The FCT-ACTech Continuous Mineral Analyser has been developed in conjunction with CSIRO (Australian Government Research Organisation) and INEL, a French manufacturer of XRD analysers.

The analyser has been specifically developed to be located in the plant at either the kiln or mill, and to continuously analyse a stream of material flowing through it in order to provide real time analytical results from which automatic control action can be initiated. Batch mode analysis is also a standard feature of the analyser.

Commercial sales have begun in the US and Australia, and on-line real time continuous mineral analysis for plant control is now a reality.

Continuous XRD analysis is made possible by an innovative combination of unique technologies.

Firstly, a means of presenting a continuous sample stream suitably prepared for X-Ray diffraction was developed. XRD analysis requires an accurately prepared flat bed of material from which diffracted X-Rays can be detected. The sample stage is specially designed to prepare and present cement direct from the cement mill (or clinker after milling, or other powdered material) as a flat, continuously moving target bed. This sample presentation system was designed to place a continuous flow of material onto a rotating table that is then flattened before passing beneath the X-Ray beam. The analyser uses a point focus X-ray tube and a graphite monochromator (or multi-layer mirror) to produce a beam of X-rays at the sample stage. After passing through the analysis zone, the material is removed from the rotating table and returned to the material stream. This table samples continuously.

By continuous analysis of many moving crystals, the analysis is averaged over many orientations giving a more representative spectral pattern.

The continuous analysis is made possible by a curved position sensitive detector, that is able to detect a full range of 120 degrees of 2-theta simultaneously.

The detector is stationary, and therefore more durable and reliable than competitive scanning goniometers. However, the real advantage in this development is its ability to register a complete diffraction pattern continuously rather than building a diffraction pattern over time from slowly traversing an arc. This means the complete spectrum of diffracted X-Rays is collected simultaneously - a task not possible with a moving goniometer which takes a finite time to scan the diffracted X-Ray pattern.

Thirdly, a modified quantitative Rietveld code provides virtually continuous spectral analysis converting the live spectral data to phase composition. Due to the analysis techniques, the instrument is free from requiring standardisation, and will only periodically (annually for example) require check calibration. The instrument has been calibrated for the mineral phases of interest after extensive research and electron microprobe analysis.

The FCT-ACTech continuous XRD analyser has been designed to be located in the plant. It will communicate directly with plant PLC and control systems. Its data output can be adjusted easily with software. In the standard form, mineral analyses are updated every minute, while averaged values are integrated (typically over a 5 - 10 minute period) to give a smoothed result for control action. Trends are displayed over a nominated time period. Batch mode analysis is also possible.

Furthermore, elemental composition (oxide analyses for example) can be calculated with good accuracy from the mineral composition. For the cement industry, this is kind of a reverse Bogue calculation.

Results from plant trials were published in World Cement in February 2000 in ZKG in March 2001. Further publications will provide more detail on the various aspects of the analyser hardware, software and analysis techniques.

The studies mentioned validate the analytical performance of the analyser.

The stability of the analyser over a 15-hour period using a single sample rotating on the sample presentation/preparation stage was shown to be excellent. Table 1 right shows the mean and standard deviations for the major clinker phases and gypsum states in a sample of cement.

The accuracy of analysis is difficult to assess because there is no method of comparison that is reliable. However, some useful comparisons were devised.

Graph 1 right shows the very good agreement between the two techniques, indicating high confidence in the mineral analysis of the XRD.

Further methods involved weighing known amounts of free lime into clinker and analysing at high temperature to limit free lime hydration. These results are presented in table 2 below right.

Other tests involving addition of limestone of known quantities as well as comparison of XRD analysis with Bogue calculation gave further evidence that the continuous XRD analyser accurately tracks changes in the mineral composition of cement and clinker. These results demonstrate that the instrument provides very robust data that could be used for plant control purposes.

The performance of a cement plant, as well as cement product, is governed by mineralogy of the materials and not their chemical composition. The burnability and grindability of materials, for example, depends on the minerals present, and not the chemical composition as determined by XRF (X-Ray flourescence).

The strength development and setting times of the final product is also determined by mineralogy.

The on-line continuous XRD analyser now provides the means for measuring the mineralogy of cement and clinker directly in real time rather than calculating a theoretical composition periodically from laboratory analysis or theoretically using the somewhat assumptive Bogue equations.

Continuous analysis of a flowing stream of material has several benefits.

Firstly, the amount of sample analysed is very large in comparison to laboratory based techniques. The continuous XRD analyser examines as much material in 10 minutes as a laboratory based automated XRD analyser examines in 48 hours. This increases confidence in the result, and greatly reduces the potential for sampling errors associated with a small sample bead analysed in a distant laboratory.

Secondly, the measured result is available in what is essentially real time. This information can be used in an automatic control loop to change kiln burning conditions or cement mill operation for example. The lag time renders distant laboratory based control systems too great for automatic control - particularly if complete phase composition information is required.

Furthermore, the continuous XRD analyser provides a trend of analytical information in the same way that temperature or pressure can be trended, this means control decisions can be based on trended information rather than single analysis points. This factor contributes greatly to the robustness of a control system using mineralogical information in the control strategy.

The continuous XRD analyser accurately and simultaneously measures C3S, C2S, C3A, C4AF, CaO, and any other minerals of interest in clinker stream ground to 20% +45micron.

Furthermore, identification of alpha, beta and gamma phases of some of the clinker minerals is also possible, leading to the expectation that relationships will be developed between the phases in clinker, the kiln conditions and the performance of the product.

The final product can be analysed directly from the cement mill. C3S, C2S, C3A, C4AF, CaO, Ca(OH)2, CaCO3, CaSO4.2H2O, CaSO4.1/2H2O, CaSO4 and other crystalline minerals can be analysed simultaneously as the cement sample stream flows through the continuous analyser.

The on-line continuous XRD analyser opens new doors for optimisation and control of cement manufacture, and makes automatic control loops for quality control a reality. Benefits are expected to flow from a number of directions, and the following are seen as the primary creators of value from this new technology.

1. Plant Optimisation from new knowledge gained

The customer effectively pays $ /Mpa (or $/psi) when buying cement. The cement producer wishes to minimise his production cost in terms of $/MPa (or $/psi).

With on-line continuous monitoring of cement product mineralogy, the relationships between cement product performance and mineralogy will begin to unravel as correlations emerge through statistical analysis.

Cement strength depends on factors such as mineral composition (including pre-hydration, gypsum states, clinker mineralogy, limestone content), mineral reactivity, and cement fineness. However, it is rarely the case that a cement plant knows the inter-relationship between these variables. Although there has been some research into these factors, the volume of information that can be processed is limited with laboratory methods and in any case, the true relationships are probably plant specific.

The first benefit is therefore expected to be an identification of what targets should be established for cement mineralogy to achieve the desired product performance with respect to strength, setting time and other key indicators. As stated earlier, this is not really established in most plants, and production targets are based on many assumptions.

Once the clinker target has been established, the next stage is to monitor the clinker from the kiln and control the operation to ensure targets are maintained. This is likely to include burner control ensuring clinker product of good reactivity as established from mineral phase analysis using XRD.

2. Product Quality Control

Mineralogical monitoring of cement product with the continuous on-line XRD analyser, (along with on-line particle size analysis) will allow prediction of cement strength and setting times, and possibly other performance parameters.

If there is a variation in the predicted cement strength development, effect mitigating action can be taken (change in fineness, change in limestone or gypsum content, change of clinker source).

Setting times will be primarily a function of C3A content and reactivity and Gypsum de-hydration states. The continuous XRD analyser measures C3A content and forms as well as the Gypsum hydration states, thus providing the primary data for prediction of cement setting times. A change in predicted values can be counteracted by adjusting cement mill temperatures and therefore gypsum de-hydration.

Prediction, monitoring and control of cement quality at the point of final manufacture is likely to be a significant marketing advantage as well as significantly lowering risk associated with poorly performing product.

3. Reduced Laboratory Testing

On-line real time analysis and control for the cement mill or kiln product will greatly reduce the amount of testing required for either mill weigh feeder control, clinker burning control or quality assurance. In some cases, this may also greatly reduce testing protocols for product accreditation prior to dispatch.

Requirements for the total system include:

• Means for extracting a continuous sample of cement or clinker from the process
  
  
• Clinker grinding mill for clinker sample
  
  
• Continuous feed of cement or ground clinker sample to the feed hopper of the XRD analyser located in an air conditioned room near the sample point
  
  
• FCT-ACTech XRD with attached feed hopper and sample disposal system
  
  
• Specially developed Quantitative Rietveld analysis software for continuous interpretation of diffraction pattern into mineral phase analysis data
  
  
• Return of disposed sample into process
  
  
• Self sufficient analysis and data reporting locally at the XRD machine with provision to report data into plant data base and/or operator interface system
  
  
• PLC interface provision with XRD machine for control of stop/start/sequencing of total system including plant and sampling system

At the cement mill, a continuous sample of cement, approximately 600grams/minute, should be provided to the sample hopper on the side of the instrument. The sample could be provided by screw conveyor from an existing or new sampling system.

The instrument has been designed as a stand-alone unit, and it can be controlled locally or remotely from the plant control system. The unit is 1.8m high, 1.2m wide and 0.8m deep and weighs approximately 750kg.

Continuous XRD analyser is fully tested and calibrated prior to dispatch from our works. Client materials are typically investigated using electron probe and microscopic techniques to identify phases for analysis.

FCT-ACTech provides client design support throughout the project implementation for the whole system and support as required for the later performance of the system. Delivery is typically 26 weeks from placement of order.