**Energy recovery in spray-drying processes can reduce production costs, but involves trade-offs. Using milk powder as an example, this article describes a spreadsheet method to evaluate exhaust-gas recirculation scenarios for optimizing spray drying**

Dairy processing is among the most energy and carbon intensive operations in the global food processing industry. Estimates say dairy farms are responsible for 4.4% of total anthropogenic greenhouse gas (GHG) emissions, which is approximately 2.1 Gt/yr of carbon dioxide equivalent (CO_{2e}) [*1, 2*]. Within dairy processing, cheese production has the highest energy requirement (8.33 MJ/kg product) [*3*], followed by milk powder processing (4.60 MJ/kg product) [*4*]. The large energy requirements in milk powder production are primarily due to the drying and evaporation processes.

With improvements in energy efficiency, it is predicted that the energy consumption of the global dairy processing sector could be reduced by 50 to 80% — equivalent to 1.05–1.68 Gt/yr of CO_{2e} [*5*]. Therefore, the food processing industry has increased attention on energy recovery and heat integration of drying processes as a main priority in efforts to save energy and reduce production costs [*6*].

## Process description

Drying systems in food manufacturing factories can produce a range of milk powder grades, such as whole- and skim-milk powder. The system considered here includes a tall form dryer, as well as a spray dryer connected to a fluidized-bed dryer (Figure 1). To enable exhaust recirculation, we investigated the addition of a dehumidification system and a wet scrubber. For the experiments, it was assumed that the plant was located in Malaysia and has an annual operating time (AOT) of 6,000 hours.Ambient air at 30°C with a relative humidity of 0.0211 kg _{water} /kg _{dry air} mixes with the recirculated air and enters the adsorber. Post drying, silica gel dehumidifies the process air (including the recirculated exhaust air), before being sent for the drying process of milk concentrate. The silica gel releases the bound water and regenerates as hot, dry air passes through it.

The integrated spray-dryer and fluidized-bed dryer system produces powder creamer of low dairy content from milk concentrate. In the spray-drying chamber, atomized liquid droplets meet hot air at 160°C with a relative humidity of 0.016 kg_{water} /kg_{dry air} to become creamer. The remaining material enters the fluidized-bed dryer with a moisture content of 0.06 kg H _{2} O per kg of creamer. Note that the drying activity mainly takes place in the spray-drying chamber, while the fluidized-bed dryer mainly cools the product and prevents the powder from agglomerating excessively.

On the other hand, the tall form dryer produces lower-grade milk powder from concentrate. In its drying chamber, atomized milk droplets come into contact with hot air at 180˚C with a relative humidity of 0.016 kg_{water} /kg_{dry air}. The difference in temperature profile for both dryers results in products with different qualities. The final moisture content of the milk powder exiting the tall form dryer is 0.03 kg H_{2}O per kg of powder. The exhaust gas from the spray dryer and tall form dryer passes through a filter bag and cyclone to recover suspended powder particles.

## Case study

In this work, we look at three concepts for the recirculation of exhaust gas with comparison with the current process operation.- Base case (no exhaust-gas recirculation)
- Recirculation of exhaust gas while considering the humidity limit of drying air only
- Recirculation of exhaust gas while considering the humidity limit of drying air and fines limit of exhaust gas
- Recirculation of exhaust gas passes through a wet scrubber its fine particles are removed

## Optimization model

** Scenario 1** describes the base case without exhaust gas recirculation. The total energy consumption (

*Q*

_{total}) of the heaters was calculated using Equation (1). The unit price of electricity (

*C*

_{elec}) is $0.08/kWh based on a typical industrial tariff [

*7*]. The total cost (

*C*

_{total}) for heaters duty is calculated by using Equation (2). The base case serves as a benchmark for subsequent scenarios.

*Q*_{total} (kW/h) = *Q*_{SD} + *Q*_{TD} + *Q*_{reg} **(1)**

*C*_{total} ($/yr) = *Q*_{tota}_{l} × *C*_{elec} × *AOT* **(2)**

Where *Q*_{SD} is the heat duty for spray dryer, *Q*_{TD} is the heat duty for tall form dryer and *Q*_{reg} is the regenerative duty. The result shows that the total heat duty is 498 kW, with a cost of $241,708 per year.

** Scenario 2** describes the system when part of the exhaust gas from the spray dryer and tall form dryer is recirculated. However, the process air for the dryers needs to have an absolute humidity (

*y*

_{ada,out}) of below 0.016 kg

_{water}/kg

_{dry air}to ensure efficient drying. By implementing the model in Microsoft Excel as shown in Figures 2 and 3, it is easy to investigate the effects of the two recycle ratios and minimize the overall energy consumption of the system:

*min* (*Q*_{total}) **(3)**

The critical variables to optimize include the air-recycling ratios for spray dryer (*R*_{1}) and tall form dryer (*R*_{2}), and the temperature for regeneration of silica gel (*T*_{Ra,in}). Because the model is linear, the Simplex (standard method of maximizing or minimizing a linear function of several variables under several constraints on other linear functions) LP method using Microsoft Excel can quickly minimize the objective function (Equation 3) given the following constraints:

0 ≤ *y*_{ada,out} ≤ 0.016 kg_{water}/kg_{dry air} **(4)**

150 ≤ *T*_{Ra,in} ≤ 180˚C **(5)**

0 ≤ *R*_{1} ≤ 1 **(6)**

0 ≤ *R*_{2} ≤ 1 **(7)**

As *R*_{1} increases from 20% to 80%, the duty for dryer heaters decreases by as much as 16%. On the other hand, as *R*_{2} increases from 0% to 60%, the dryer heaters’ duty reduces by up to 15%. However, as *R*_{1} and *R*_{2} increase, the regeneration duty increases as well. When *R*_{1} increases from 20% to 80% and *R*_{2} from 0% to 60%, the regeneration heater duty increases by 2.3% and 2.1%. As *R*_{1} and *R*_{2} increase, the absolute humidity of the air into the adsorber increases, leading to a faster saturation of the silica gel. As a result, the silica gel requires additional hot air from a heater to regenerate. For this specific case, the optimization suggests that that the entire exhaust gas from the spray dryer (*R*_{1}) and 18% of the exhaust gas from the tall form dryer (*R*_{2}) should be recycled. This optimum combination of recycling ratio has a total duty of 409 kW with an annual cost of $198,369.

** Scenario 3** assumes the adsorber has a fines limit (

*x*

_{f,r}) of 45 parts per million (ppm). As a result, the amount of fines within the recirculated gas should be limited. An additional constraint in the model, Equation (8) provides for the fines limit. The exhaust gas from cyclone and filter bag is assumed to contain 100 mg/m

^{3}and 50 mg/m

^{3}fines.

0 ≤ *x*_{f,r} ≤ 45 ppm **(8)**

Now the objective in Equation (3) can be minimized subject to the constraints in Equations (4) to (8). In this case, the optimum recycling ratio obtained is total recirculation from the spray dryer (*R*_{1}) and 2% from the tall form dryer (*R*_{2}). Under these recirculation conditions, the heater duty reduces to 420.10 kW. The total cost for heater duty is $203,865/yr.

** Scenario 4** proposes a wet scrubber to actively remove fines in the exhaust gas. The wet scrubber has an estimated removal efficiency of 90% for a liquid-to-gas ratio (L/G) for the wet scrubber of 0.94 kg/m

^{3}[

*8*]. A final assumption is that the wet scrubber operates adiabatically.

Last of all, the model can be re-run to achieve the condition in Equation (3) subject to the constraints in Equations (4) to (8) and a fines upper limit in the recirculated air of 4 ppm. The optimum recycling ratio now reduces to 59% from the spray-dryer exhaust air (*R*_{1}) and no recirculation from the tall form dryer (*R*_{2}). Although the purpose of the wet scrubber is to remove fine particles to increase the recycle ratio, the result shows that the maximum recycling ratio decreases with the introduction of a wet scrubber. This is because when the exhaust gas passes through the wet scrubber, it adds water to the air, increasing its absolute humidity and decreasing its temperature drastically due to water evaporation [*9*]. Due to the high absolute humidity of the wet scrubber outlet, and the constraint of process air absolute humidity of less than 0.016 kg water/kg dry air, the recycling ratio decreases in Scenario 4 compared to the other two recirculation scenarios. For Scenario 4, the heater duty is evaluated to be 489 kW. On the other hand, the total cost for heater duty is calculated to be $237,142/yr.

## Comparison of results

The results from the base-case model and those from the different scenarios are shown in Figure 6. The required heater duty decreases as the inlet temperature increases, due to the increase in recycling ratio. The regeneration heater duty remains stagnant throughout four scenarios, as presented in the bar chart. On the other hand, there is an obvious decreasing trend for dryer-air heater duty compared with the base-case model. As a result, the increased regeneration heater duty is not as significant as the decrease in dryer heaters duty. Moreover, as heater duty is depending on the electricity cost, both the total heater duty and annual heater cost have a similar trend. The optimum recycling ratios that yield the lowest total duty are summarized in Table 1.To conclude, we have presented four scenarios — three dryer-air recirculation scenarios and the current system design without recirculation (Scenario 1). Scenario 2 achieves the lowest heater duty, corresponding to an 18% reduction, with annual savings of $57,785/yr. However, the fines in the dryer exhaust air streams may clog the adsorber and lead to a reduction in energy efficiency in the long run. In Scenario 3, a constraint for the fines limit was applied and the reduction in annual heater duty and savings was determined to be 16% and $50,457/yr. The total elimination of fines in the exhaust air is proposed to enable an energy-efficient, total closed-loop drying system. In Scenario 4, the introduction of a wet scrubber to eliminate fines was analyzed. The annual heater-duty decreases was 2%, but it was not enough to compensate for the cost of water required for the wet scrubber unit. This is because, with the assumptions of L/G ratio of 0.94 kg/m^{3}, the water flowrate required is 3,457 kg/h, which leads to high effluent disposal charges. The freshwater tariff at the location of the plant (Johor, Malaysia) is $0.81 per ton of water. As a result, the cost of water required for the wet scrubber adds up to $22,402 annually, which is not enough to be covered by the savings from the decrease in heater duty ($6,087 per year).

*Edited by Scott Jenkins*

## References

1. Kaviani, A., A. Aslani, R. Zahedi, H. Ahmadi, M.R. Malekli, A new approach for energy optimization in dairy industry,

Clean. Eng. Technol.8, 100498, 2022.

2. De Vries, M., W. Al Zahra, A.P. Wouters, C.E. van Middelaar, S.J. Oosting, B. Tiesnamurti, T. V Vellinga, Entry Points for Reduction of Greenhouse Gas Emissions in Small-Scale Dairy Farms: Looking Beyond Milk Yield Increase,

Front. Sustain. Food Syst.3, 2019.

3. Todde, G. M. Caria, F. Gambella, A. Pazzona, Energy and carbon impact of precision livestock farming technologies implementation in the milk chain: from dairy farm to cheese factory,

Agriculture. 7, 79, 2017.

4. Ladha-Sabur, A.. S. Bakalis, P.J. Fryer, E. Lopez-Quiroga, Mapping energy consumption in food manufacturing,

Trends Food Sci. Technol.86, pp. 270–280, 2019.

5. Xu, T., J. Flapper, Reduce energy use and greenhouse gas emissions from global dairy processing facilities,

Energy Policy, 39, pp. 234–247, 2011.

6. Marcotte, M., S. Grabowski, Chapter 16: Minimizing energy consumption associated with drying, baking and evaporation, in: J. Klemeš, R. Smith, J.-K.B.T.-H. of W. and E.M. in F.P. Kim (Eds.),

Woodhead Publ. Ser. Food Sci. Technol. Nutr., Woodhead Publishing, pp. 481–522, 2008.

7. Tenaga Nasional Berhad, Pricing & Tariffs, 2014. (accessed August 15, 2022).

8. Mussatti, D.C., EPA Air Pollution Control Cost Manual, North Carolina, 2002.

9. Sparks, T., G. Chase, Section 6, Other Separation Processes and Equipment, in: T. Sparks, G.B.T.-F. and F.H. (Sixth E. Chase (Eds.), Butterworth-Heinemann, Oxford, U.K., pp. 361–382, 2016.

## Acknowlegement

The authors would like to thank Chee Hong Lee for his support in this work.**Au****thors**

Yick Eu Chew, is a chemical engineering graduate from the University of Nottingham Malaysia (B34, 43500 Semenyih, Selangor, Malaysia). He is now undertaking postgraduate study at Swinburne University of Technology Sarawak Campus. His research works focus on process design and optimization.

Wei Wen Wee is a chemical engineering graduate from the University of Nottingham Malaysia (). Her research interests are in process optimization, waste treatment and management.

Timothy Gordon Walmsley is a senior lecturer in process and energy engineering in the School of Engineering and assistant director of the Ahuora Centre for Smart Energy Systems at the University of Waikato in New Zealand. He is a chartered engineer and member of the Institution of Chemical Engineers (IChE).

Dominic C. Y. Foo is a professor of process design and integration at the University of Nottingham Malaysia ([email protected]). He is a past president of the Asia-Pacific Confederation of Chemical Engineering, a fellow of the Academy of Sciences Malaysia and the Institution of Chemical Engineers.