I Proc. Int. Soc. Sugar Cane Technol.

, 24: 395-397

PERFORMANCE INDICATORS
2001

STEPHEN J . CLARKE
Florida Crystals Corporation. USA

Abstract
The information required by sugar technologists for more complete understanding and control of
process operations in a cane sugar factory is described, including basic material balance data and
factors related to individual unit operations. Discussion items are listed with questions related to
basic operating conditions, recognising that there are no single or simple answers. The need for good
on-line process measurements is stressed, especially as factories increase in size and complexity, and
energy issues become more important. The final discussion issue is the potential of benefits from
statistical and model-based performance indicator systems.

Introduction comments and/or questions related to our state of

The goal of the cane sugar factory technologist is knowledge in the particular area.
to recover as much sugar as possible at high produc-
tion rates, of optimum quality and with minimal costs 1. Material balances and data acquisition
in equipment, energy and personnel. Cane quality Good material balances, for each component of
plays a major role in factory performance but is not the process streams, are an essential starting point in
essential to this discussion which should focus on performance monitoring. Without both mass and
how we can know whether we are meeting our goals composition data, other indicators of process con-
of performance efficiency. Inadequate cane supply, ditions are of dubious value. A good balance of
often a significant problem, is also not really relevant materials in and out of process is not necessarily an
and the techniques we choose as performance indica- indication of good performance, e.g. high sucrose
tors should apply both when the mill is operating content in final molasses, perhaps as a consequence
steadily at full capacity and when there are problems. of minimal sucrose degradation, is not an indication
We need to avoid feeling comfortable when opera- of an efficient process. Conversely, low final
tions are steady and production meets routine expec- molasses purity may well be the result of excessive
tations. inversion but this could be unknown unless a reliable
The third edition of the classical 'System of Cane mass balance is known.
Sugar Factory Control' was published in 1971 Does the factory have sufficient scales to
(Clayton, 1971). The following statement is from the monitor all materials in and out of process? This
preface to the first edition: 'Your committee should is more complex when an annexed refinery
strive for three main objectives, namely, accuracy, takes the raw sugar and returns run-off to the
clarity and simplicity; and of these three I consider raw house.
the last as important as the first two. In striving for Are flow meters and other on-line devices suffi-
greater accuracy, formulae have become so complex ciently accurate to replace scales? What level of
that they are practically.useless.' The purpose of this accuracy is necessary in this instrumentation, if
workshop discussion of performance indicators is used for this duty as well as for process control?
neither to reinvent what has already been done nor to Can any losses (or gains) in material in process
replace the traditional means of performance compar- be determined with confidence? For example,
ison but rather to look as some of the issues and can the quantities of non-sucrose, especially
opportunities presented to us at this time. Although invert, be reconciled between that in the juice
we now have the technology to avoid the drudgery of into process and that leaving in molasses, etc.?
complex calculations, we must strive for simplicity of How much confidence do we have in the
concept and relevance to operations in any means estimates of undetermined (or unaccounted for)
that we use to monitor factory performance and avoid losses? (An observation is that factories with
approaches that are 'black boxes'. We must avoid the good data on material balances show higher
problems of a disconnect between operators and undetermined losses than for factories where the
'techies' that can occur if the basics of the measure- balance is estimated.)
ments used are not appreciated by the users. Are the routine analytical methods sufficiently
I accurate to provide the required data and in a
Discussion topics timely enough manner? If not, what additional
The discussion of performance indicators has methods are required? Are these analyses per-
been arbitrarily sub-divided on the basis of the type formed with a purpose or merely as established
of information involved and each section ends with practice?

KEYWORDS: Losses, Material Balances, Analysis, On-line Sensors.

-

Clarke, S.J., Proc. Int. Soc. Sugar Cane Technol., 24: 395-392001

Are on-line analytical methods (NIR, etc.) good Evaporation

enough for these purposes? What level of Sucrose losses and colour development during
accuracy is required? Are on-line systems evaporation can adversely impact recovery and
merely faster and less demanding of personnel? sugar quality. Do we have sufficient informa-
How useful are trends in composition rather tion to be able to relate time, temperature, pH,
than absolute numbers for specific components? etc to losses during evaporation?
Are data collected on carryover into condensers, Should data obtained by automatic evaporator
leaks and floor drains, etc.? control systems be used to give on-the-run values
for heat transfer coefficients and so monitor (and
2. Data evaluation and equipment performance minimise) the effects of scale formation?
The data described in the previous section should
be used to evaluate process efficiency and the calcu- Crystallisation and centrifugation
lations described in the 1971 ISSCT publication Should crystal yields be monitored for each
(Clayton, 1971) are acceptable for comparison of massecuite and included in crystalliser opera-
factory performance. However they are perhaps tion? This can be very tedious but how close are
rather limited by arbitrary assumptions (e.g. Winter- we to having this data directly, preferably on-
Carp) and their primary application to the total line, or at-line? .
process rather than to the individual unit operations. * Can optical systems be used to monitor crystal
These are the sites of inefficiencies and should be size distribution and improve centrifugal
investigated individually. throughput and sugar quality?
Equipment design and limitations vary widely and Can data obtained by automatic vacuum pan
we should also consider limitations in support control systems be used to give on-the-run
services, especially steam and electric power. The values for crystallisation rates and heat transfer
critical questions are 'How do we know that we are coefficients and optimise pan performance?
using the existing equipment to the fullest extent pos- What are acceptable molasses purity increases
sible?' and 'How do we decide on what are the during centrifugation?
process problems and bottlenecks?' We need to dis- Are on-line (or at-line) massecuite consistency
tinguish between equipment limitations and material measurements practical and useful?
limitations. A factory set up to handle cane with high How do we apply final molasses target purity
purity juices will almost certainly have problems with values to our operations?
high levels on non-sucrose entering the boiling
house. Raw sugar quality and storage
* Is raw sugar quality significantly affected by
Clarification and filtration incoming cane quality or process upsets? If so,
Does anyone have a practical definition of how can the process be controlled more
'clarification efficiency'? efficiently to minimise such problems?
What are optimum liming techniques and Is raw sugar quality monitored routinely and
clarified juice pH? How much variation in pH is what factors are used to decide on optimisation
acceptable, especially with varying cane of this facet of operations?
quality?
Sucrose losses in clarifiers may become exces- 3. Modeling and statistical systems
sive with long residence times and these can be In a paper presented at the 231d ISSCT, the dis-
measured using chemical tracers. CFD tech- tinction was made between primary technology
niques can be used to evaluate and enhance (mills, clarifiers, centrifugals, etc.) and secondary
clarifier performance. technology, information technology based systems
Should clarified juice turbidity be a measure of for design, control and optimisation of processes
clarifier performance? If so, how is this (Clarke, 1999). This approach does not supercede
measured, preferably on-line? Can these traditional systems but does allow sugar technologists
measurements be used to optimise addition of to understand inore fully the variables involved in
phosphate and flocculants? maximising factory performance.
Is the measurement of mud level and density in The chemical processing industry is increasingly
the clarifier a useful measure of how well the relying of information technology for process optimi-
clarifier is being operated? sation. Data mining and statistical process control
The retention of solids at the filters is easily . offer the raw sugar industry the opportunity to
measured and gives a measure of the recycle of improve our understanding and control of processes.
insoluble material to process. Data reconciliation is a branch of process monitoring
What are acceptable purity differences between and consists of evaluating robust on-line measure-
mixed juice, clarified juice and filtrate? Is the ments with the object of determining reliable
measurement of degradation products, e.g. estimates of process variables. When a process is
lactic acid, useful? running optimally the measurements and the values
Is sucrose % filter cake a sufficient measure- derived from them (by statistically sturdy computer
ment without data on the quantity of filter cake models) stay within established values. Deviations
and filter retention? from optimum become quickly apparent and remedial
 Clarke, S.J., Proc. Int. Soc. Sugar Cane Technol., 24: 395-397 2001

action can be taken. Examples could be massecuite Conclusion

consistency and continuous centrifugal operation. Attention to the details of the sugar production
The achievement of these goals will depend on process and careful choice of performance indicators
the development of the appropriate sensors to should allow factories to improve performance. The
measure the critical parameters in the process vessel final question is whether the significant effort
(Bullock et al., 2001). Work On such Sensors is involved can be financially justified,
ongoing but we will need to put much effort into their
practical application to achieve our goals.

REFERENCES
Bullock, G.E., Edye, L.A. and Doherty, W.O.S. (2001). Recent advances in unit process design and measurement system development at
1 the Sugar Research Institute. Proc. Sugar Industry Technol., 2001, in press.
.ke, S.J. ( 1999). Productivity increases in sugar mills through new technology. Proc. Int. Soc. S u g a ~ :Tec:hnol.,
rton, J.L. ed. (1971). System of cane sugar factory cc~ntrol.Published for the ISSCT by the QSSCT (3 rd editio