Capturing, recording equipment inspection data for FMEA

Artificial intelligence can assist with data capture for failure modes and effects analysis and reliability programs. ozcan yalaz / DigitalVision Vectors / Getty Images

Reliable equipment and processes are key to sustainable productivity and success in a metal fabricating shop. Programs like reliability-centered maintenance (RCM), which analyzes equipment breakdowns to determine the best maintenance methods and schedules for each machine, help keep shops on track with minimal downtime.

Smart manufacturing can help elevate RCM and other reliability programs even further, offering useful life cycle curves, predictive maintenance alerts, suggestions for scheduled maintenance, and suggestions for maintenance of nearby components when you are working in a difficult-to-access area.

RCM uses your data and experience to anticipate and plan required maintenance events. Artificial intelligence (AI) can work as your on-site assistant to monitor conditions and predict machine failures. AI also can help you maintain a reliability program that reduces machine failure, eliminates stoppages, and supports safer operations.

The traditional method to diagnose and anticipate failures is failure modes and effects analysis (FMEA). FMEA identifies potential functional failures, modes of failure, and the effects and severity of failures. From this, maintenance managers can prioritize maintenance activities to prevent the worst and most probable failures.

Introducing AI allows your staff to employ historic data to supplement their experience and intuition. FMEA is dependent on both, using supplier information, employee knowledge, and historic data to document the causes, consequences, remedies, and prevention of failure events.

Many manufacturing reliability managers agree that diagnosis of machine maintenance issues is a partially intuitive skill on the part of maintenance employees and machine operators. But while some employees have the intuition for it, many do not. With accurate and precise monitoring and data collection, AI can offer tremendous assistance to predict and prevent production and safety failures.

The key to success, whether using AI or not, is that component status and experiences be captured. It’s unrealistic to monitor every bearing, interface, and motor. That’s why it is critical that maintenance and operations employees understand the need to record degradation, maintenance, and failure events accurately. This data is crucial to support the integrity of your AI/smart manufacturing application.

Data Collection Options

Figure 1 show an example of FMEA on a robotic subsystem. Even though this system is simple, it includes quite a few data points. The first row, Fails to Supply Power, includes several potential issues that would require monitoring: heat, material corrosion, component vibration, impact warnings, and fuses. Multiplied across a complex system, comprehensive component monitoring can become expensive.

Periodic visual inspection can fill the role of expensive monitoring devices in these situations. Visual inspections can easily include several potential sources of failure in a single review. The inspector must thoroughly document the results of his visit, and it must be captured to maintain the integrity of your smart manufacturing analysis.

Keep in mind, however, that as you move into a more data-focused business model, capturing accurate and precise information becomes a mission-critical activity. For the following critical component measures, you’ll still need to use devices and components with digitally enabled monitors that store their readings:

Figure 1. Even a simple robotic subsystem includes many data points for FMEA.

• Component dimension
• Input material and component properties
• Equipment operating state
• Hole placements and dimensions on flanged surfaces
• Energy consumption trends

Measuring devices must be clean and calibrated to ensure accurate readings. Each poor reading becomes a part of your data history and will influence future decisions. And a single dirty or poorly calibrated monitoring device can lead to stoppages and serious safety issues.

Capturing Inspection Findings

Capturing equipment maintenance procedures is a manual task that is often neglected in the rush to move to the next task. Inspection observations, observed failure modes, and maintenance activities are important factors in defining equipment life cycle curves and predictive condition-based maintenance.

When performing equipment maintenance, it’s important to capture and record your inspection findings, component problems you’ve observed, repairs completed, the time since the last repair, and the cycle count for the equipment.

These manual tasks can be supported with apps that direct you in recording maintenance observations and activities. Voice-activated AI also can offer promise in capturing the important events in a machine’s maintenance life cycle.

As data becomes even more critical in decision-making, disciplined processes will become necessary. Each poorly recorded event, each imprecise or inaccurate entry, will be baked into your data history, leading to incorrect conclusions from your decision support system.