Need of analytics for Business
Business have many questions in mind regarding how they can leverage their existing infrastructure to come up with advanced analytical solutions to solve there some of problem statements like identifying the key problem area in their shop floor operations where business can focus on to improve the operational efficiency, to have an authentic analysed data which can help them to take proactive decisions which can eventually help them to reduce downtime and operating cost.Business are first looking at their existing infrastructure to come up with such analytical solutions or are in search of tools which can help them to come up with advanced analytical solutions.
Type of analytics currently in market
Currently in market there are wide variety of analytical solutions available which business are looking at to solve
- Descriptive Analytics 2. Condition based Monitoring 3. Predictive analytics 4. Situational awareness 5. Machine l
- earning 6. Pattern matching 7. Regression and multivariate analysis
This all type of analysis gives insight into operational activities providing business authentic and variety of data co-relation. To come up with more accurate results acquires wide variety of data which can be time series data or even unstructured data like shown in figure below
Accuracy of Analytics will depend on how effectively we will be able to utilize all this wide variety of data.
How we leveraged OSISoft Analytics for solving the business problems
Simple Calculation based analytics
Customer had existing infrastructure of OSISoft and Siemens DCS system from which Alarm and Event data was coming and they were looking at solutions to build alarm management solutions based on ISA 18.2 standard leveraging advanced analytical capabilities of OSISoft Asset Framework analytics.
Above example shows how we leveraged the OSISoft AF functions like mathematical ,logical, statistical to come up with simple and advanced analytics where Alarm string tag coming from PCS7 was broken into different attributes which further were utilized in AF Event frame.This gave analysed data results which are nothing but KPI defined by ISA 18.2 standard such no of alarms per hour area/unit and plant wise,No of chattering alarms, severity wise breakup of alarms, top 10 equipment alarms. All this kind of KPI's were derived using OSISoft AF analytics .
Visual analytics (Dashboard and Reporting)
In visual analytics solutions dashboard and reporting was created using .net web pages by using AF SDK to retrieve the AF analysed data and show KPI on dashboards like shown below
Visual analytics (Area/Plant/Asset comparison)
Another method of visual analytics is bench-marking the performance of Equipment/Area/Plant and do comparison analysis with different Equipment/Area/Plant like we did for comparing the no of alarms generated depending on selection and doing the comparison.
Visual analytics (Forecasting)
There is another type which is Visuals analytics for forecasting to evaluate the accuracy of predictive model under various conditions.
Predictive Analytics (Advanced Pattern Recognition) –
Also for along with help of other analytical tools such as R, Matlab, and Azure Machine learning can be used to build and train the model. The predicted data can be pushed back to Future data archive in PI Historian. Monitor multiple asset variables for anomaly detection like in sample below for Air compressor based on co-relation of compressor operational parameters failure pattern was derived which was compared with live streaming data to arrive at probability of failure of Air compressor.
Condition Based Monitoring –
In the condition based monitoring example , asset like Pump was taken by retrieving it's maintenance information from Maintenance management system and based on number of hours operated after installation/maintenance necessary notification to concerned deportment were sent to take action on same. For Example if Pump has operated more than specified limits after its maintenance notifications are triggered or if pump if there large deviation in operational parameters from specified limits of Pump then notification are triggered.
We think with evolving architecture of OSISoft can give flexibility to collect data from other type of data sources to arrive at more accurate results.
With increasing need of analytics in industries business are looking out at tools with analytical capabilities and also to maximize their existing infrastructure utilization for analytics solutions without any need of further investment. OSISoft which is currently the leading data historian widely used in industries has strong analytical capabilities as we experienced in our project.This also provides flexibility to business to leverage existing OSISoft infrastructure and use this as backbone to come up with Analytical solutions and realize the benefits out of this.