What Big Data Can Reveal About Your Equipment: The Case of an Electric Motor

June 26, 2025

Blog

Few understand the phrase “time is money” better than industrial companies, especially when both time and money are wasted during unexpected equipment downtime. For automakers, just one hour of unplanned downtime caused by sudden equipment failure can cost up to $2.3 million. In fact, equipment failure accounts for 80% of all unplanned production halts.

To streamline operations, many manufacturers are turning to Predictive Maintenance (PdM) — one of the most effective ways to apply Big Data in production. This approach involves continuously collecting and analyzing data on equipment condition to detect faults early, often before they result in any costly failures. One luxury car manufacturer, for instance, managed to reduce unplanned equipment downtime by 25% simply by leveraging data analytics.

In this article, I’ll use the example of an electric motor to show exactly what insights Big Data can reveal about your equipment — and how you can start using it to your advantage.

What the Data Reveals

Depending on their type, IIoT sensors track various physical parameters, such as vibration, temperature, pressure, current, and rotational speed. These measurements are converted into digital signals and transmitted to a cloud platform. 

Below are some examples of deviations in electric motor performance parameters captured by sensors and the potential faults they signal:

• The Appearance of High-Frequency Vibrations Above 2 kHz. Under normal operating conditions, electric motors rarely experience high-frequency vibrations. A sudden spike in these vibrations can indicate a deterioration in bearing lubrication. The vibration sensor, mounted on the bearing housing, detects these anomalies. It records the vibration data along with a timestamp — depending on the task, measurements can be taken up to 32,768 times per second.
• A Gradual Increase in Bearing or Motor Housing Temperature Without a Change in Load. The normal operating temperature for most electric motor components is up to 70-80°C. A rise of 10-15°C could indicate issues such as lubrication problems, initial bearing wear, contamination, or shaft misalignment, causing excessive friction. This anomaly is detected by a temperature sensor, which transmits temperature readings along with timestamps. Typically, measurements are taken once every minute.
• A Gradual or Sudden Increase in Current Consumption. The current should not deviate more than 10% from the motor’s nominal value specified by the manufacturer. If the value consistently exceeds this threshold, it may signal bearing wear or poor lubrication, causing the motor to consume more energy to overcome friction. It could also indicate starting issues, such as a slipping belt. This anomaly is detected by a current sensor, which records the current in amperes along with timestamps, often for phases A, B, and C if the motor is three-phase. Measurements are typically taken once per second or more frequently.
• Vibration or Decrease in Rotational Speed (Revolutions per Second or Minute) Under Stable Load. If the motor’s rotational speed deviates by more than 5-10% from the nominal value, it indicates mechanical resistance in the system. Possible causes include shaft misalignment, bearing wear, slippage in the drive, or even unstable operation of the frequency converter. In some cases, this could be a sign of overload or power loss. This anomaly can be detected by a rotational speed sensor, such as an encoder. The encoder records the number of pulses, typically voltage spikes generated as marks pass over the rotating element. Using this data, the system calculates the rotational frequency. Measurements are typically taken 1–10 times per second.
• Micro-Vibrations in the Motor Housing Under Stable Load. Invisible to the naked eye, micro-movements can be early signs of loose fastenings, bearing wear, shaft misalignment, or damage to the motor housing. Such deviations can be detected by a video monitoring system using Motion Amplification technology. A high-speed camera records the motor’s operation, while a specialized algorithm processes the footage, amplifying micro-movements by hundreds of times, making them visible to the human eye. Based on the processed data, vibration amplitude maps and frequency spectra are created. Measurements are typically taken at a recording frequency of 30 to 500 frames per second. Motion Amplification is one of the latest advancements in Predictive Maintenance, combining traditional vibration diagnostics with specialized software for visualizing vibrations.

As you can see, a single parameter can signal multiple potential issues. The task of a PdM system is to analyze all the indicators collectively. AI and ML algorithms simultaneously analyze several parameters, compare them with historical data on normal and abnormal conditions, identify the likely cause, and propose it as a hypothesis.

Currently, AI cannot guarantee 100% accuracy in diagnosis. For a thorough and final diagnosis, various specialists may be involved, such as vibration analysts, acoustic diagnostic engineers, thermography experts, and others. However, in my forecast, fully automated diagnostics will become a reality in the coming years, as this field is developing rapidly.

How to Start Using Data in Manufacturing

Implementing an effective PdM system that works with Big Data requires careful planning. Here are the key steps you'll need to follow if you're planning to implement PdM on your own:

• Identify critical equipment whose downtime could lead to significant costs.
• Install IIoT sensors that are compatible with your IT infrastructure and enable high-quality data collection.
• Set up network infrastructure to ensure stable real-time signal transmission.
• Configure a database for processing large volumes of time-series data, scalable to accommodate increasing numbers of machines and sensors.
• Choose the right analytics tool that suits your needs, operates in real-time, and has a user-friendly interface.
• Invest in training for engineers and analysts.

Data collection alone doesn’t guarantee success. However, if you approach the implementation of PdM as a strategy and take a gradual approach, starting with critical equipment, using Big Data and advanced AI algorithms will help you shift from reactive to proactive management of production assets. Companies utilizing Big Data report an 8% increase in revenue and a 10% reduction in costs. In 2023, the use of AI tools by the National Oil Company of Abu Dhabi (ADNOC) resulted in an additional $500 million in profit. And this is just the beginning.

---

Illia Smolienko is the Chief Software Officer at Waites, a leading provider of condition monitoring and predictive maintenance solutions for industrial enterprises. He has over a decade of experience in industrial IoT and the implementation of PdM strategies. Under Illia’s leadership, large-scale IIoT-based monitoring projects have been deployed for global companies such as DHL, Nike, and Tesla.