•  

     

    Engineering Grade Analytics

    A Predictive Failure Platform with Engineering Diagnostics

  • A Platform Built for Engineers

    Built by Engineers for Flexibility, Scalability and Price-Performance

    The DecisionIQ platform was built by our industry savvy team of mechanical, electrical and industrial engineers with a world class team of data scientists. They built a platform for engineers to solve problems that limit industrial performance. The team transformed the proven engineering methods that are foundational to today’s industrial operations into a state-of-the-art machine learning, problem solving platform. This approach produces results that are distinctively superior in accuracy of prediction, quality of diagnostics and quick adoption by technical operating teams. We keep it simple for the engineers -- we hide the data science.

  • Robust Connectivity

    Because Nothing is Standard with Industrial Applications.

    Any Process in Your Operation

    Any problem in any operational process can be tracked if it shows in your underlying data. We are expert and production and maintenance process.

    Any Data, Any Sensor

    We can join multiple wired sensor types, wireless IoT sensors, any condition data, PLC data and process data from any API and historian type.

    Rapid Results

    We typically can have you deployed within 30 days of of receipt of data, meaningful operational results within 30 more days.

    Low Sample Rates

    Our engineering analytics enables us to build data models with low data sample rates, ideal for a light footprint on IT infrastructure. No edge computing required.

    Automation at Scale

    Our engineering analytics enable your technical workforce to solve more problems across each site. They also automatically benefit from learning at other sites.

    More Than Failure Anomalies

    There are many types of failures and other anomalies. We find them all, including operational startup and shut down, data corruption, streaming corruption.

  • Powerful Diagnostics Dashboard

    For each alert, engineering diagnostics are provided that are useful for root cause analysis. This helps engineers determine the type of failure underway and how to prepare. In this case, time-series supported vibration data from a 3rd party wireless sensor by a waterfall presentation of waveform frequencies with failure modes.

    How We Enable Root Cause Analysis

    Our platform consolidates much of the data required for Root Cause Analysis (RCA) and Six Sigma Continuous Improvement Process (CIP). We assist with process measurement and cause of failures. We typically connect to multiple data sets from production process to condition monitoring from maintenance. We include work orders and spare parts when relevant. Predicted failures show all the offending variables from these data sets. We show feature extracted data as shown in the diagnostics panel above. All this works to give engineering much, if not all, the data they need to perform a structured CIP or RCA by an experienced engineer. But we automate much of the process so an engineer can make a final determination.

    A Powerful, Easy to Use Platform

    With our crawl, walk, run strategy, the DecisionIQ platform can be deployed quickly and supplying your operating team active performance alerts. It can rapidly scale across the enterprise as value is demonstrated. DecisionIQ does the deployment work including connection to your data, curating the data and building the machine learning models. Our customers do not need any data science skills. DecisionIQ has deep knowledge of industrial process, so your vital operational staff is not bogged down setting up a platform.

    Cloud-AI Data Pipeline

    We install and manage the platform for you. We do this to minimize the impact we have on your existing operations and IT workforce. Our platform hosted in the Amazon Web Services Cloud (AWS). Our AI model building pipeline (shown below) was built to specifically to process disparate silos of time series data in real time. For each component or process, there are over one million models evaluated with 300+ separate optimizations that generate 100’s of generations and 1000’s of total children, tests & iterations to produce a final ensemble optimized for very high detectability and very low false positives. Multi-plant tested, deployed and in operation.

    • No Data Science Skills Required
    • Automated Model Building
    • Automated Learning and Re-Learning
    • Support for Any Data, Mixed Data Types
    • Supports Asynchronous Data Feeds
    • Optimized for Time Series Data
    • No Proprietary Sensors Required
    • Engineering Based Feature Extraction
    • Engineering Data Set Curation and Unification
    • Diagnostics Shown with Engineering Data
  • Our Partnership with Amazon Web Services

    We've partnered with Amazon Web Services (AWS). To be an AWS partner means we have both demonstrated success with customers and our platform has met a series of stringent technical requirements. The AWS infrastructure sets a powerful, scalable foundation for DecisionIQ to enable customer digital transformation in production and maintenance operations.

     

    Our customers can use data from virtually any source, sensor and data type in their operation and upload to their data lake. AWS has deployed a powerful suite of software tools that have enabled DecisionIQ to build and operate an analytics platform with the power to solve complex industrial operating problems. It enables our customers to rapidly scale in crawl, walk, run strategy from a single point use case at a single site to many use cases across the site and even to a hundred sites or more across the enterprise.