Compact cross-flow three-fluid heat exchanger with plain rectangular fins is experimentally investigated for both thermal and hydraulic efficacy under steady-state condition. The elaborative investigation is conducted for all the four viable flow arrangements of cross-flow configuration which comprises the effect of Colburn factor and effectiveness ratio of central and adjacent fluids, and friction factor with Reynolds number. It is learnt that thermal efficacy is sentient to the operating conditions, type of flow arrangement and variation in flow rate. The comparison of the efficacy between the plain duct (without fin core) and with fin heat exchanger core is presented to analyse the impact of fins in enthalpy exchange. The regression correlation is developed to estimate the thermal and hydraulic performance which are accurate within $\pm$16% and $\pm$5.8% respectively. The investigation is further assisted with the ANN modelling for the prediction of the thermo-hydraulic efficacy with two input (Reynolds number and flow arrangement type) and four output performance parameters (Colburn factor, friction factor and effectiveness ratios). The fraction of 77% and 23% of data is considered for training and testing the neural network respectively. It is learnt that ANN predictions are more promising and accurate than regression predictions and the magnitude of the value of an absolute fraction of variance equals unity.

实验研究了具有扁平矩形翅片的紧凑型横流三流体热交换器在稳态条件下的热效率和水力效率。对横流构型的所有四个可行的流动布置进行了详尽的研究，其中包括Colburn因子和中心流体与相邻流体的有效比以及具有Reynolds数的摩擦因子的影响。据了解，热效率与操作条件，流量布置类型和流量变化有关。比较了普通管（不带翅片芯）和翅片热交换器芯之间的功效，以分析翅片在焓交换中的影响。开发回归相关性以估算在以下范围内准确的热力和水力性能$\pm$16％和 $\pm$分别为5.8％。该研究进一步得到了ANN建模的协助，该模型具有两个输入（雷诺数和流量布置类型）和四个输出性能参数（柯尔本系数，摩擦系数和效率比）来预测热工液压效果。分别考虑了77％和23％的数据分别用于训练和测试神经网络。据了解，与回归预测相比，ANN预测更有希望且更准确，并且方差的绝对分数的值的大小等于1。