Erkan Teskancan
Corporate
- Thread Author
- #1
Predictive Maintenance: Is Observation Enough?
The primary goal of predictive maintenance is to prevent unnecessary downtime by detecting machine failures in advance. While many advanced monitoring technologies such as vibration, infrared, and ultrasonic sensors are available, it is often observed that alerts from these sensors do not translate into work orders.
According to Deloitte's 2023 smart factory research, 72% of manufacturers use condition monitoring systems, but only 26% have automated the integration of alerts with CMMS (Computerized Maintenance Management System). The rest operate with manual methods such as email, voicemail, or notes.
The Gap from Observation to Action
Condition monitoring systems can tell you that a bearing's vibration value has exceeded 7.1 mm/s or that a compressor's discharge temperature has been increasing for 72 hours. However, they do not provide information on:
- Who to notify
- Whether spare parts are in stock
- How to schedule maintenance within production windows
- How to record the maintenance activity for compliance reporting
While monitoring is diagnostic, maintenance is an operational activity. These two activities are performed in different software systems, by different departments, and on different timelines. When the transfer of responsibility from one activity to another depends on personal initiative, there is a risk that the correct actions will not be taken. This situation can lead to the sensor detecting the fault on Tuesday, but maintenance being postponed until the following Wednesday, causing the fault to occur and production to stop in the interim.
How CMMS Closes the Gap with Automated Work Order Creation?
The core function that CMMS adds to predictive maintenance is workflow automation triggered by real-time data. Today's CMMS software can be integrated via API with IoT gateways, SCADA systems, or condition monitoring software.
Threshold-Based Automated Triggers
When sensor data exceeds a predefined threshold (e.g., pump motor vibration exceeding 6.5 mm/s), the CMMS automatically:
- Creates a work order including asset ID, problem description, and sensor data.
- Assigns the work order to the correct technician based on their skill, shift, and location.
- Checks the inventory status of required spare parts and alerts if there are shortages.
- Determines the priority level based on the severity of the threshold violation.
- Alerts the technician via mobile notifications, SMS, or messaging services.
This eliminates situations like emails, phone calls, or "I thought someone else was handling it."
Multi-Layered Escalation Logic
A single threshold may not always be sufficient. The ideal approach is to use an escalation method:
- Warning threshold (e.g., 5.0 mm/s vibration): Creates an inspection job to be completed within 7 days.
- Alarm threshold (e.g., 6.5 mm/s vibration): Creates an urgent work order to be scheduled immediately.
- Critical threshold (e.g., 8.0 mm/s vibration): Issues an emergency work order and alerts the maintenance manager about potential equipment isolation.
This escalation method helps prevent "alert fatigue," the biggest enemy of predictive maintenance. When everything is considered critical, nothing is critical.
A Practical Framework for Integration
Connecting condition monitoring and CMMS requires a detailed plan that considers four levels:
- Asset Hierarchy Alignment: Ensure that asset names and hierarchies are consistent across both systems.
- Threshold Definition by Criticality: Automated alerts should be set for critical assets that could disrupt the business process, while regular inspections are set for others.
- Data Flow Architecture: Decide how sensor data will reach the CMMS (direct API integration, middleware, or manual upload).
- Feedback Loop for Continuous Improvement: The results reported by technicians after work orders should be integrated into the system, and issues such as alert accuracy and threshold appropriateness should be evaluated to optimize the system.
Voice-Activated Work Orders: Solving the Last-Mile Adaptation Problem
Despite perfect system integration, there is still a human bottleneck: the technician in the field. Maintenance technicians often work with their hands, and using a tablet can be time-consuming. Voice-activated work order creation allows technicians to create work orders, update statuses, and record findings hands-free. This enables faster data capture, higher work order completion rates, and richer maintenance histories, ensuring that predictive models become more accurate over time.
Lessons from Scaling Across Industries and Sites
Organizations that have successfully closed the observation-to-action gap share some common approaches:
- Start with a single critical asset class: Don't try to connect every sensor to every workflow from day one. Start with the asset class that causes the most unplanned downtime (often rotating equipment like pumps, motors, or compressors). Prove the workflow, refine the thresholds, then expand.
- Standardize across sites: In multi-site operations, standardizing the integration of CMMS and condition monitoring systems ensures consistency and efficiency.


















