Manufacturing Sector: White Goods
Keywords: Mobile Device, Decision Support System
The trial address the problem of improving the human decision process in terms of speed and efficacy through an event driven real time communication system that filters events captured in a process such as a production line, generates content rich alarms and send them through a mobile device to interested users such as supervisors, quality mangers etc. The challenges that this trial has to cope with are basically:
- Interface with variety and complexity of real time data series produced in the production environment
- Ensure a good reactivity between event generation and alarm communication in order to reduce delays in decision take
- Allow for a flexible management of data, events, subscriber in order to adapt to different and/or changing situations and increasing mobility of the users.
Innovative Services and Functionalities
- Data handling and Event Generation: gets data from the field and apply basic to medium complexity algorithm defined by end users to generate appropriate alarms (e.g. threshold, SPC rules, etc.)
- Event Delivery: send alarms with context data to mobile devices of subscribers
- Event and Subscribers management: manage subscribers (username and passwords), events, event groups and links.
Impact and Benefits
The trial impacted positively both on intangible and tangible aspects. Decision maker involved in the trial improved their awareness and knowledge of the production plant and this was reflected in the medium terms into the improvement of two basic KPI chosen: internal Fall-Off Rate, representing the amount of defective parts produces and detected, and Service Incidence Rate, representing the overall defectiveness emerged in the market. Both KPI was positively impacted with a double digit improvement in less than 6 months.
The trial took place in a Whirlpool Italian plant insisting on the decision chain related with Quality improvement in the washing machine production line. The production line is made of more than 20 stations most of them involving highly automatic operations driven and controlled by computers and PLCs. The amount of data generated was huge but not properly used by the decision makers, who turn to act in delay and with poor information in response to an unexpected variation in the production line.
To improve the situation, the real time data recorded for a group of critical selected stations has been processed through the FITMAN SMART Factory platform applying basic rules to generate alarms (e.g. machine errors, process drift) to be sent to mobile devices of a group of selected decision makers.
In this new situation Quality supervisors can be now aware of a specific alarm wherever he is and just few second after the event happens. With the context rich data communication he can quickly decide what to do: go to the production line and examine the situation, request the maintenance intervention, get to original data and perform a more detailed analysis and correlation. With this new approach some apparently irrelevant events are now treated properly and the overall behaviour of decision chain has improved.
Technical Challenges & Trial Platform
The main challenge was processing the event flow in real time, and sending out problem notifications within a narrow time window (max 20’’) in order to enable an effective and timely reaction. This is a typical Smart Factory scenario, so our architecture was based on FITMAN’s SMART Factory platform. The baseline system was implemented as a collaboration of three GEs from FIWARE’s IOT Chapter: DataHandling, ConfMan and IOTBroker. The DataHandling component is where the business logic for problem recognition runs: events are fed to it by a custom-built event generator software (the integration layer which collects real-time data from the legacy ICS of the assembly line) and are filtered so that only anomalies are forwarded. ConfMan and IOTBroker act as a transparent decoupling layer, the former as a plug-in hub for event sources, the latter for managing multiple event subscribers. On top of this baseline system, the Whirlpool Mobile Notification trial-specific component adds a presentation layer, transforming raw events into content-rich push messages that reach the Quality Supervisors on their own mobile device.
The Whirlpool solution is entirely cloud-based, with the sole exception of the integration layer (i.e., the event generator) which runs on a local server on Whirlpool’s premises. It was deployed on Engineering’s FIWARE-enabled cloud computing infrastructure, which is currently an experimental lab that paves the way for a production-grade facility that will be available in the future as part of Engineering’s commercial offering.
Whirlpool Europe, Middle East & Africa (EMEA) – with approximately 26,000 employees, a sales presence in more than 30 countries in Europe, Middle East and Africa and manufacturing sites in nine countries, Whirlpool EMEA is a wholly-owned subsidiary of Whirlpool Corporation.
Whirlpool Corporation is the number one major appliance manufacturer in the world, with approximately $20 billion in annual sales, 100,000 employees and 70 manufacturing and technology research centers throughout the world in 2014.
The company markets Whirlpool, KitchenAid, Maytag, Consul, Brastemp, Amana, Bauknecht, Jenn-Air, Indesit and other major brand names in more than 170 countries. Whirlpool Corp.’s European Operations Center is located in Comerio (VA), Italy.
Engineering Ingegneria Informatica S.p.A. [ENG] is the head company of the Engineering Group, headquartered in Rome, Italy. Engineering Group was founded in 1980, and it is currently the first IT group in Italy with over 7.300 employees and 43 branch offices in Italy and abroad. The group delivers innovative IT services to more than 1.000 large clients, with a complete offer combining system and business integration, outsourcing, cloud services, consulting, and proprietary solutions. Engineering Data Centres offer business continuity and IT infrastructure management to about 15.000 servers and 230.000 workstations. Since 1987, ENG’s innovation capability is supported by its Central Department of Research & Development, with around 250 researchers currently involved in over 70 research projects. ENG holds different responsibilities within the international research community, including technical and overall co-ordination of large research projects and consortia. ENG is also an active member of several international open source communities and founder of SpagoWorld, a free/open source initiative managed by ENG. The company is corporate member of the OW2 Consortium and of the Eclipse Foundation.
Reference Papers and/or Books
Petrali, P., An improved decision support system in Factory shop‐floor through an IoT approach. Enterprise Interoperability (2015), Wiley, ISBN: 9781848217997. http://onlinelibrary.wiley.com/doi/10.1002/9781119081418.ch5/summary