Temperature Field Reconstruction And Electrical Compensation For Phased Array Antenna

During the service of the phased array antenna,the influence of the external ambient temperature and the excessive power consumption of the internal T/R components will cause the uneven temperature of the antenna array,which will affect the output current of the antenna feed network,resulting in electrical performance decline.In order to solve the above problems,this paper proposes an electrical performance compensation method based on the reconstruction of the temperature field with incomplete information.The main contents of the paper include:1.A method for dynamic reconstruction of antenna temperature field under incomplete information is proposed.This method realizes the real-time calculation of the temperature field and achieves the purpose of predicting the temperature field one moment in advance.In this chapter,the intrinsic orthogonal basis decomposition(POD)method is used to map the high-order model of the antenna into a low-order model about the coefficients of the POD mode,which is used to quickly calculate the temperature field of the antenna.However,because the antenna electrical performance compensation needs to be fed back to the background for adjustment one time in advance,it is proposed to use recursive least squares to perform time series calculation on the POD mode coefficients to achieve the purpose of predicting the POD mode coefficients one time in advance In this way,the temperature field can be calculated one time in advance,and the dynamic reconstruction of the temperature field can be realized,so as to pave the way for the subsequent antenna electrical performance compensation.2.A temperature field calculation method based on transfer learning is proposed to solve the problem of inaccurate temperature field reconstruction caused by model errors and measurement errors in practical applications.This method adds an error correction model based on the calculation of the temperature field with an intrinsic orthogonal basis.The error correction model is obtained through transfer learning.The specific steps are:First carry out a priori simulation of the model through the software to obtain a large amount of simulation data(source data);then experiment with the sample to obtain a small amount of experimental data(target data),and then use affine transformation and adaptive weighting algorithm to make the source data The target data is learned to obtain a large amount of pseudo-simulation data.Finally,the extreme learning machine algorithm is used to construct an error correction model to correct the temperature field reconstruction result,reduce the reconstruction error,and make the calculation result closer to the actual situation.3.A method to compensate the uneven temperature of the antenna is proposed.This method combines vector fitting and extreme learning machine algorithms to obtain the amplitude and phase of the excitation current at any temperature,and then to compensate for the electrical performance of the antenna.First,the temperature drift characteristic experiment of the power amplifier device is performed to obtain the relationship between the scattering parameter of the power amplifier and the frequency change at discrete temperature;according to the idea of black box modeling,only the input and output are considered,and the S21 is constructed by vector fitting and extreme learning machine.Mechanism functions of amplitude and phase with respect to temperature.It can realize the prediction of the amplitude and phase of the S21 parameter of the excitation current at any temperature.Finally,the method is applied to a 64-element antenna array,and the electrical performance after temperature is compensated according to the calculation results.Comparing the electrical properties,the results show that this method has better compensation effect.