Selection criteria for autolab of a modern cement plant
An automatic laboratory (Autolab) for cement plants comprises of five sub systems. A prospective buyer will have to focus on all these aspects for him to realise the true benefits of the investment in the automatic laboratory.
Sampling: Most products normally sampled in a cement plant are not homogeneous and therefore sampling bias can be far greater than analysis bias. Most plants have little understanding of what representative sampling is and therefore have only samplers that inherently have a bias that will exist till the analysis whose information is used to control the plant. This typically results in the process operators not believing the QC results and using other parameters such as temperature, pressures, flows, etc. to control the process. Whereas QC data is essential as it really provides information as to the true nature of process changes.
A representative sampler is one where each particle of the product flow has the same probability of being taken by the sampler. It has to be rugged, reliable and accurate. It is best advised that the primary sampler for powder sampling (raw meal, kiln feed, cement, fine coal) is located in the vertical chute just above the bucket elevator. This ensures that the sampled product is the final product that goes to the storage silo or the process. This vertical chute should have a minimum height of 2 m so that the primary auger, sample mixer and the secondary sampler can be fitted and the excess material can be returned to the same chute by gravity. Care should be taken during the engineering phase to ensure that this vertical height is available.
Some examples of proper sampling systems arrangements:
Sample cartridge transport: The representative sample fills up by gravity the cartridge that is in the sending station. One of the key selection criterions should be the life of the cartridges. The cartridges should have runners (the portion of the cartridge that is in contact with the transport pipe) that are long lasting and when required replaceable (without having to replace the whole cartridge which are costly).
The sealing system on the cartridges should be fail-safe, so that there is no risk of the cartridge opening during transport. If the cap opens during transport material will come out of the cartridge and risk blocking the cartridges later in the transport tubes (such as with cartridges using rubber sealed capsule closing caps).
The sample tubes should be of stainless steel and not galvanised steel to prevent rust. They should also have long length connectors, so as to avoid wear on cartridges if the tubes are not aligned properly. The diverters should be only two positioned and not multi positioned. Manufacturer's in an attempt to save cost offer multi position diverters that actually cause the cartridges to wear out faster as the pipes are not positioned correctly
Receiving station: The mechanism should be simple with as few actuators as possible. A full cleaning cycle is a must to remove all sample contamination. The sample should be unloaded automatically and the cartridge should be cleaned before being sent back to the sending station Sample preparation: The sample is prepared by using a mill and a press.
For XRF preparation, raw meal, hot meal, cement and clinker should be ground at an average particle size distribution of 3 to 4 microns as this will remove all the quartzite effects of silica. Only then will the XRF analyser be able to give us the right measurement. No wax or cellulose should be added to the sample preparation as this will dilute the sample and, depending on the type the dosing of additive to product ratio, will have a negative effect on the XRF measurement.
For XRD preparation for clinker or cement analysis, the sampler should be ground to an average particle size distribution of about 25 to 30 micron so that the peaks are visible under XRD analyser. If the sample powder is ground too fine, no peaks will be visible. Therefore different sample preparation techniques should be used to prepare sample for XRF or XRD analysis.
Thus no compromise should be allowed by way of a"one average" particle size distribution for XRF or XRD analysis. The press should be having a high pressure to ensure that the strength of the pellet and ensure that it does not break in the vacuum during XRF/XRD analysis. Once the preparation cycle is over an efficient cleaning cycle should ensure no sample contamination is taking place.
PLC and vision system: The PLC offered should be modern and the screens should be customised to customer's requirements. They should be easy to operate and update of the software should be seamless. Ideally the whole system (sampling, sample transport, preparation, interface with XRF and XRD) should be PLC based only, with no PC. A PLC based system is much more rugged than a PC based system and is less prone to software glitches or virus attacks.
Furthermore the XRF and XRD analysers should have all the required features (hardware and software) to integrate into a full lab automation configuration, depending on the type of sample preparation chosen.
ABOUT THE AUTHOR: The article is authored by B Venkat Kaushik, Operations Director, ITECA Instruments.