Study on thermal conductivity of improved soil under different freezing temperatures.

Based on the influence of moisture content, dry density and temperature (≦ 0°C) on the thermal conductivity of lime-modified red clay, the thermal conductivity was measured by transient hot wire method. A total of 125 data were obtained and the evolution law of thermal conductivity with influencing factors was analyzed. The fitting formula of thermal conductivity of lime-modified red clay and a variety of intelligent prediction models were established and compared with previous empirical formulas. The results show that the thermal conductivity of lime-modified red clay increases linearly with water content and dry density. The change of thermal conductivity with temperature is divided into three stages. In the first stage, the thermal conductivity increases slowly with the decrease of temperature in the temperature range of-2°Cto 0°C. In the second stage, in the temperature range of-5°Cto (-2)°C, the thermal conductivity increases rapidly with the decrease of temperature. In the third stage, in the range of-10°Cto (-5)°C, the thermal conductivity changes little with the decrease of temperature, and the fitting curve tends to be stable. The fitting formula model and various intelligent prediction models can realize the accurate prediction of the thermal conductivity of lime-improved soil. Using RMSE (Root Mean Square Error) and MAPE (Mean Absolute Percentage Error) to evaluate the model, it is found that the GBDT decision tree model has the best prediction effect, the RMSE value of the predicted value is 0.084, and the MAPE value is 4.1%. The previous empirical models have poor prediction effect on the thermal conductivity of improved red clay. The intelligent prediction models such as GBDT decision tree with strong universality and high prediction accuracy are recommended to predict the thermal conductivity of soil.

Double-loop control and intelligent parameter tuning for the temperature control system of a DBR semiconductor laser.

Aiming at lower startup power consumption, stronger thermal load adaptability, easier parameters adjustment, and higher parameter tuning efficiency for the temperature control system of a distributed Bragg reflector (DBR) semiconductor laser, this paper employs the double-loop control and intelligent parameter tuning methods. First, the thermal equivalent circuit model is established for the laser temperature control system, which has stronger thermal load adaptability than the traditional transfer function model. In order to improve the modeling speed and accuracy, a mean impact value (MIV) quantum particle swarm optimization (QPSO) intelligent algorithm is proposed to tune the model parameters. A double-loop temperature control system is set up on this basis. Then, the MIV-QPSO intelligent algorithm is used to tune the control parameters, which shortens the settling time, increases the tuning efficiency, and improves the temperature control effect. The feasibility and effectiveness of the proposed methods are verified through the MATLAB/Simulink simulation of the laser temperature control process.