Product Insight: How to Increase Productivity and Customer Satisfaction with Machine Monitoring

Satisfaction with Machine Monitoring

Machine monitoring - what means renewal of the own service strategy for the machine manufacturer, creates a new level of clarity in production for factory operators, as a software solution for data acquisition. Two different perspectives in mechanical engineering, which are united by one goal: Creating transparency and efficiency and ensuring quality with quantifiable data. With the introduction of machine monitoring, mechanical engineering is taking an important first step towards Industry 4.0 and making effective use of digitalization as it continues to advance. 

What does this actually mean for factory operators?

Manufacturing companies with distributed operations and sometimes manual production steps are faced with the challenge that existing machines rarely have data interfaces. In addition, the machines have a high diversity in terms of technology and age, so many different interface protocols need to be supported. Often, there are special solutions that have evolved over time and that are difficult to adapt, which do not make machine integration any easier. In order to be able to digitalize individual machines and manual workstations, a system of worker terminals for recording machine states, sources of error and NIO and IO parts has become standard. 

Machine and data integration — but how?

What is carried out by the operator at manual workstations, is handled by a controller or integration solution at complex or modular systems. In this case, common industrial interfaces such as OPC UA, MQTT or TCP/IP are used. Here too, a flexible configuration of the machine interface for rapid commissioning is the goal, as is the easy connection of existing systems with IT systems.

Handshake between PLC, host computer and MES 

From field level (signals and actuators) to control unit level (PLC, automation) and process control level (SCADA, line controller) to operations control level (MES, monitoring) and company management level (ERP) — interfaces are everywhere and the challenges of seamless integration are correspondingly great. The requirement for more configuration instead of complex programming is therefore clear and can be intelligently solved by suitable frameworks such as Node-RED and FabEagle®Connect. 

Machine monitoring in real time — how mechanical engineering benefits

Ultimately, companies want to increase process transparency in production, by determining production volumes and downtimes, unknown factors and the causes of rejects and bottlenecks, as well as hard-to-detect dependencies. 

A single production site primarily benefits from the introduction of machine monitoring by the factory operator:

• Reduced downtimes of expensive machines
• Production figures that can be evaluated and compared
• Motivated employees thanks to fact-based feedback on production status
• Site-wide transparency

There are also many positive effects for machine manufacturers who want to quickly get a comprehensive overview of their customers’ machines and systems distributed all over the globe at any time, and can sell machine monitoring as a service for their machine:

• Verifiable data-based acceptance criteria during commissioning
• Reduced service and support cases thanks to predictive maintenance 
• Higher optimization potential through feedback on machine utilization 
• Performance values that can be evaluated and compared 
• Cross-type transparency

Even though the benefits of machine monitoring are apparent, from a production and operational perspective, production managers, process engineers and team leaders are right to ask themselves what machine monitoring can and must achieve before choosing an appropriate system. From the manufacturer's perspective, service managers, product managers and digitalization managers are also faced with the same challenge. The advantages can be categorized into three core elements

1. Visualize

Visualizing machine data means making correct state changes (productive planned, downtime unplanned, downtime engineering) and highlighting alarms over time according to frequency or duration and to classify whether these are notifications, warnings, or critical errors. Tracking these process values over time or being able to classify production output into good or bad parts also requires uniform visualization across all systems. 

Displayed as clear graphs and compressed and dynamic dashboards, information can be retrieved and processed in a well-presented style.

2. Analyze

The analysis of machine data goes one level deeper. Here, machine monitoring helps to process the data to be able to better identify correlations. For example, the correlation of status and alarms or throughputs as well as the correlation of specific process values. The correlation of process values is so detailed that relevant parameters can be created as predefined profiles. This quickly makes the most important parameters available to the entire team as well as to the end customer with just one click. Features such as zooming, displaying and hiding values or streaming them live alongside the plots make it easier to get to grips with the visualized data sets and identify any anomalies that stand out. 

Thanks to the mapping of industry-specific or even customer-specific key figure models, the overall equipment efficiency (OEE), availability and process performance of machines can be determined on a company-specific basis. This increases the comparability of machines and also gives mechanical engineers the opportunity to set established key figures individually for the site or customer. The system adapts to the company, not the other way around. 

Act with forsight

Being able anticipate and act accordingly requires that users have all the necessary information available at the right time to make clear decisions and act efficiently. Machine monitoring can also support this by using important machine data to inform users at an early stage of deviations or anomalies (e.g. limit values exceeded, critical alarms or status changes) via email or push notifications to their mobile devices. 

In addition, machine data in real time can be a helpful trigger for scheduling maintenance and servicing. It is not uncommon for unforeseen developments to occur, such as deviating process values or alarms, actual operating time required, the overrunning of a counter or switching state. When these are anticipated they can be scheduled directly as upcoming maintenance in addition to routine maintenance work.

On-premises or cloud hosting?

There are two schools of thought here. Although the acceptance and distribution of cloud applications (public, private & hybrid) is continuously growing across all industries, on-premises hosted applications continue to have their firm place in the IT infrastructure. There are also different requirements between factory operators and machine manufacturers when it comes to hosting machine monitoring. The differences mostly relate to data protection, costs, maintenance, scalability, time-to-market and mobile availability:

Where factory operators prefer on-premises hosting, machine manufacturers prefer cloud hosting, and either FabEagle®Monitoring or EquipmentCloud® are used accordingly. Both product variants share the same database structure, but the hosting and range of functions vary. In addition, customer-specific adaptations and the use of individual functions in FabEagle®Monitoring are an integral part of the product, whereas EquipmentCloud® stands up as an agile ready-to-use solution.

Regardless of which solution companies in the mechanical engineering sector choose, one thing is clear: Machine monitoring increases productivity through transparency in production, equally benefitting system manufacturers and operators. This paves the way to predictive maintenance and machine learning for both manufacturers and operators — two areas in which there is still a lot of potential. Yet another reason for both sides to meet on common ground and strive for a secure but open exchange of machine data.