Exxaro Energy Mangement - IS³ | AVEVA Select Partner

Industry: Mining

People to make the difference between energy usage and waste at Exxaro’s Grootegeluk Mine


  • To continually reduce power consumption and increase operational efficiency with the aid of KPIs recorded as historical plant data.

Solutions and Products

  • System Platform
  • InTouch
  • Historian
  • Skelta BPM
  • Flow Software


  • Representation of the plant through two different model viewpoints: The Control Model and the Information Model


  • Predicted savings of more than R5 million over one year based on 5% reduction of energy usage above set targets (based on 2012 data)
  • Improved decision making as implemented through Skelta BPM
  • Increased operator cooperation

Lephalale, Limpopo Province, South Africa − Today’s process control systems gather huge amounts of valuable data from the processes they are monitoring. For the humans involved in the control loop, however, making sense of it all isn’t always easy. But by combining the strengths of various advanced control software solutions, they can be advised on the best possible decision to make under current circumstances.


Exxaro’s massive coal mining operations at Grootegeluk supply the existing 4000MW Matimba power station and the future 4800MW Medupi power station.  Ironically, one of Exxaro’s immediate goals is to reduce production cost by minimising electricity usage from these same power stations. Since a large part of that energy reduction is the result of human decisions, workflow management software along with a versatile and novel historical data retrieval and analysis solution were used to prompt operators as to what best action to take within 5 minutes of a sub-optimal event. The only way they can do that is to emulate the proven effectiveness of control loops – get the information they need and know what to do with it within the context of a human-based process control mechanism.

Humans have always been part of production control loops but rarely in formal ways. “What to do next” has usually been left to the wisdom and experience of whoever’s in charge. But wisdom is an increasingly rare commodity with today’s skills shortage. What action to take relies on the analysis of the correct information delivered at the right time followed by a formal implementation procedure involving all the relevant personnel – because they’re the ones who will contribute the most to energy savings.

Studies have shown that technological changes account for only 10% of energy savings whereas behavioural changes in human stakeholders can increase this to a possible 30%. Since plant operational personnel control a large portion of the power used at Grootegeluk, they have the potential to contribute the most towards achieving energy-saving targets.

So Exxaro decided to empower operators to make the right moves by providing them with the real-time and historical information they need (using Flow Software) collated from the existing real-time Wonderware Historian and giving them a framework of action and responsibilities (using Wonderware Skelta BPM) so that “what to do next” became a far more predictable and optimised way of achieving the company’s energy-saving goals. In other words, a system that could provide operators the necessary “wisdom” that would otherwise have had to be developed over years of on-the-job practice.

“Achieving this ‘wisdom’ is more complicated than it sounds,” says CJ Barnard, system engineer at system integrator Advansys. “Energy consumption depends on a multitude of variables so that comparisons between shifts or different production configurations are very difficult to address unless apples are compared with apples.

The ‘garbage in/garbage out’ principle applies and must be avoided above all else. Relevant recommendations to improve poor performance can only be generated if smart ways to group, sift and sort information have been implemented.”

The main goal of the project was to reduce input cost by reducing energy usage while maintaining or improving production throughput. “So the aim of the project was not necessarily to increase the production throughput but rather to decrease the associated cost through more effective equipment use,” says Barnard.


The project was started in September 2014 and completed in June 2015. To meet Exxaro’s goal, information model objects had to be developed in such a way that they could seamlessly integrate with the existing control model objects. A well-designed containment structure was developed to logically represent plant information through the ArchestrA, Flow Software and Skelta BPM software layers.

“Wonderware’s System Platform, based on ArchestrA technology, provided the hub for the merging of historical plant data, its analysis and operator interaction. The project proved that existing plant data could be leveraged to identify inefficiencies and to optimize Grootegeluk’s use of energy.”
– CJ Barnard, System Engineer, Advansys

The plant benefited from an extensive existing installed base of industrial IT assets centred on the Wonderware System Platform. Information already contained and reported in the Wonderware Historian provided the foundation on which decisions about the plant’s optimisation could be made. “With appropriate engineering, the flexibility of the Wonderware System Platform provided the customisation demanded by this project and also allowed the Skelta BPM component to be incorporated,” says Barnard.

Flow Software serves as the data gathering and information-generating engine that queries the historian to determine the best performance levels over a period of 3 months. The results of this analysis are accessed by the appropriate objects in the System Platform and used in the Skelta BPM rule-enforcement procedures. This, in turn, uses the representation of the various process modules at the level of the ten InTouch (SCADA/HMI) control stations to ensure that the correct messages are delivered to the relevant operators. The Flow database also contains the data on which reports are generated and hosted with Microsoft Reporting Services.

The system was developed and tested separately from the actual plant prior to changeover. Testing was conducted in the Cloud (Microsoft Azure Services), providing a unique way in which contributors from both Advansys and Exxaro could design, collaborate, implement and view progress on the project. The solution could be deployed on the live plant as the information model does not control the plant and does not affect the deployment of on-scan production objects.

Apart from engineering done in the System Platform as well as Skelta BPM and Flow Software, a customised trend client was developed in C# as a user control that could be embedded into ArchestrA graphics. In addition Microsoft Reporting Services and custom stored procedures were used to generate the various reports pertaining to the project.

System architecture – the key building blocks and feedback components

Role of Wonderware Skelta BPM

The incorporation of Skelta BPM allowed for the implementation of institutional knowledge as it relates to plant optimisation based on measured past performance. The configurability of this application facilitated the design of templates with which optimisation rules could be implemented and enforced.

Continuous plant improvement opportunities are possible as analysed and processed information is fed back into the system to generate workflows based on the previous best-performance information. This reduces the time the plant spends in poor performance situations and continuously adjusts targets towards an optimum performance level.

The design of workflows allowed the following rules to be implemented consistently:

  • Rule 1: Switch off power-consuming components if production is low.
  • Rule 2: Distribute production load over one or more processing components to reduce effective power consumption.
  • Rule 3: Switch off power-consuming components if silo levels are low

Since operator accountability is paramount to real energy savings, rule enforcement triggers operator prompts as to which operator responses are obligatory. Their responses and IDs are stored in the historian. The rule prompts and operator comments are monitored on the custom-built trend client and an audit trail can be created of the plant’s performance, the activities of operators and the constraints against which optimal production was being driven.

Optimised energy usage takes place according to the rules determined by the company’s engineering management team. These high-level rule definitions are made specific to the operators controlling the plant by adapting to the current production and power consumption scenarios. This two-layered approach provides the feedback on the rules and optimisation parameters that will contribute to the overall performance of the plant.

Role of Flow Software

Flow is a configuration-based calculation and reporting engine that integrates directly into and extends various data sources such as Wonderware Historian, Microsoft SQL, OPCHDA and others. Calculations, manual data entry, limits, reports and forms are configured in a drag-and-drop manner. There are three phases to the Flow information delivery process; collect, calculate and visualise.

On this project, the inability to perform any one of these tasks would have resulted in failure. The single Flow Repository is the source for all reporting. Pre-aggregated and pre-calculated information enhances the performance of reporting and increases business intelligence. For this project, the Flow information was integrated into the Wonderware Skelta BPM application and can also be used in the company’s ERP and other environments.

Role of Wonderware Historian

This project would not have been possible were it not for Historian which supplied Flow Software with the necessary data for analysis and aggregation. This enabled the automatic generation of empirical mathematical models representing the efficiency of a particular process module.

Historical measurements, equivalent to 3 months of data sampled every 5 minutes (or 43200 samples), is used to evaluate the production configuration in each of Exxaro’s eight process modules.


A significant challenge was the representation of the plant through two different model viewpoints; i.e. the control model and the information model. The control model comprises plant representations extending from the System Platform down into the PLC and field layer. The System Platform is the layer in which the control and information models meet and coexist in the galaxy. The information model extends upward to the PIMS layer containing Flow Software and beyond to SAP and other business layers.

“Good judgement comes from experience and experience comes from poor judgement.”


  • At launch, Exxaro predicted savings of more than R5 million over one year based on 5% reduction of energy usage above set targets (based on 2012 data)
  • Improved decision making as implemented through Skelta BPM
  • Increased operator cooperation

About Exxaro Grootgeluk Mine

The mine is one of the most efficient mining operations in the world and operates the world’s largest coal beneficiation complex, where 9,000 tonnes per hour of run-of-mine coal is upgraded in six different plants.

The $1.3 billion Grootegeluk Medupi Expansion Project (GMEP) will increase throughput to 14,000 tonnes per hour, to supply the Medupi power station with an average of 14,6Mtpa of power station coal over the next 40 years.

Exxaro has a strategic target to become a US$20-billion company by 2020 and a global, diversified company with top-quartile returns. To help achieve this it has set targets for operational excellence, aiming to reduce its carbon footprint, improve the safety and empowerment of its people, become an employer of choice, and an agent for change in South Africa.

The Project in a Nutshell

  • Supervision is done on the energy usage of Exxaro’s Grootegeluk coal mining plant as it relates to the production of coal.
  • This is followed by monitoring the response of operators to automated recommendation prompts designed to optimise the performance of the plant in its current configuration.
  • This process is repeated in order to provide a platform for continuous and incremental optimisation of power usage by those who control its use.

Standard Work Processes

  • Consistent execution
  • Institutional knowledge
  • Plant optimisation
  • Execution trail and audits
  • Production configuration

Work Flow Execution

  • Responds to real- time data
  • Triggered on sub-optimal events
  • Leveraging of optimisation opportunities
  • Cost savings

Operational Consistency

  • Best practices enforced as operational rules
  • Operator responses are obligatory on workflow execution

Minute-by-minute Data Aggregation

  • Consistent execution
  • Institutional knowledge
  • Plant optimisation
  • Execution trail and audits
  • Production configuration

Weekly Data Analysis

  • Responds to real- time data
  • Triggered on sub-optimal events
  • Leveraging of optimisation opportunities
  • Cost savings 

Manual Data Entry

  • Best practices enforced as operational rules
  • Operator responses are obligatory on workflow execution

Big-data Source

  • All data
  • All modules

Store Information Model Tags

  • Total production
  • Total power
  • Average efficiency
  • Production configuration

Expose Tags to Flow Software

  • Aggregated measurements
  • Data analysis
  • Target evaluation