Optimal Sensor Placement For Vertical Quench Furnace Based On Chaos Shuffled Frog Leaping Algorithm

Abstract:Large-scale vertical quench furnace was a kind of critical equipment for thermal treatment of high-strength aluminum alloy components. Because the structure of large-scale vertical quenching furnace was complex and temperature distribution was distributed parameter, non-homogeneity, strong coupling and nonlinear characteristically, it’s very difficult to realize high precision and homogeneity of quench temperature control. It is necessary to configure the quenching furnace’s temperature sensor measuring point, in order to ensure the further improvement of temperature control accuracy of the large-scale vertical quench furnace.In this paper, we constructed a Fisher information matrix based on the parameters estimated minimum error criteria and the furnace control model. This Fisher information matrix has nothing to do with the sensor’s actual measured value but the sensor’s measurement position. We use chaos shuffled frog leaping algorithm (CSFLA) to find the best sensor measurement position to maximize the determinant of Fisher information matrix.Firstly, considering the shuffled frog leaping algorithm (SFLA) local search capability is not strong and easy to fall into local optimal solution, we carried out two aspects of the improvement of algorithm. The first one is adjust three poor individuals’position. This will increase the information exchange between individuals in the frog subgroup. The second one is embed chaos optimization algorithm into frog leaping algorithm, using the chaos optimization algorithm’s characteristic of randomness, ergodicity and sensitivity to the initial value, to improve the search ability of SFLA. Secondly, we proved the CSFLA’s execution efficiency and convergence speed are vastly superior to the SFLA by using simulation test results.Finally, analysis is also given in the configuration under the condition of different number of sensors and the corresponding optimization target function, using CSFLA carries on the optimization search. The simulation results proved that using CSFLA in quenching furnace sensor optimization configuration is feasible and effective. This article has18pairs of figure,9tables and61references.