多普勒雷达资料的体积速度处理VVP（Volume Velocity Processing）风场反演方法可反演风场的3维结构，但由于算法的系数矩阵病态问题易导致反演风场产生误差。本文针对VVP算法中反演参数的性质，进行了简化算法的模拟检验和误差分析。选取量级最大的3个主要参量进行反演，引入随机的观测误差，通过改变模拟风速确定了反演算法的适用范围。对比结果发现，简化算法的反演结果对观测误差并不敏感，而且从低仰角到高仰角的均方根误差基本不变，当风速较大时，反演的精度会更准确。对0608“桑美”台风的风场反演表明，该算法较真实地反演出了台风中心及眼区外围的风场，并与Rankine台风模型相符。研究表明，简化VVP算法可清晰地揭示台风内部水平风场的3维结构，可以应用于台风等灾害性天气的风场反演与分析。

The Volume Velocity Processing (VVP) retrieval method can be used in the three dimensional structure analysis of wind field, but the accuracy of retrieval results is always affected by the ill-conditioned coefficient matrix. Error analysis and tests with synthetic data were performed by using a simplified VVP method to analyze the characteristics of fitted parameters and to confirm the fitted wind field in retrieval. In the simplified method, only three parameters with particularly large magnitude were selected. The coefficient matrix is a function of the analysis volume position. Hence, the retrieval errors and condition number varied accordingly. Selected partial parameters were substituted for full model retrieval and retrieval errors caused by neglected parameters were qualitatively analyzed to reduce solving difficulties. Observation and calculation errors are inevitable in retrieval. According to the spread of errors in the solving process, a demonstration through mathematical deduction was provided to explain that the large analysis volume can improve retrieval accuracy. The performance of the simplified VVP method was tested using synthetic data with random observation errors to find the fitted wind velocity. Wind model parameters with small magnitude or small position coefficient can be abandoned to simplify the wind field model. The results of the sensitive tests indicate that the robustness of the simplified VVP method is not highly sensitive to random errors. The accuracy of retrieval results at different elevations is almost similar and may improve when the wind field velocity is large. Considering the observation and calculation errors with normal or uniform distribution, we can reduce the influence caused by these errors in an analysis volume. These errors are spread and amplified in the retrieval process. Therefore, preventing the spread of these errors at the beginning of retrieval can improve the accuracy of retrieval results. For the wind field that is suited to use the simplified VVP method, the test result shows that this method can be used to retrieve the wind field with wind velocity above 20 m/s, which contains 1.0 m/s random observation error. In addition, the analysis volume size is 10° ^*20 Gates. The retrieval results of the 0608 "Saomai" typhoon shows that the wind field can be retrieved reasonably and is consistent with the Rankine typhoon model. The retrieved wind profiles indicate that the point of inflexions for wind velocity and direction can be used to confirm the eye region and typhoon height, which are suited to analyze the three dimensional wind field structures of a typhoon. Parameters with small magnitude are abandoned to simplify the wind field model. Thus, difficulties in solving are decreased and retrieval results become more robust. After confirming that the fitted wind field is suited to retrieval by using the simplified VVP method, the wind field of the 0608 "Saomai" typhoon is retrieved for a real test case. The three dimensional wind field structures of the 0608 "Saomai" typhoon can be obtained with this method and can be used for disastrous weather analysis.