Predictive Analytics — Bringing Data Together So You Can Take It Apart
In Project X, we go behind the scenes on our current projects to give you real examples of our solutions.
This time it’s about tools powered by predictive analytics that let you break data out of silos and enable everyone on your team to become a data scientist.
GrayMatter’s goal is to help you view data differently with interactive visualizations, intuitive dashboards, custom alerts and other tools. Compare how temperature and humidity impact metal forging. View voltage versus energy consumption in a pulp dryer. Predict how ingredients will impact the quality of food. Or see vibration readings in relation to flow rates at a pumping station.
Predictive analytics super-charge the expertise that’s already in the boardroom, or in the warehouse or at the well pad. It can answer a question no one thought to ask, or offer an early alert about a pressure reading that could become a ruptured tank.
It’s about finding warning signs, plotting trends and using a virtual space to dig into real-world problems. Simply, we bring the data together, so you can take it apart.
OIL & GAS
FOOD & BEV
GrayMatter is collaborating with an oil and gas customer to analyze how much work its equipment is doing in the field, instead of monitoring only whether equipment is running.
By predicting equipment failures and optimizing how diesel engine pumps are deployed and used, our solution generated an annual savings of nearly $1 million. In addition, GrayMatter helps the customer generate multi-million-dollar savings by identifying and eliminating instances when operators taxed equipment beyond recommended limits, causing failures and shortened equipment lifespans.
In another predictive analytics solution, a GrayMatter oil and gas customer discovered it is spending extra money on equipment that’s not meeting expectations.
GrayMatter’s solution uses company data to reduce high variability in how often key components fail, which ranged from a few hundred hours to a few thousand hours before equipment failure. We also made predictions about equipment performance based on how workers use it in the field, generating significant six-figure savings.
GrayMatter worked with a third oil and gas company to predict shut-ins on some wells two hours in advance.
We also documented that company logs are not capturing all downtime events. On older wells, GrayMatter gave the company confidence that investing in a solution to increase data resolution (the rate data is collected from wells) would enhance the ability to predict shut-ins.
Unplanned shut-ins can cost roughly $150,000 per incident, depending on whether it affects a single well or an entire pad of well sites.
Monitored all process variables from the well
Machine Learning & AI
Computer models predict costly downtime events (well shut-ins)
Utilization of equipment; predicting when equipment is used beyond recommendations and run to failure
Eliminates $150,000 in lost productivity each time a well shut-in is averted
Saves nearly $1 million a year via equipment failure prediction and performance optimization
Empowers field workers to work alongside data center analysts & improves on previous industry standard for data collection
Extends operational lifespan of components worth hundreds of thousands of dollars
Previously, the food and beverage company relied largely on sampling, which added up to thousands of pounds of product “lost” to sample testing per year at a cost of more than $1 million at a single facility.
The company is reducing product sampling by roughly 75 percent and might be able to eliminate it completely – creating a significant savings that it plans to replicate in multiple facilities. When one of the quality measures begins to drift out of specifications, the GrayMatter predictive analytics tools warns employees an hour or more in advance so they can correct the process and avoid a production stoppage or lapse in quality.
Our solution breaks down siloed data about production conditions, process settings, cooking time, temperature, drying time and other factors to allow the customer to predict product quality and to meet quality goals the first time, eliminating the creation of sub-par material that must be scrapped.
Data plots over custom time periods to help spot trends
System learns from months or years of data to flag problems and provides at least an hour of warning to allow workers to correct the issue
Machine Learning & AI
Product quality models predict performance outcomes, reducing reliance on expensive, traditional sampling methods
Saves more than 100,000 pounds of product previously used for testing
Creates a replicable model for additional facilities to unlock multi-million-dollar savings
Accrues a savings of $1 million
At one utility GrayMatter works with, predictive tools continuously ingest third-party data from weather forecasters.
As well as current system conditions and surrounding river conditions to prepare for challenging conditions.
Alerts allow employees in the field to quickly address emerging storm water overflow conditions that can impact compliance with government regulations.
Flexible, scalable system adapts to huge operating area and thousands of established systems from a variety of industrial technology providers
SCADA operators can receive alerts remotely or at a hub
Allows for rapid troubleshooting and system adaptations to redirect water to underutilized areas; workers in the field can receive updates and make adjustments via tablets
GrayMatter’s smart, digital utility solution prevented 1.4 million gallons from overflowing
Preventing this overflow decreased cost 23-40 times lower than capital-intensive infrastructure
Increases compliance with government regulations and avoided potential fines
Creates quick-response system to changing weather events