As presented by Jakab Rózsa at the 2019 SIT Annual Conference in Durban, South Africa

Batch cooling crystallization is a widely practiced production process in several industries. In the sugar industry it is mainly used as the first step to produce massecuite (footing magma) to seed batch or continuous evaporating crystallizers and to increase the final exhaustion of sugar from the basic raw material (beet or cane).
During batch evaporative crystallization, widely used in the sugar industry advanced instruments (process refractometer, microwave and SeedMaster devices etc.) can provide quite many online data (including supersaturation) on the most important parameters of the massecuite. Contrary to this practice, instrumentation in cooling crystallizers used in the sugar industry is rather poor: monitoring and control of the process mostly relies only on the measurement of temperature.
This post reports on a series of tests carried out in a sugar mill in Central Europe to monitor online the most important parameters (supersaturation, crystal content and mean crystal size, product yield, mother liquor purity etc.) by using a built-in process refractometer and a SeedMaster device.

When visiting sugar mills and refineries in different countries it is common to find a series of batch evaporative crystallizers equipped with a large set of instruments. The times are over when the operation of the vacuum pans relied exclusively on the expertise of the pan-men and on the exclusive use of the vacuum gauge and the mercury thermometer. Even the more sophisticated devices measuring boiling point elevation, density, or some electrical property of the massecuite (simple conductivity or radio frequency probes) have lost ground on the pan floors. Their roles have been taken over by process refractometers and microwave instruments, which are widely used to measure liquid and total solids concentration of the massecuite, respectively.

It is well known that the most important parameter of crystallization is supersaturation. It is certainly true not only for the case of sugar crystallization, but also for the crystallization of other products, like pharmaceuticals as well.

Attempts to develop an instrument for the online measurement of supersaturation in batch vacuum pans, which are widely used in the sugar industry, had limited success. The reason: supersaturation is a function of several parameters of the technical sugar solution, like concentration, temperature, non-sugar content and syrup quality parameters, all of which have considerable influence on it. However, without reliable information on the “driving force” of crystallization, its control still was and is quite often even today reminiscent of the “trial and error” practice of the old times.

The situation with batch cooling crystallization in the sugar industry is somewhat different. When filled to full volume, there is no further syrup feed in the crystallizer. There is no evaporation, and the amount of the different components (sugar, non-sugars, and water) remains constant all over the time. However, the role of supersaturation during batch cooling crystallization remains the same: it is the most important parameter of the process, and its monitoring and control is of primary importance. During the cooling crystallization of sugar, the only control parameter is temperature, so most of the papers published on the subject are devoted to the discussion of the development of an optimal time dependent temperature trajectory. There are several ways to develop this trajectory, namely:

By the familiar and traditional trial and error method, which is sensitive to process disturbances (for example: changes of the feed syrup parameters) and can be a never-ending process lagging by one batch time behind real time.
By using a model-based batch-to-batch control strategy. It is a step-by-step method hopefully leading to a good temperature reference profile.
By implementing supersaturation-based control of the process temperature.

Batch cooling crystallization of sugar

One can find many papers and presentations about batch cooling crystallization. Most of them are trying to develop a kind of control which results in optimal process parameters. These parameter requirements are usually:

Product crystal size and size distribution meeting target requirements,
Crystal content and product yield,
Time of crystallization.

The methods of control range from simple temperature or concentration control using prescribed trajectories (set point values) to different, more elaborate model based iterative (batch to batch) learning control solutions. Development of temperature or concentration profiles for control however is a time-consuming process and similarly to model-based control, they are prone to changes of the different parameters (initial concentration, temperature, purity etc.) of the highly nonlinear process of crystallization.

The different methods of control used for cooling crystallization of sugar in the industry still rely on a very restricted use of instruments. Online data on the process in most cases is reduced to temperature, with the addition of occasionally taking samples for laboratory analysis (to determine the time of correct seeding, for example). The result is a lack of online information on important parameters, something like the list of those available when using process refractometers and SeedMaster instruments.

Sensing the need to provide more online information on the process, first on supersaturation for its advanced control, we decided to develop a version of the SeedMaster-3 device, capable to fill this need. Later, we developed the cooling crystallizer versions of SeedMaster-4 based on the experiences gained with this prototype. In this post we discuss the results of a series of tests carried out in a sugar mill in Hungary to monitor the process of cooling crystallization with SeedMaster-3.

Production of footing magma for the product pans

It is common practice since a long time in the mill where the tests had been carried out to use a cooling crystallizer to produce footing magma for the product pans. The process starts with concentration of standard syrup in a normal product pan. Part of the concentrated syrup is fed to the cooling crystallizer (typical target parameters: concentration: 77 %, temperature: 64 °C, purity: 94.5 %, supersaturation: 1.07). Having filled the crystallizer, the pan is filled again to seeding level, and using previously prepared footing magma for seeding from the magma receiver, continues operation as a normal product pan.

Instruments

According to local practice, crystallization is controlled by using online temperature data from a temperature transmitter and occasional laboratory data based on massecuite samples.

Testing of the experimental SeedMaster-3 version was based on the use of

one K-PATENTS process refractometer and
one SeedMaster-3 device with added new software to serve cooling crystallization applications.

The sensor head of the refractometer, which provides online data on syrup/mother liquor concentration and temperature, is mounted in the bottom of the crystallizer.

The most important parameters provided online by the SM-3 instrument are listed below.

Temperature (°C or°F)
Mother liquor concentration (%)
Supersaturation (--)
Mother liquor purity (%)
Crystal content (% vol.)
Crystal size (mm)
Massecuite density (kg/m3)

Test results

During testing the monitoring of cooling crystallization, process data on 10 consecutive strikes were collected. What follows is a summary of the results with some comments on the recorded data.

Massecuite temperature

Recorded from 10 strikes.

Mother liquor concentration

Recorded from 10 strikes.

According to the trends on temperature, crystallization is carried out roughly in the 62...64 to 28...30 °C range of temperature. When reaching the low end of the range, cooling is not continued anymore, but dropping the massecuite may be delayed due to lack of receiver capacity.

Liquor concentration changes in the 77 to 70 % range.

Trends on temperature and liquor concentration during the 10 recorded strikes were similar, except those which belong to strike No. 6. Temperature was controlled according to the same pre-defined curve, but there was a slight delay at the beginning of the strike. We will see on supersaturation and crystal content trends what a huge difference this makes.

Supersaturation

Recorded from 10 strikes.

Mother liquor purity

Recorded from 10 strikes.

Supersaturation is defined as sugar in solution over sugar at saturation (both at the same temperature). After seeding the crystals begin to grow and purity of the mother liquor drops accordingly. It is natural that this drop in purity is considered when calculating supersaturation in the SeedMaster devices. The trends on supersaturation show data in the SS = 1.01 to 1.14 range, where spontaneous nucleation will not take place in the bulk of the massecuite (feed syrup purity: ~ 94.5 %).

Based on the data on supersaturation full seeding with slurry is practiced in this mill.

In three of the ten strikes supersaturation reaches, or even is much less than SS=1.035, the value which is the high limit of the “NO CRYSTAL GROWTH AREA”.

Trends of the characteristic crystal size clearly show that when supersaturation approaches (drops to) the high limit of the “NO GROWTH AREA”, crystal growth will be ended accordingly.

Crystal content

Recorded from 10 strikes.

Crystal yield

Recorded from 10 strikes.

The growth rate of crystal size drops fast as the operating temperature decreases. So, to continue cooling crystallization to too low temperature will not bring much increase neither in crystal content, nor in product yield, but the time of crystallization will be increased considerably.

Massecuite density

Recorded from 10 strikes.

Conclusion

Production of good quality footing magma for the product pans is an important step in sugar manufacturing. The correct operation and control of the process needs reliable online data on its most important parameter that is on supersaturation.

It can be seen on the recorded trends that strike No. 6. performed very poor compared to all other strikes. That is because there was an issue with cooling water at the beginning. Although the usual temperature curve was used, results were far worse than usual. With supersaturation-based control, this issue could be eliminated, that is why it is the most important parameter of crystallization.

Besides supersaturation, however, there are quite many important parameters which by considering the local constraints, can be used to select the optimal operating parameters (for example: feed syrup concentration and temperature, temperature range of the cooling water, seeding practice etc.). These parameters have a very important effect on the final parameters (crystal size and content) of the magma to be used for seeding in the product vacuum pans.

It is well known that advanced control of crystallization needs reliable online data on supersaturation. It is also well documented, that supersaturation is a function of several variables (liquid concentration, temperature, purity, and quality parameters), therefore reliable data on it can be provided only by calculation, considering all these parameters.

Development of the experimental version of SM-3 and later SM-4 was aimed to provide an instrument capable to serve the needs of advanced vacuum and/or cooling crystallization control, based on reliable online data on supersaturation and other important process parameters.