【正文】 g heat exchanger length and wind speed. For purposes of paring trends, a correlation for recirculation remended by the CTI [58CT1, 77CT1] is also shown in figure . It should be noted that this correlation is 內(nèi)蒙古工業(yè)大學本科畢業(yè)論文 15 applicable to induced draft cooling towers although the authors do state that they expect the recirculation of a forced draft system to be double the value of the correlation shown. Duvenhage et al. [96DU2] show that the addition of a solid walkway along the periphery of the aircooled heat exchanger (at the fan platform elevation) tends to improve the mean flow rate through the fans (see figure ). Figure : Walkway effect According to the abovementioned findings the reduction of performance in a long forced draft aircooled heat exchanger may generally be ascribed primarily to a reduction in air flow through the fans along the windward side of the bank when crosswinds prevail as shown in figure (a), and to recirculation of hot plume air as shown in figure (b) when the winds blow in the direction of the major axis of the heat exchanger. Fahlsing [95FAll observed reverse rotation of out of service fans on the windward side of a large aircooled condenser when crosswinds prevailed. Figure : Flow patterns reducing performance. RECIRCULATION AND INTERFERENCE As in the case of banks of aircooled heat exchangers， recirculation of hot， moist plume air is known reduce the performance of rows of cooling tower units or cells [77CI1，內(nèi)蒙古工業(yè)大學本科畢業(yè)論文 16 88BS1]. Furthermore, when several banks of aircooled heat exchangers or rows of cooling tower cell are located next to each other， the plume of one bank or row may be drawn into an adjacent one． This phenomenon is referred to as interference. Ribier [88RI1] conducted recirculation tests on models of induced draft cooling towers cells similar to the types shown in figure , but without a diffuser. Initial tests were conducted on a row consisting of three cells with fills in counterflow and crossflow respectively. The results of these tests are shown respectively in figures (a) and (b) as a function of different wind directions and ratios of wind speed(measured 10 m aboveground level)to plume exhaust speed Vw/Vp． The percentage recirculation is less for the counterflow arrangement than for the crossflow arrangement． When a diffuser was added to the counterflow unit a reduction in recirculation was observed． Figure : Recirculation in threecell counterflow and crossflow cooling tower. A further set of tests was conducted by Ribier in which two rows of counterflow cooling towers each consisting of three ceils were first arranged end to end (six ceils) and then systematically spaced one, two and three cells apart. Of these tests the continuous row of six cells experienced most recirculation with results as shown in figure . Recirculation appears to be a maximum at Vw/Vp ＝ . 內(nèi)蒙古工業(yè)大學本科畢業(yè)論文 17 Figure : Recirculation in sixcell cooling tower. Figure : Recirculation in a counterflow cooling tower consisting of two threecell rows, two cells apart. As shown in figure recirculation is considerably reduced when the two rows of three cells each are separated by a distance of two cells. Further separation does not reduce recirculation much. By placing two rows of three cells each side by side, recirculation is relatively high as shown in figure . 內(nèi)蒙古工業(yè)大學本科畢業(yè)論文 18 Figure : Recirculation in cooling tower consisting of two rows of three cells located side by side. If the two rows of three cells are separated by one cell width only a relatively small reduction in maximum recirculation is experienced as is shown in figure . Based on these results it may be concluded that a row of induced draft cooling tower cells should be arranged inline with the prevailing wind direction. A high air outlet velocity and the addition of a diffuser will also tend to reduce recirculation. Bender et al. [97BE1] numerically analyzed the air flow into a counterflow induced draft cooling tower consisting of two adjacent cells of the type shown in figure (b) with a view to reducing or eliminating ice formation at the tower inlet during windy periods in winter. The dimensions of the tower they studied were m (width), m (length) and m (height) with an intake height of m. The stack or diffuser diameter was m and its height was m. Ice buildup tends to be most prevelant at the windward facing intake where the entering air flow rate is higher than on the leeward intake. By placing a 10 percent porous wall m in height, m in front of the cooling tower inlet, the air flow entering on either intake was found to be essentially the same. Tesche [96TEl] conducted model tests to determine the effect of recirculation and interference on the performance of rows of induced draft hybrid cooling tower cells (similar to the unit or cell shown in figure ). His results are applicable in cases where the wind velocity distribution is given by Vw/Vwr = (Z/Zr). It is found that the recirculation of individual cells in a row consisting of twelve ceils varies as shown in figure . All wind speeds are specified at 10 m above ground level. 內(nèi)蒙古工業(yè)大學本科畢業(yè)論文 19 Figure : Recirculation in cooling tower consisting of two rows of three cells separated by one cell width. Figure : Re, circulation in a row consisting of twelve hybrid cooling tower cells. The lowest recirculation is observed when the wind blows in the direction of the major axis of the cell row. The influence of the number of cell