Is Machine Learning the Emergency Preparedness Secret Weapon?
I read an excellent article that discusses the potential for machine learning to help electric utilities improve outage restoration metrics. The key takeaway is that when it comes to this type of technology, it’s all about the data!
How Electric Utilities Can Leverage Machine Learning
In its most basic form, machine learning takes data – lots of data – and performs continuous analysis with the objective of identifying the exact combination of variables that will lead to progressively desired results. The type of data that can be incorporated runs the gamut from weather data to equipment lifecycle data to outage data and human resource data.
According to the article, there are multiple ways that utility companies can leverage this type of learning. For example, machine learning can analyze outage root cause data to predict when the grid might be at an increased risk of failure, and can also help predict the path of cascading failures in a variety of situations. Armed with this knowledge, grid operators can take proactive, preventative measures.
Really, the benefits of machine learning are only limited by the degree to which utilities can obtain the underlying data. What will it take to increase the amount of available data and optimize the benefits of machine learning? Utilities will need to focus on three primary things:
- Continuously increase the percentage of customers with smart meters.
- Deploy an increased number of sensors along the grid for monitoring and control purposes.
- Integrate data from off-grid generation sources like solar and wind.
It won’t be easy, but over time electric utilities will have the data infrastructure necessary to facilitate machine learning within their operational processes. And when it happens, we’ll all benefit tremendously.
