29 May - 28 July, 2000


E-mail conference on
"Small Scale Milk Collection and Processing
in Developing Countries"

Discussion Paper 2.1: Liquid Milk Processing







Pictures / Map

FAO-Dairy Page

FAO Home
Search FAO


Issues raised by conference moderators:

  1. “Solar/absorption refrigeration based on ammonia as refrigerant has rarely gone beyond the prototype testing”
  2. “UHT milk plants are too expensive for small-scale operations”
  3. “complete low-cost modules for liquid milk below 5,000 litres per day are not available”
  4. “If milk, that is consumed the same day, is boiled before consumption, like it is the case in many Asian countries, there is no need for pasteurisation” 



Discussion Paper 2.1: Small scale processing technologies: Liquid milk

Lusato R. Kurwijila, Department of Animal Science and Production, Sokoine University of Agriculture, Tanzania


Worldwide, liquid milk occupies a dominant position in milk processing, marketing and consumption. This is because liquid milk is a basic food in many societies where cattle and other milch animals form an important part of the agricultural production system. The form in which the milk is presented for sale, ranges from on-farm and traded raw milk sales in most of the poor developing countries to pasteurised or UHT milk in countries with developed dairy industries. If you review the history of dairying in most the European and North American countries, one sees the share of liquid milk products in the dairy industry changing over time in proportion to level of economic and technological development in their societies.   Consumers in richer societies demand and consume a more diversified mix of milk and milk products. Thirty years ago liquid / market milk constituted more than 70% of the quantity and values of marketed milk and milk products in these countries. To date this has dropped to no more than 50% as more and more milk is directed towards products such as fermented milks, cheese, butter, dried milk products etc.  In developing countries, depending on the level of development of the dairy industry, products range from a few indigenous dairy products and raw milk to pasteurised / UHT liquid milk and a small proportion of other  “luxury” products. Thus liquid milk is still the most important base for developing the dairy industry in developing countries.  This is due to population growth and rapid pace of urbanisation in most developing countries.  (See introductory paper of this conference).

The capacity to handle, process and market liquid milk must therefore be expanded to match the growing demand for processed liquid milk by the expanding (urban) populations in developing countries.  Given the nature of the dairy industry in most developing countries and the infrastructure, within which the industry must operate, this challenge is by no means an easy one. For the purposes of this e-conference, I intend to introduce the discussions on this subject by highlighting on the conditions which make “Small scale” processing, with all economic and technological considerations in mind, often the most viable if not the only option available for successful and profitable liquid milk processing in many situations with a poorly developed or emerging dairy industry.  I will then follow it up by discussing why, even as inevitable as small scale processing seems to be, its operations may not always be as easy as one may be tempted to think. I will conclude by highlighting the technical and economic challenges that we, in developing countries face, as we try to promote small scale processing.  This paper will dwell only on liquid milk processing. Other milk products will be discussed in discussion paper 2.2.


Conceptualisation of small scale processing

On the basis of the definition that has been given, small-scale processing units are those units handling more that 500 litres per day but less than 5000 litres per day. Very small scale or micro-dairies are those handling less than 500 litres per day. At the very bottom line we should not forget about “Household level technologies” which are important for household food security and traditional processing of liquid milk. These definitions are important, for what may be small scale in one country could be medium scale in another, depending on the level of development. Besides this has a big influence on choice of technology that may be available.

Developments in small scale processing

In the 60s, dairy processing in developing countries was largely based on the European model of centralised processing with plant capacities ranging from 30,000 to 60,000 litres per day.  These plants required electricity, the pasteurised milk was packed mostly in single use containers which had to be imported; the volumes of locally available milk was too small or too scattered so recombination of imported was used to fill the deficit as long as milk powder and butter oil donations were available. It soon became clear, the mostly state run large dairy plants were unsustainable. In the late 70s and early 80s commercial manufacturers of dairy plants came up with the concept of  “Mini dairy” which were essentially “block mounted”, turnkey projects of 500 litre per hour capacity (up to 5,000 per 8 hour shift).  Although in itself a very smart idea, the operations of these small plants were in most cases I have seen not very successful, mainly because of mismanagement of state run enterprises. In some cases, even these mini dairies were still too large for the small volumes of locally produced milk (often less than 2000 litres) that could be made available.  For some one in Europe or America it may sound inconceivable that local milk producers cannot produce enough milk to satisfy the requirements of such as a small plant! The reasons are varied, but some of the very common ones are:

i)   The plant is located in a town far away from the rural based milk producers where road networks are so bad that is uneconomical for the processor or milk producers to transport the milk to the plant, particularly during the wet season.

ii)  The milk marketing system is highly unregulated with informal milk traders offering a cheaper product to (mostly poor) consumers in the form of raw milk which they boil at home any way (In countries like Tanzania informal milk marketing of raw milk account for more than 90% of marketed).  The majority of the low income, in urban centres cannot afford the added cost of pasteurised milk especially when expensive packaging systems  costing more than 20-30% the retail value of raw milk are used.

iii) High cost of processing due to inadequate plant capacity utilisation, high electricity tariffs, taxes on processed milk versus no tax on raw milk. 

In spite of the big technological advances that have been made in the dairy industry and the subsequent drive for large and more efficient plants in developed countries the concept and need for small scale milk processing remains as valid and relevant today in most countries with a developing dairy industry as it was 30 years ago when the “mini-dairy” concept came to the fore (in developed countries to serve the needs of the on-farm producer-processors with spin-offs for developing countries). 


The subject of small scale milk processing was one of the topics covered at recent workshops held by FAO at Morogoro in Tanzania in 1995 and by FAO/IDF at Anand in India in 1997.   During the former it was noted that a stepwise scaling up of technologies is essential for any dairy or country wishing to grow a successful dairy industry. Unfortunately, simple solutions are either not always available or are ignored in favour of more sophisticated processing equipment which may not always be appropriate for specific situations.  For anyone trying to develop or set up a liquid milk-processing unit, one cannot avoid considering the following:

i) Milk Cooling and collection system.

Whether one is dealing with on-farm small scale processing or small scale milk processing based on collection of milk from many individual small scale milk producers one has to think of the most appropriate and efficient method of cooling and/or collecting the milk. In 1990 IDF published a Manual of Milk Collection in Warm Developing Countries. The time tested immersion coolers  (ICE BANKS) have now been largely replaced by direct expansion vats. The benefits of the immersion coolers in a developing dairy industry are obvious:

  • Milk is cooled in individual milk cans, avoiding the mixing of milk of inferior milk from one supplier into the rest of the milk.
  • In the event of electricity breakdown, the ice bank provides a buffer, which can still keep the milk for several more hours
  • Armed with a compressor, condenser, copper pipes for an evaporator and other essential standard components of a mechanical refrigeration system, any good refrigeration technician /engineer in a developing country can construct a well insulated ice bank using cement bricks or metal/metal boxes!

In spite of these advantages, people hesitate to use this simple technology in favour of an imported, stainless steel direct expansion vat. 

ii) Milk separation

Milk separation is a necessity for anyone trying to optimise product mix and profits by producing butterfat-standardised milk. The excess butterfat may be converted to more highly prized products such as cream, butter or ghee.  The centrifugal cream separator remains the equipment of choice. Manually or electric operated Models of 50 litres per hour to 500 litres per hour are available from various manufacturers. During the Anand IDF/FAO conference attempts to develop household level milk separator based on the common household blender/food processor machine was reported by Agrawala (1997). It is not clear whether or not commercial production of the described designs has come into production as of now.

iii)  Milk Homogenisation

Homogenisation of milk is standard operation for commercially processed milk. Homogenisation has the effect of breaking down the fat globules to sizes so small that the milk fat ca no longer forms a cream layer. While this improves the organoleptic and physical properties of pasteurised milk, the loss of ability to form a cream layer is sometimes viewed negatively by consumers in developing countries, who may erroneously think that such milk may be skimmed milk. This partly contributes to consumer preference for raw milk from which the consumer can still skim off the cream for other household uses. Homogenisers are essentially very expensive, the smallest unit may come in the range of 250 - 300 litres/ hour and they may cost as much as US $ 15,000 even as second hand equipment. Use of homogeniser may not be economical for plants of less than 5,000 litres per day. 

iv) Milk pasteurisation

Batch pasteurisation
Milk may be pasteurised by use of the age-old double-jacketed batch pasteurisers. All stainless steel vats of various sizes may be used.  Schulthess (1995) described a cheaper version (locally fabricated in Kenya) made of a stainless steel inner surface (which comes in contact with milk) covered in mild steel outer jacket and operating on low-pressure steam.  In-can pasteurisation of several hundred litres per day may be done in brick lined charcoal/firewood stoves (See FAO 1998, Schulthess in FAO, 1995).  Agrawala (1997) describes a mini-jacketed kettle with a double jacket filled with water that is heated, I guess, by use of an electric element. The milk is filled and heated to any desired temperature and holding time. The excess steam generated is whistled out from a vent, cautioning the operator on control of the pasteurisation temperature of the milk. Batch pasteurisers may easily be constructed by any mechanical engineering workshop with capability for stainless steel welding and metal sheet work facilities.

H.T.S.T. pasteurisation

Mini H.T.S.T pasteurisers are based on plate heat exchanger technology and much more sophisticated in their design and construction. The smallest units have capacities ranging from 500 litres per hour upwards. Major dairy equipment manufactures have each versions of their own. India is emerging as a strong source of cheaper but good quality versions as was evident at the dairy equipment exposition at Anand, 1997.

In-container pasteurisation

Milk may also be pasteurised after filling in the final container. This may be a plastic sachet or plastic bottle. The milk sachets or bottles are filled, then placed in a hot water bath or spray with or without agitation.  The batch method of pasteurisation are usually employed (63 - 65 0C for 30 minutes). See, for example, the in-pouch pasteurising system described in the poster paper on the Village Milk System that you will receive together with this paper.

V) Milk packaging

Pasteurised milk, after cooling, has to be packaged immediately. For batch pasteurisation installations of up to 2000 litres per day, manual plastic sachet sealing machines may be used at the rate of up to 300 sachets per hour. Sachets are usually pre-formed and sealed on one end. Semi-automatic versions with mechanised dosing of the milk are available from different manufacturers. Full automatic versions that form and fill various sizes of plastic sachets from a roll are available.  Unlike the pre-formed sachets, the automatic form and fill versions provide and opportunity for disinfecting the plastic film before milk is filled in. Batch pasteurised milk may also be filled in plastic or glass bottles or 3 - 5 litres plastic gallons and capped with aluminium foil, screw or cork cap.

vi) Bulk vending

After batch pasteurisation and cooling milk may be sold in bulk via bulk vending machines. This system is reported to be successfully in use in India. It cuts down the high cost of single use packaging.

vii) Milk storage and transportation

Pasteurised milk has to be kept cold throughout the marketing chain. Hence provision for refrigerated storage is mandatory. Transportation and distribution has to be done in insulated vehicles to avoid excessive temperature increases.

viii) Process utilities

Hot water/steam generation
As described above low-pressure steam boilers or electric hot water generators may be used to provide process heat. A combination of solar water panel for preheating the process water may reduce the cost of electricity in hot water/steam generation by as much as 30 - 40% but this is rarely used!

Mechanical refrigeration appears to be the only option available.  Solar/absorption refrigeration based on ammonia as refrigerant has rarely gone beyond the prototype testing. Am I right?

Renewable energy
Although small scale processors in developing countries often claim that the cost of electricity accounts for up to 30% of their operational costs, very little effort has been made to incorporate renewable energy sources of energy  (solar, Biogas) in small scale dairies. Much of what is reported in literature is a pilot/prototype test model. What is hindering commercial application?

ix) Sterilised/UHT milk

While its is possible to use batch retorts to produce sterilised milk in bottles. This technology, to my best knowledge rarely in use today. Reasons may be due to unfavourable economics of milk in-bottle milk sterilisation today and the extensive organoleptic and nutritive changes imparted on the milk. In can sterilisation is definitely uneconomical.  UHT milk plants are perhaps too expensive for small-scale operations. I do not know of any commercial UHT plant of less than 5,000 litres per day.


  • While small scale processing remains inevitable in developing countries, complete low-cost modules below 5,000 litres per day appear not to be available. One has to assemble various units from different manufactures to come up with a complete processing plant of ones choice/requirement. Even then most potential users of such equipment have no information on where to source for relevant equipment. 
  • While the basic technologies are well established at large scale operations, the challenge for developing countries is to miniaturise these technologies to a level that makes them accessible to even the smallest of dairy plants in remotest area of our developing world. 
  • The possibility to commercially incorporate renewable energy sources in small dairies has to be explored in commercial proportions beyond the research laboratories. 
  • A number of policy issues that seem to discourage development of milk processing in favour of raw milk marketing may also need to be addressed.


The author wishes to thank all scientists and dairy plant manufacturers who have worked and continue to work on small-scale milk processing technologies Worldwide. Availability of such information in various publications/brochures has enabled me to make this modest contribution to the conference. I hope that in the course of this conference we shall provide answers to some of the questions raised in this paper. Anybody aware of any technology /supplier of dairy equipment that may contribute towards technology transfer/development of small scale processing in developing countries may contribute that information to the moderator for incorporation in the Equipment database being developed as part of this conference. Thank you.


Agrawala, S.P. 1997. Technological aspects of dairy processing equipment:
Equipment design-Introduction. In: IDF, 1997. IDF workshop on Small-scale dairy processing and indigenous milk products. Proceedings, Anand, India, December 4-6 1997. Pg. 17 - 27

FAO, 1988 .Village milk processing. Animal Production and Health paper No. 69, FAO, Rome Italy.

FAO, 1997. FAO Workshop on Market oriented dairying: Role of Producers Organisations and NGOs, December 1- 3, Anand, India

IDF, 1990. Hand book on Milk collection in Developing Warm Countries. 

IDF Doc. No. 9002

IDF, 1997. IDF workshop on Small-scale dairy processing and indigenous milk products. Proceedings, Anand, India, December 4-6 1997.

Schulthess, W. 1995. Personal experiences with the promotion of milk processing in developing countries. In: FA0, 1995. Strategies for Market Orientation of Small scale Milk producers and their organisation.  Proceedings. 20- 23 March,  Morogoro, Tanzania, 1995.

Back to top

FAO, 2000