EMI as a provider of Big Data analytics for the Process Industry
Whether you are looking to improve your oil & gas exploration and production efficiency, reduce non-productive time or predict equipment failure before it happens; EMI can put you on the right track with more than 25 years of experience in real-time data, predictive and prescriptive analytics.
Data Acquisition, Ingestion, Warehousing and Retaining
EMI has been helping customers collect and historize real-time data since 1992, with data sources ranging from OPC servers to flat files and relational databases and into historians such as OISSoft PI, Wonderware, Aspen IP21.
Ever since that time, Enterprise firms have collected huge amounts of data, and even though traditional historians can handle the archiving and analysis of this data, the need to integrate with other systems transformed these historians into another data sources. Data sources for what? Data sources for Big data platforms, such as Hadoop HDFS.
In addition to its vast experience in collecting data into real-time historians, EMI has also developed tools and implemented scripts to ingest data from various data sources into big data lakes, starting with Hadoop HDFS, on both premise and cloud and now Cloudera Data Platform. These data sources range from real-time historians such as PI into relational databases, document management systems, unstructured data, media files, drilling and well management systems…
Data Quality is a must
EMI started with automatic data profiling and cleaning through its “Tag Data Validator” (TDV) product. TDV will automatically scan and subscribe to your OSIsoft PI Data (real-time and archived) and mark your data as either good, have quality issues or not available. The quality issues range from simple marks such as below or above range, into complex ones such as noisy or frozen. Newly added features include machine-learning enabled algorithms to eliminate false positives and deliver the most accurate data quality profile. Historical data will be studied and analyzed to produce a distinct profile for each tag, allowing TDV to study the optimal values to consider a tag as frozen, noisy or not available.
Descriptive, Predictive, and Prescriptive Analytics
In the race to Industry 4.0 and IOT, predictive maintenance is crucial in reducing downtimes and preventing catastrophes as problems can occur suddenly and in remote areas where downtimes can be very costly.
Predictive maintenance a technique for preventing/predicting equipment failures by analyzing the patterns in real-time data and comparing them to historical data to identify anomalies and predict downtimes.
Until recently, engineers and operators carried out scheduled maintenance and regularly repaired machine parts to prevent downtime, this is called preventive maintenance. Studies and statistics have shown that half of all preventive maintenance activities could be ineffective, and it consumes unnecessary resources and leads to production losses.
Due to this, EMI is taking the initiative to lead its clients towards the implementation of predictive analysis either by using its state of the art Predict-It™ Software from Engineering and Consultant Group ECG or by exploiting the historical data and AI algorithms to predict failures, lower service costs, maximize uptime, and improve production throughput
EMI didn’t stop here, in addition to predicting the failures, recommendations and actions must be advised in order to avoid or prepare for the upcoming failure. This is known as Prescriptive analysis, because it “prescribes” a set of different possible actions to the users and guides them towards a solution.