Asset managers often ask how online condition monitoring fits into digital transformation. It can support an asset-management strategy, but sensors are only one part of the system. The practical challenge is collecting usable data and making it available for informed decisions.
Why should we choose online condition monitoring?
Transformer condition depends on interacting electrical, thermal, mechanical and chemical stresses. Moisture in cellulose can accelerate ageing and lower the temperature at which bubbles may form, but there is no universal rule that each one-percentage-point increase halves the life of a complete transformer. CIGRE TB 761 discusses moisture alongside temperature, paper type and other ageing factors. Online monitoring can help track selected indicators between laboratory tests, provided alarms and interpretations reflect the sensor, the asset and the quality of the data.
Rating the digital monitoring market with pros and cons
Online condition monitoring can reveal changes between scheduled tests and help plan investigations. Its value depends on which failure mechanisms the sensors can detect, the quality of their readings and the response to a warning. Those capabilities can support availability and maintenance planning, but installation alone does not establish a reliability or life-extension benefit.
However, lack of proper training, high cost, and relatively shorter lifespan of monitors are practical concerns to assess. Often triggering of false alarms and erroneous data interpretation due to sensor malfunctions or temporary fluctuations are also challenging. An interesting use-case is integrating aged assets with compatible monitors to the supervisory control and data acquisition (SCADA) of transformers.
Is online data collection sufficient?
During one of our Deepbrains conversations, we asked the same question. The current market is booming with various sensors and monitors for capturing condition monitoring data online. Since it may be challenging to explain some or all fault modes, people assume capturing more data is the key. The lifespan of online monitors is often less than the machine and require replacement. This adds more complexity to the problem at hand. An ideal diagnostic platform should have a dashboard for visualization along with an analytical tool to draw insights from it. Monitoring systems can combine sensors, data storage and visualization. A useful display must also make the diagnostic method and its limitations clear; appearance alone does not establish the quality of an interpretation.
It suffices that only relying on online monitors is not sufficient. Additionally, the real challenge is storage, management and analysis of the large volume of data generate by these monitors. It can be more challenging for organizations that lack necessary infrastructure or data analysis capabilities.
What can be the best practices in data management for online monitoring?
- Data collection: Digital sensors generate a vast amount of data regarding on machine’s condition. After identifying the key condition alerting parameters, select and commission devices with accuracy, measurement range and sampling frequency appropriate to the failure mechanisms being monitored.
- Data storage and processing: Choose cloud, on-site or hybrid storage according to access, resilience, data-governance and processing needs. Cost and environmental impact depend on the actual infrastructure, energy supply, utilization and data-transfer requirements; neither architecture is inherently preferable on those measures. Without processing, it can be challenging to effectively utilize this data in its raw form. Suitable analytical tools can perform data cleaning to extract meaningful insights and identify relevant trends or anomalies.
- Access to Actionable Information: Analytics software plays a crucial role in processing and interpreting the collected data. It helps transform raw data into actionable information by applying algorithms and models that can detect patterns, identify potential faults, and provide insights for decision-making. Without such software, the collected data may remain unutilized or lack the necessary analysis to drive proactive maintenance actions.
- Early Fault Detection: An analytics software can detect and diagnose faults by analyzing sensor data and identify early warning signs of developing faults. By overlooking faults, risk of unexpected failures and subsequent damages increases.
- Efficient Maintenance Planning: Analytics software enables the development of predictive maintenance strategies based on the condition of the transformer. It assists in determining the optimal timing for maintenance activities and resource allocation. Without this software, maintenance planning may rely on less effective methods, such as reactive or time-based maintenance, resulting in suboptimal utilization of resources and potentially increased maintenance costs.
- Trend Analysis: Trend analysis is crucial for understanding the long-term performance of a machine. It helps in identifying patterns and deviations from normal behavior and help in identifying anomalies. Without it, trend analysis becomes difficult, hindering the ability to track and interpret changes in the machine’s condition over time.
- Reduced Manual Effort: Without data management, the analysis and interpretation of sensor data may rely heavily on manual efforts. This can be time-consuming, labor-intensive, and prone to human errors. Automated analytics software streamlines the data analysis process, enabling more efficient and accurate interpretation of the sensor data.
- Improved Scalability: Effective data management can help in scaling up the condition monitoring system to accommodate a larger number of machines or additional sensors, that can otherwise challenging. The lack of automated data analysis and management tools may limit the scalability of the monitoring infrastructure, making it difficult to handle the increased data volume and complexity
Try Ronin AI for improved experience on transformer condition monitoring
Analytics software can help organize sensor readings and laboratory results for review. Ronin AI provides a platform for transformer condition data and fleet screening. Supply the required gas measurements, sample dates and units, together with relevant oil-quality and asset information. Check input quality, make missing fields visible and review what changes the result. A dashboard score supports an engineering decision; it does not guarantee a particular accuracy gain or an effective maintenance outcome.
So don’t wait and grab your free trial right away!
