Tubular Heat Exchanger Fouling by Milk During Ultra High Temperature Processing K. R. SWARTZEL ABSTRACT The effect of milk fouling deposits on the heat transferrate of a tubular heat exchanger was investigated for varying process heat treatments. Product input and discharge temperatures were main- tained constant while steam temperatureincreased as deposits formed. Processing conditions included product entrance tem- peraturesof 347.6 and 366.4”K for heater exit temperature of 410.7’K and product velocity of 1.58 m/s; 363.9 and 366.7 for 410.7 and 2.70; 364.6 for 410.7 and 3.25; 370.6 and 357.8 for 427.4 and 1.58; 378.5 for 427.4 and 2.70; and 379.5 and 376.3 for 427.4 and 3.25. Predictive expressions representing product deposition were determined for process variables, heating curve variables, and equivalent times and temperatures representing the thermal treatment. Product constituent losses during the fouling process andaspects of fouling kineticswereexamined. INTRODUCTION MILK DEPOSITS on heated surfaces have long been recog- nized as a normal byproduct of milk processing. When product deposition (fouling) occurs on heat exchanger surfaces three problems occur. First, processing rates may be reduced. Deposit may restrict the flow of product or deposit may decrease the heat transfer rate to the extent that product flow rate must be reduced if processing tem- perature is to be maintained. Second, product quality is often impaired. And third, removal of the deposited ma- terials is often a difficult part of the cleaning operation. This problem can be as serious as having to shut down to clean after only 2 hr of operation or having to flush with acid, after 2 hr to extend the run time (Casey and Swientek, 1982). The formation of the deposits result in increased steam consumption and appearance of solid particles in the final product (Jelen, 198’1). To minimize formation of the deposits heating media temperature is usually limited to a few degrees above product temperature (Farrall, 1963). Since heat exchange area must be increased if heating media temperature is reduced, this adds to the size and expense of the heating equipment. An additional technique known in industry to extend run time calls for lengthtening the I preheat by adding a holding section to the preheater. Some processing plants have found it more economical to add a second sterilizer in parallel to their existing one than to shut processing completely down for cleaning (Jelen, 1981). In this arrangement one sterilizer is cleaned while the other continues the processing. These approaches are I expensive practical answers to a difficult industry problem. In an effort to understand the basic mechanism inves- tigators have given attention to isolating product and process variables which are contributing factors to the deposition problem. Product variables that have been re- lated are content (Gynning et al., 1958; Burton, 1961, 1968), pH (Gynning et al., 1958; Burton, 196.5, 1968; Lalande and Corrieu, 1981; Claesson et al., 1974), seasonal Author Swartzel is affiliated with the Depts. of Food Science and Biological & Agricultural Engineering, North Carolina State Univ., I Raleigh, NC 27650. . variations (Burton, 1967), “age” of the milk (Burton, 1968), and ammonia concentration (Lalande and Corrieu, 198 1). Compositional variations between milk-based prod- ucts (flavored milks, ice cream mix, etc.) have been dis- cussed because increased viscosities lead to an increased probability of non-uniform heating, low heat transfer rates and an increased product depositional rate (Jelen, 1981). Single process variables have been associated with foul- ing rates for a variety of products. Among the most com- mon process variables are product velocity (Bunchero and Gordon, 1960; Perry, 1963; Gynning et al., 1958; Thonie, 1958; Kern, 1966, Morgan et al., 1959; Gonionskiy et al., 1970; Crozier, 1982), exposure time (Burton, 1961), wall temperature (Morgan et al., 1959; Gonionskiy et al., 1970; Burton, 1968; Lalande and Corrieu, 1981), process- ing temperature (Thonie, 1958; Burton, 1961) and bulk fluid temperature (Morgan et al., 1959; Taborek et al., 1972; Kern, 1966). Some researchers have combined process variables such as heating surface - product temperature difference (Thonie, 1959) and heating media - product temperature difference (Gynning et al., 1958; Gonionskiy et al., 1970; Kern, 1966). Others have made use of nondimensional groups such as the Reynolds number (Lund and Bixby, 1975). Dimensionless functions of average deposition temperature, hot tempera- ture difference, average milk velocity and temperature of pasteurization have also been examined (Lalande and Cor- rieu, 1981). The heat exchanger wall surface has been shown to affect the deposition rate as noted by Kern (1966), Morgan et al. (1959), Taborek et al. (1972), and Ling and Lund (1978b). Pretreatment has been shown to affect the rate of foul- ing. Bell and Sanders (1944), Morgan et al. (1959), Jelen (198 1) and Ball (1977) have all noted that preheating prod- uct reduces the rate of increased resistance in the heat ex- changer. Lund and Bixby (1975) demonstrated that by prefouling a heat exchanger with one species of foul deposit a reduction of deposit by other sources could be achieved. With a variety of product and process variables shown to contribute to product deposition, development of basic mechanisms has been a major unresolved problem for researchers. Many of the variables or functions of variables are interrelated. Attempts to reduce contributing variables in statistical models have tended to reduce the usefulness of the model (Lalande and Corrieu, 1981). General con- siderations for a possible mechanism were proposed by Bur- ton (1968). He suggested that two separate processes occur. First, the temperature effect produces a condition where milk solids are no longer in true solution. At this state they either absorb to a surface or aggregate. He proposed that the presence of an available surface would determine if the solids form deposit in the heat exchanger or sediment in the processed product. However, Burton (1968) noted that the literature was conflicting as to the relation of the degree of product deposition during processing to sedi- ment formation during storage. Having a site available for deposition may depend on many of the variables listed above. Crozier (1982) noted that if the stress at the wall exceeded the bond strength Volume 48 f1983)-JOURNAL OF FOOD SCIENCE-1507