Some Case Studies Indicating the Benefits Derived

• Ball Mill Optimisation : A cement plant had three cement mill circuits producing PPC and PSC and none of them was operating at its rated capacity.  They tried to optimise them in every conceivable way and ultimately commissioned our services for analysis and optimisation.  Upon using our softwares BMSIMUL, CLASANAL and SIZEANAL, it was discovered by us that fundamentally, it was not a grinding problem at all!  It was essentially a heat transfer problem and accordingly, using these softwares (BMSIMUL considers heat transfer in mill circuits in great detail), alternative operating strategies were worked out. These resulted in increased throughputs of 30% thereby enabling them to achieve higher than rated throughputs.  It was after this that they were able to sell their plant to an MNC for Rs. 8.2 billion.
  
• Classifier Bypass : In a closed raw mill circuit operating at high air speeds, the use of CLASANAL showed that while the classifier per se was giving a decent circulating load of nearly 100% of its fines product stream, the total amount circulating in the circuit was extremely low at 20% of the fresh feed to the mill.  It was therefore decided to remove the classifier altogether from the circuit and operate the mill in open-circuit mode.  As predicted by BMSIMUL, there was hardly any effect on the throughput but there was a significant saving in power consumption.  It was thus ironic that the classifier software CLASANAL was used to remove the classifier!   CLASANAL is thus useful to know the classifier performance parameters and  BMSIMUL is useful to study their impact on mill circuit performance and the two softwares together are extremely useful for evaluating possible conversion of open-circuit mills into closed-circuit ones and vice-versa.
  
• Raw Mill Stabilisation and Optimisation : In one cement plant, they were puzzled by the fact that while their Raw Mill 2 circuit was apparently similar in all respects to Raw Mill 1 circuit at the same location, similar operating conditions led to excessive power consumption and highly unstable operation.  We were the ones to correctly identify the "finer differences" between the two mill circuits using BMSIMUL, CLASANAL and SIZEANAL and optimise the RM2 circuit.  This led to not only stable operation, but also a reduction in power consumption by as much as 700 kW.
  
• Aluminium Ore Plant Optimisation : Their wet grinding mills were not giving the rated outputs.  Without addition of any capital equipment, purely through simulation, we were able to effect an increase of 30% in the throughput.   However, we also gave them detailed simulations on the possible effects of installation of hydrocyclones using our softwares BMSIMUL, CLASANAL and HCYCLONE.
  
• Fertilizer Granulation Plant : The overall plant software GRAND which includes the module RDSIMUL for rotary dryers has been successfully used to simulate and optimise over a dozen rotary dryers and granulation plant installations.  The same has also been used to design the whole process and sizing of equipments for a number of plants which have been successfully commissioned and are in operation in various parts of the world.  The packages have led to significant increases in throughput, reductions in fuel and power consumption, and improvement in product moisture content by controlling moisture re-absorption.
  
• NPK Plant Optimisation : A large NPK granulation plant had two rotary dryers in series because the CRH value  was low, only 42% at 60°C.  The dryers processed about 200 t/h of 19-19-19. The solids should not get heated up beyond 70°C at dryer outlet, otherwise they degrade.  It was shown through simulations using RDSIMUL that indeed two dryers in series were required, to first dry from 2.8 to 1.8% followed by 1.8 to 1.4% moisture content.  The software RDSIMUL provided vital answers to improve the throughput by roughly 10%.  It was shown that installation of an air pre-dryer was not useful, dryers could not be run in parallel, increasing air flow would lead to greater dust entrainment, increasing the throughput would lead to higher outlet moisture content, and the only option was to increase the air temperature marginally (by 10 deg.C) with air mass-flow rate kept constant.  All the other performance and quality parameters were also predicted by RDSIMUL and the same were found to be acceptable and manageable.  This actually led to the throughput increasing by 10%.  Further, the plant avoided wasteful expenditure of almost USD 1 million on an air pre-dryer which they would have installed otherwise.
  
• Organo-NPK Plant Optimisation : An NPK plant using up to 30% organic materials in feed was giving a production of only 4.3 t/h although the plant was rated for a higher capacity.  The outlet solids and gas temperatures were 72 and 88°C respectively.  The rotary dryer could dry a combined feed (including recycle of 6 t/h) of 10.3 t/h from 12 to 3.1% moisture content on wet basis.  GRAND was used to increase the plant throughput from 4.3 to 7 t/h.  The simulations indicated a low material filling of 7.7% and solids residence time of 12 min.  Since the hot air generator fan was operating at only 60% damper opening, it was suggested to increase the same to effect a 30% increase in hot air flow, increase in hot gas temperature from 400 to 500°C, and a reduction in dryer rpm from 6 to 5 to obtain a higher material filling of 13.6% and solids residence time of 14.8 min.  The resultant higher power consumption was fortunately possible with the same drive motor.  For a further increase in throughput to 8 t/h, it was suggested to increase steam addition rate in granulator to enable solids entering the dryer to be heated up to 60°C from 37°C otherwise.
  
• Particle Size Analysis : In one case, implementation of SIZEANAL resulted in a direct saving of Rs. 3 million in expensive instruments' cost besides improvement in the operation of equipments.  This case has been published in ZKG, 41, 82-86 1988 and World Cement, 13, 38-44, Oct. 1991.  The theory of the same has been published in a number of reputed research journals including those of the American Chemical Society and summarised in our series of technical reports on the subject.  SIZEANAL has greatly helped to understand the extent of fines in sub-sieve ranges in a number of powders which has, in turn, helped to optimise a number of grinding circuits.  Further, the size distributions predicted by SIZEANAL have enabled understanding the operation of a number of classifiers (by measuring feed, overflow and underflow) leading to their optimisation.
  
• Cyclone Design and Optimisation : In one process plant, for collection of dust for pollution control purposes, a cyclone followed by a multiclone was being employed.  Yet, the emission norms were not being met.  The cyclone design was analyzed using the CYCLONE software.  Various combinations of cyclone dimensions were tried and the one that met the required collection efficiency was chosen.  With this, not only were the emission norms met, but the multiclone was also rendered redundant.  Numerous such studies have been carried out in which the companion software SIZEANAL has also been extremely useful.
  
• Energy Conservation Using TEAM : The implementation of TEAM software in a cement plant led to the identification of optimisation potential in each individual equipment, over-loading and under-utilisation of specific equipments, identification of optimum throughput range for each equipment, elimination of idle running of auxiliaries etc. which has guided them to effect a reduction in overall specific energy consumption from 120 to 90 kWh/t of cement.  This improvement was gradually achieved over a period of 3 years by the plant personnel whose diligence and commitment in implementation can not be underestimated.
  
• Brick Lining Life Optimisation : The refractory brick linings in the kiln get damaged and fall which need replacement.  Sometimes this can become frequent and pose serious problems to the operation of the plant itself.  In one 0.6 million TPA cement plant, this problem was so acute that the bricks would fall within two weeks of their laying in the kiln.   The problem was analysed and traced to faulty raw-mix design which was giving poor liquid content and coating index, and improper estimation of coal consumption.  The problem was solved using RAWMEAL software.  The brick lining life eventually reached the normal figures of  6 months enabling the plant to improve its capacity utilisation from a dismal 35% to over 90%.
  
• Technical Forecasting : Kiln feed is controlled by controlling the differential pressure in cement plants.  Conventionally, a linear fit is attempted to establish the relationship using the few data available.  It was found in this plant that while most of the time, the kiln would operate stably with a linear fit, on some days especially with low throughput, the kiln would become unstable with the clinker getting hard-burnt.   The software GITA  was used to simply determine the best "curve" for the data points.  The predictions of this curve were shown to match well with those of the straight line fit for most of the cases (of high throughput), but the same deviated markedly for low throughputs.  Further, GITA's predictions clearly indicated that for achieving a certain low throughput, one needed to actually set a higher differential pressure than was being done following the straight line fit which explained why they were getting hard-burnt clinker.  A simple change from a straight line fit to the best curve thus resulted in improved and stable kiln operation as well as avoidance of loss of hard-burnt clinker.
  
• Sale Forecasting : Sale forecasting is an application where the software GITA has been found to be extremely useful.  This has been actually done for an FMCG company making personal care and hygiene products.  Here, an innovative method was employed before applying GITA.  The "weekly sales" figures were converted into "cumulative sales till the week" figures, which then increase monotonically with the week number.  This monotonic relationship is amenable to mathematical analysis using GITA.   The cumulative sales for future weeks can then be predicted from past values.   The sales for any future week N equals the cumulative sales predicted for week N minus the same for week N-1.  The accuracy obtained using this method was in excess of 95%!  This seems to be an excellent procedure because the influence of all the important parameters are already factored in into the past sales figures and therefore the future trend predicted using GITA can be used to take policy decisions like advertisement budget, field sales personnel, sales outlets etc.  This method of sale forecasting using GITA is especially useful for FMCG, pharma companies etc.