Particle swarm and multi-objective optimization algorithm for heat sink configuration of a two-stage precision thermoelectric refrigeration

Two-stage cascade thermoelectric coolers (TECs) face performance limitations in practical applications due to inefficient heat sink configurations, which hinders the simultaneous maximization of cooling capacity and coefficient of performance. To solve this problem, this study establishes a one-dimensional thermoelectric energy model integrated with a heat sink thermal conductance. Combined with the Particle Swarm Optimization and the multi-objective optimization algorithm, the single-objective (maximum cooling capacity Q_c or maximum cooling coefficient of performance COP_c) and multi-objective (COP_c-Q_c trade-off) optimization problems are analyzed, respectively. The research results show that for single-objective optimization, it reveals the relationship between the optimal thermal conductance distribution ratio (Z_UA) and input current. In the optimization for maximizing Q_c, when the total thermal conductance (UA_hx) is 12 W/K, the results show that the optimal Z_UA is 0.685, corresponding to an optimal input current of 8.12A. And the Q_c reaches its maximum value of 13.89W and COP_c is 0.107. In the optimization of maximizing COP_c, the optimal Z_UA is 0.645, corresponding to the optimal input current of 2.52A. Meanwhile, COP_c reaches the maximum value of 0.314 and Q_c is 4.78W. Furthermore, high Z_UA could enhance COP_c but limit Q_c, while low Z_UA is beneficial to Q_c but weakens the enhancement effect of COP_c. The TOPSIS decision indicates that under the given operating conditions, the performance COP_c and Q_c of the two-stage cascade thermoelectric cooler are 0.219 and 9.36W respectively, and the Z_UA and input current are 0.66 and 4.49A, respectively. Finally, in practical applications, the optimal solution should be selected based on the actual demand. For instance, when prioritizing cooling capacity, a high Q_c solution should be chosen, while in energy-constrained scenarios, a high COP_c solution should be emphasized.