Answer: Both use predictive analytics.
In 2012, the Obama campaign notably used analytics to identify which voters could be influenced through campaigns such as calls, knock on the door, flyers, and ads. With less than 15 months to the 2016 presidential election, analytics is once again set to take center stage in the campaign process. While predictive analytics models may not yet be able to forecast 2016 results with certainty, they are playing a transformative role in the campaign process. They are helping candidates make strategic business decisions and identify voters who can be persuaded to switch allegiance.
The energy, resources, and utilities industry uses predictive analytics to proactively identify assets that require maintenance in order to enhance outcomes. Product manufacturers are under continual pressure to minimize unplanned maintenance, achieve higher asset uptime, and meet regulatory compliance mandates. Unplanned outages can prove to be expensive. So, organizations have to stay on top of their assets, and perform a balancing act between reducing costs and offering reliable services.
Traditional monitoring approaches are inadequate
Organizations have traditionally used two approaches for monitoring and maintenance – reactive and preventive. Reactive maintenance comes into play after the occurrence of an event, while preventive maintenance deals with fixing a problem at specific intervals. The problem with this is the interval may be too long or too short. Both approaches are cost intensive and eat into the operating budget of an organization. Also, the basic monitoring tools used in these approaches offer limited scope and isolated feedback. They do not consider the holistic picture or understand the relationships between components, often providing biased judgements and false alerts.
Predictive analytics – a statistical approach to asset maintenance
One of the most prominent applications of predictive analytics is monitoring equipment health to avoid asset failures. Analyzing factors such as frequent load fluctuations can clue in power utilities to potential transformer failure. Similarly, for water utilities, pressure changes in water pipelines could indicate a leakage. Early detection of such issues can be achieved by monitoring for suboptimal performance or deviations. Once identified, proactive actions can be taken to resolve the issues – even before their occurrence.
Asset health management is another area where predictive analytics can add substantial value by providing answers to some critical questions. How much longer is the asset likely to perform at the same quality level? Is the asset underperforming under certain conditions? When should it be replaced? Predictive analytics helps gain insights into such questions with data mining methods such as time series. Nexen, a Canadian oil and gas company, uses predictive analytics to determine when its equipment needs to be replaced. In doing so, Nexen shifted from a traditional approach of replacing its assets at fixed intervals to a more accurate and dynamic model based on asset health monitoring.
Accurate data is the key
Data is at the center of analytical models such as advanced pattern recognition and trend analysis. While leveraging the potential of predictive analytics for asset monitoring, it is equally important to ensure accuracy of the predictions and patterns. Sensors provide insights into parameters such as flows, temperature, and pressure in an industrial facility. This is then combined with historical data and other existing asset management systems to come up with a data pool that can feed into analytical models.
Hybrid models such as analytics network process (ANP) and fuzzy neural network (FNN) can be used to predict failure of parts. The output generated by ANP hybrid model can be used as input for a FNN model. Critical range and limits can be defined for these models based on historical data, allowing maintenance teams to take immediate action when they are breached. Machine learning is making this approach much easier with the ability to self-learn and reset new limits based on the tools’ performance.
Analytics as a business enabler
As a tool that generates futuristic insights for enterprises, predictive analytics is quite powerful, and has the potential to influence strategic outcomes. It is estimated that by 2023, the utilities industry will spend $50 billion on asset management and grid monitoring technology. Given this huge investment, it is important for industry players to focus on maximizing the bottom line. Predictive analytics can deliver this through better inventory management, and by reducing investments in new assets due to better lifecycle management and timely maintenance.
How does your organization leverage analytics to stay ahead of the curve?