SIMULATION OF WATER EVAPORATION ON ULTIMONITORED THERMOMIX®: POTENTIAL APPROACH FOR STUDIES ON MILK EVAPORATION Simulação da evaporação da água em Termomix® multi-monitorado: abordagem potencial de estudos sobre evaporação de leite

It was simulated the industrial evaporation process using a multi-monitored Thermomix®. To validate the hypothesis that this equipment could act as a laboratory-scale evaporator, we evaporated water from milk to produce dulce de leche. Using this equipment, we were able to precisely determine mass changes throughout the process, observe the effect of the exhaust system on the evaporation, obtain dulce de leche with market composition, and determine the effect of sugar addition on water evaporation. The error between the observed boiling point temperature and that predicted by theory was ~ 2%; as a consequence, the equipment could be used to establish Dühring curves. It was found that the Thermomix® multi-monitored configuration was a simple and inexpensive equipment for simulate evaporation in industrial dulce de leche production.


INTRODUCTION
Evaporation of dairy liquids is an important step in the dairy industry. Sweetened condensed milk, dulce de leche, evaporated milk, and dairy powder products are the main dairy products obtained by evapora tion (SILVEIRA et al., 2013;SILVA et al., 2015;SILVEIRA et al., 2015;SCHUCK et al., 2016).
The energy expenditure is a problem involving this technique due to the removal of water by its phase change. For example, to concentrate atmospheric pressure 1 kg of 10 % sucrose solution to 20 % sucrose requires a total of 1439 kJ eliminating a total of 0.5 kg of water. Therefore, industries explore ways to reduce energy consumption and consequently lower costs.
It is admitted that this operation of concentration induces strong modifications of the physicochemical conditions not only impacting negatively the concentrates but also the process (TANGUY et al., 2016). Upon milk evaporation, lactose crystallization, Maillard reaction and an increase in viscosity are consequences of the heating and the concentration of the solids (SILVA et al., 2015;SOUZA et al., 2015;STEPHANI et al., 2015).
In view of the above, Thermomix, which corresponds to a process simulator, appears as a tool to study the manufacture of dairy products such as dulce de leche. This equipment has the advantage of less time of manufacture, besides its ease of handling, with use of smaller amount of ingredients, re sulting in less waste. Although it has as a disadvantage the high value and need for thorough washing.
Thermomix is equipment that can be used on a laboratory scale for the purpose of developing new products. By enabling the screening of certain technologies, reducing energy and material costs thus optimizing costs.
The aim of this study was to apply and validate the Thermomix as a laboratory scale evaporator, by optimizing water evapora tion and dulce de leche production.

MATERIAL AND METHODS
To emulate the industrial evapora tion pro cess, we used the Thermomix ® TM5 (Vorwerk, Wuppertal, Germany) coupled to a load cell (Ramuza IDR 7.500, San tana de Parnaíba, Brazil) with a precision of 1 g, a PT-100 temperature sensor, and an exhauster (Blower NáuticoSeaflo 3", XIAMEN HUILIYUAN, Xiamen, China). The process simulator can be visualized in Figure 1. The Sitrad software version 4.13 (Full Gauge Controls, Canoas, Brazil) was used to collect and record data from the load cell and thermometer at 1-s intervals. The supple mentary files include a photo of the equipment. To maximize the rate of water eva poration, we analyzed evaporation in light of heating power and the presence or absence of the exhauster. The experimental parameters are described in Table 1.  Aiming to evaluate the Thermomix as an emulator of evaporation in the food in dustry, we produced dulce de leche using two different formulations. Similarly, to maximi ze the water evaporation rate, the parameters of heating power and the presence or absence of the exhauster were optimized during dul ce de leche production. The experimental conditions for the dulce de leche produc tions are presented in Table 2. Three repeti tions were done for each treatment.
Data were collected according to Figure 2.
The evaluated parameters were: initial boiling temperature, T i (ºC); final boiling temperature, T f (ºC); initial time of the evaporation, t i (min); final time of the evaporation, t f (min); mass of evaporated water, M 1 (g); mass of total evaporated water, M 2 (min); heating rate, a h (ºC . min -1 ); constant evaporation rate, a e (g . min -1 ); total evaporation rate, E r (g . min -1 ); and air flow rate in the exhauster, V (m 3 . h -1 ). The total evaporation rate (E r ) can be calculated as where E r = total evaporation rate (g . min -1 ), M 2 = mass of total evaporated water (g), and t f = final time of the evaporation.
It is possible to divide Figure 2 in two zones. The first zone relates to the hea ting of the product or water until the boi ling tempera ture (heating zone). The second part relates to the water evaporation (evaporation zone), begins at t i (initial time of the evaporation) and T i (initial boiling temperature), and ends at t f (final time of the evaporation) and T f (fi nal boiling temperature). The amount of evaporated water is represented by the difference between M 2 (mass of total evapo rated water) and M 1 (mass of evaporated water).

Dulce de leche composition analysis
Moisture, water activity, fat content, and percentage of soluble solids were measured  Table 3 -Results for water evaporation and dulce de leche production experiments (n = 3) T  to characterize dulce de leche. The de termition of the moisture content was performed by the gravimetric method, while the lipid con tent was obtained by the Gerber method according Pereira et al. (2001) (both in duplicate). Aw analysis was performed in duplicate, and conducted on Aqua Lab 4 ATE equip ment (Pullman, USA). The quantification of soluble solids content was obtained in triplicate by refractometry using the Reichert AR200 equipment (Buffalo, USA).

Statistical analysis
The results were evaluated by variance analysis (ANOVA) and Tukey test for comparison of means (p < 0.05). Data were analyzed using the statistical program Statistical Analysis System (SAS Institute Inc., 2006) version 9.2, licensed to the Federal University of Viçosa.

Theory/calculation
According to Roos (2007), the increase of the boiling temperature, as a consequence of the concentration of the product, can be calculated as 2 * ( ) ln where DT b = boiling point elevation, R = gas constant (8.314 J . mol -1 . K -1 ), T wb = boiling temperature of water (373 . 15 K), a w = water activity of the product, and Dh v = molar latent heat of vaporization (J . mol -1 ).
The Dh v value can be calculated via equation 3 according to Morison;Hartel (2007).
Dh v = 57222 -44.3 * T wb (Eq. 3) Table 3 presents the results of water and dulce de leche evaporation experiments.

RESULTS AND DISCUSSION
As expected, the water evaporation occurred at a constant boiling temperature. During evaporation, the exhauster presence was related to the rate of water evaporation, and to the time at which evaporation began. In all tested configurations of the Thermomix, the exhauster delayed the beginning of evaporation; on average the time until evaporation start was 11.4% higher, with the minimum from W 105 (5.4%) and the maximum from W 115 (14.9%). However, during the eva poration phase, the exhauster contributed to an increased rate of water evaporation. On average the rate of evaporation was increa sed by 12.6%, with minimum for W varoma (0.76%) reaching 25.8% for W 100 . The duration of the evaporation was on average 10.7% lower, reaching 21.2% of reduction for W 100 . In dulce de leche industrial production, the presence of the exhauster is mandatory, primarily becau se it increases the evaporation rate, consistent with the present results. Simulations of the evaporation process should investigate the exhaust system or pressure reduction (vacuum evaporation), as demonstrated by Silva et al. (2015), Silveira et al. (2013), Rovedo et al. (1991), Martinez et al. (1990), andPauletti et al. (1990). Silveira et al. (2015) compared the efficacy of a pilot scale single stage evaporator in the evaporation of water and skim milk. In this study the heat transfer coefficient did not differ according to the product, while the flow behavior was modified. It concluded that the behavior of a product during the evaporation process can not be predicted by the global coefficient of heat transfer alone, requiring a wide range of information to understand the evaporation process, such as residence time distribution, product viscosity, and surface tension.
For the dulce de leche evaporation the exhauster reduced the time of evaporation by 25.5%, increased the evaporation rate by 33.5%, and delay the beginning of boiling by 6.4%.
The rate of water evaporation during the dulce de leche production experiments was not affected by the amount of added sugar (treatments D 20 and D 25 ), and no reduction in evaporation time was detected. The values reported here for the mass of water evaporated were, on average, higher than those found by Silva et al. (2015).
The compositions of the dulce de leche obtained in the experiment are presented in Table 4.
The dulce de leche produced using a multi-monitored Thermomix to emulate in dustrial evaporation of water presented compositions of fat, water activity, moisture, and soluble solids similar to those of industrial pro ducts from Brazil, Argentina, and Uruguay (GAZE et al., 2015a;ZARPELON et al., 2016). In a study by Gaze et al. (2015b) was found a variation between 3.56 g / 100 g at 6.99 g / 100 g fat and 17.49 g / 100 g at 29.67 g / 100 g moisture, while Silva et al. (2015) found a mean soluble solids content equivalent to 66.7 ºBrix (± 2.1). These results are in agreement with the present study.   Table 4 presents the relation between concentration and boiling temperature for dulce de leche.
The maximum error between the predicted value for boiling temperature, calculated from equations 2 and 3, and the boi ling temperature observed using the multimonitored Thermomix was of 2.08%. This result implies that the multi-monitored Thermomix can be used to establish Dühring curves for different foods during evaporation. Figure 3 shows rates of water evaporation for different configurations of the multi-monitored Thermomix.
Among all of the tested treatments, four sets of experimental conditions yielded ma ximal water evaporation (W' varoma , W varoma , W' 120 , and W 120 ).

CONCLUSIONS
The food processor brand Thermomix ® enabled us to precisely determine the varia tion of mass during the process, show the effect of the exhaust system on the evaporation, obtain dulce de leche with market composition, and determine the effect of sugar addition on evaporation of water during production. The error between the final boiling point tem perature and that predicted by theory was ~ 2%; as a consequence, the equipment could be applied to establishing Dühring curves. The Thermomix ® multi-monitored confi gura tion is a potential tool to emulate evaporation in the food industry, as demonstrated by our laboratory-scale production of dulce de leche. Futures studies should apply statistical methods to validate the Thermomix ® multi-monitored as an industrial simulator of evaporation.