. MULTISCAN collects the equipment and inspects its components to identify those that need to be replaced or upgraded to new versions.
MULTISCAN will exploit the potential of the AIDEAS AI technologies as a transformation tool to improve sustainability, agility and resilience of its customers providing AI-based industrial equipment. MULTISCAN aims to use the AIDEAS Machine Passport to conform large datasets consisting of labelled captured images, industrial variables and X-Ray inspection results collected during the use stage. Expected contributions of AIDEAS AI Solutions to the different stages of the product life cycle are described below:
  • DESIGN phase: MULTISCAN will use the AIDEAS Machine Design Optimiser to optimise the lighting conditions of the computer vision system based on physical models. The modules will be integrated in CAD tools (through the AIDEAS CAx Addon) to ensure a seamless adoption of the optimisation modules by the engineering team. The AIDEAS Machine Synthetic Generator will be used to simulate operating conditions not found in the AIDEAS Machine Passport datasets, for instance when the product design needs to be adapted for new types of fruits.
  • Manufacturing phase: The main objective is to use the AIDEAS Fabrication Optimiser to optimise the assembly and quality control processes, supporting the agile manufacturing methodologies used currently in the factory with fast planning and sequencing tools. This will help the company make an efficient use of their human and technical resources while at the same time shortening the delivery dates of the production orders. Additionally, the AIDEAS Procurement Optimiser will be used to minimise component stocks and ensure material flows.
  • USE phase: Based on the datasets obtained through the AIDEAS Machine Passport, the objective is to automate the initial calibration of the lightning PCBs, learning the relationship between the information collected on customer premises describing the installation and its environment and the optimal configuration parameters of the PCB lightning board. The approach is to use the AIDEAS Machine Calibrator solution to calibrate the machine parameters, yielding shorter installation times and requiring no advanced skills nor training for the installation. Another objective is to use the AIDEAS Adaptive Controller to detect and correct errors in the set-up of the machine and the AIDEAS Quality Assurance time series forecasting features to predict the quality breakdown (quantity of each product category produced), from the information coming from the sensors placed in the machine, the surrounding environment, and other connected systems. Moreover, the labelled images in the datasets will be used to improve the performance of the machine learning models that perform the sorting and grading processes. The availability of large datasets will significantly improve the system resilience and the ability of the models to learn the properties of products under different conditions, allowing to work more efficiently with new varieties or under new seasonal characteristics.
  • REPAIR/REUSE/RECYCLE phase: The AIDEAS Smart Retrofitter will be used to improve the retrofitting of components, allowing to effectively detect components that need to be replaced and facilitating the inspection process in decommissioned equipment. The AIDEAS Prescriptive Maintenance can use the AIDEAS Machine Passport datasets to estimate the remaining useful life (RUL) of mechanical components and maintenance plans adapted to the individual characteristics of the machine.