Process Data Analysis

MagicCorr:  new toolbox for fast historic data analyses

In order to provide a handy (historic) Data-Analysis solution to the daily hectic of operating a power plant, the former PTM team of KEMA – currently CerTa Veritas – developed the MagicCorr® ToolBox. This MagicCorr®-ToolBox is a fast and user friendly ‘App-based’ data-consulting package, designed and innovated in close cooperation with clients in the energy market to handle large data sets (such as the historical process measurements and control data from control systems (e.g. DCS) of power plants, archived in an – often multi-year – historical data base).

Principle diagram of historic process data analysis with MagicCorr, that enables fast analysis of large datasets (time series from DCS and/or process information systems) using an intuitive graphical user interface for tedious data operations (such as filtering out disturbing, irrelevant start-up data) in seconds, that would take hours when carried out in e.g. Excel. This makes complex analyses possible (and affordable), with clear graphical overviews (with color as third dimension) and quick identification of numerical correlations between process variables, which otherwise cannot be made or can only be produced at very high costs. The correlations can amongst others be used to predict certain performance variables using one or more other (independent) variables as input values in the polynomial mathematical expressions produced by MagicCorr.


New toolbox MagicCorr for fast data analyses

Press Play button for MagicCorr overview


Energy companies are constantly looking for ways to make their processes more efficient and reliable in order to reduce costs and environmental impact. A powerful step is to gain insight and understanding, to find out what is really going on in the sometimes complicated processes of a power plant. To do so, a process engineer of a power plant needs to keep track of the historical data for thousands of para­meters. A complex and time consuming task. However, this effort can lead to significant savings, process and efficiency improvements.

MagicCorr® ToolBox

In the past 10 years, the former PTM team of KEMA (nowadays CerTa Veritas) has developed a powerful datamining solution called MagicCorr®. In a blink of an eye, MagicCorr® is able to visualize all the possible correlations and trends among process parameters on screen in a fast, clear and logical way. With feedback from process engineers, PTM designed the MagicCorr® ToolBox. This new ‘App-based’ ToolBox version of MagicCorr runs on MS Windows based computers and are fully adapted to process questions in the daily practice of operating a power plant. All ‘Apps’ are user friendly and optimized for automated and fast data-visualization as desired.


MagicCorr load GT vs Tamb_color_Pamb
MagicCorr analysis of gas turbine power output data in relation to ambient conditions. Upper curve; histogram of ambient pressure. Lower curve: Y-axis: power output, X-axis: ambient temperature, ‘Z-axis’ (color): ambient pressure; for each (middle) ambient pressure from histogram a curve is fitted through the correspondingly colored points (click to enlarge)

What clients say

“The new magicCorr toolbox is a very user friendly and unique in its fast and powerfull data visualization. This makes it an ideal tool for trouble shooting and performance analysis in modern power plants.” – Ellen Tuinman, Process Engineer, E.on-benelux

“This accessible tool supports us very well in the different steps of analysing the huge amount of data available in our plants” – Jasper Kusters, process engineer, Essent Energy B.V.

About MagicCorr®

Besides all benefits, MagicCorr® is an expert tool, which needs training, and experience. Unfortunately this is less applicable in the daily hectic of operating a power plant, since there is great demand for fast, easy, intuitive tools which can be applied without any training or manual. CerTa Veritas has gained much experiency in analysing process data and can support power plant owners with analysing their data.

Effective and fast toolbox

Sick of all effort and endless waiting before your data can be viewed? MagicCorr visualises process data in an easy way and it is way faster than conventional methodes.

Upper graph: histogram of IGV position. Lower graph: x = ambient temperature, y = power output, z = IGV position (indicated by different colors)

Efficient root cause analysis tool

By combining speed and user friendly ‘Apps’, MagicCorr opens up the pathway to high efficient root cause analysis. As an extra benefit, users report to discover new and unexpected causes for old problems.

Increase plant performance (95% savings on preheating gas)

MagicCorr process analysis point out how lost energy could be regained by applying unused pump capacity in combination with two heatexchangers. The result of this improvement are better cooling water conditions and 95% savings on preheating gas.

What is the purpose of maintenance? –> Minimise interventions costs

MagicCorr analysis of process behaviour before and after interventions (overhaul, repair work, process improvement) shows the value of it. Insight- by use of MagicCorr- helps to minimise costs significantly.

Analysis of effect on gas turbine performance of IGV optimization (pre and post).

Key Benefits MagicCorr

  • Fast, graphical analysis of large data sets (time series)
  • Quasi 3D-plots by using color as 3rd dimension
  • Easy graphical selection of relevant data from time series (e.g. only full load cases by graphically filtering out part load data)
  • Easy finding of correlations between process variables and fitting regression curves through related data
  • Easy graphical removal of outliers

MagicCorr Visual datafiltering

histogram of GT-efficiency

‘Polar XY-plot’

In MagicCorr a number of analysis apps are available (XYZ-graph, Split-graph, Matrix-graph, Multi Y-graph and Trend-graph). The presentation of the data in these apps is in a common linear XY-format, but for some processes and measurements the presentation in polar format gives direct insight in the relation between the individual variables. This benefits analysis of data in the situation that for one type of variable more than one sensor is available (and in particular if the sensors are mounted on a (imaginary) circle), for example exhaust gas temperature sensors of a gas turbine. The new polar plot enables easy analysis of GT combustion uniformity over time and provides insight in GT performance/tuning. In the MagicCorr picture above, historic data of 24 exhaust thermocouples of a gas turbine are plotted for an operating range from 80 MW to 290 MW in a time window of approximately one day (standstill time is already filtered out). The connection between the individual points (light blue line) are made on a time basis and at the given time (easily selected by a slider bar above the variable list). The gas turbine is in a ‘ramp-up’ state at a load of approximately 120 MW. The green circle is the averaged exhaust temperature for all thermo couples and is clearly noticeable that the temperature distribution is not uniform which indicates an unevenness in combustion of the GT combustors. The same can be concluded for full load (the purple points in the graph. We put the (selectable reference) Y-value (blue circle) on the lowest exhaust temperature during full load and it is directly clear that almost half of the thermocouples show an, at least, 20°C higher temperature which could indicate that the fuel distribution to the various combustors is not optimal.

Also, at the bottom a linear XY-graph is included in the polar plot which enables the user to include a fourth variable for analyzing or view one variable in the polar plot in more detail in dependency of this fourth variable. The time graph below the polar plot shows a time plot of one of the variables (in this example the GT exhaust temperature 1). In principle instead of time another variable can be choosen at the horizontal axis. Then the polar coordinate in the various pie wedges of the polar plot will be no longer time but the selected variable.

The polar plot shows in one graph an overview of the course of the exhaust gas temperatures over (e.g.) a full day and enables to focus on interesting or ‘suspicious’ values on a particular time by easy selection of that time using a slider bar. This graphical presentation tool will help power plant operators and GT service providers in easy detection of non-uniform GT combustion, and can be the basis for condition dependent maintenance, enabling (requests for) adjustment of the combustion to optimize GT performance and GT expander lifetime.

MagicCorr_plant_efficiency_before and after major overhaul
MagicCorr Analysis of efficiency data before (purple) and after (green) major overhaul (click to enlarge)

Contact for more information or quotation: