The short-circuit current of renewable energy resources presents non power frequency characteristics with limited amplitude, resulting in performance degradation of differential protection. Based on the analysis of the fault characteristics of renewable energy resources and traditional synchronous generators, this paper puts forward a high-speed protection of a Teed line connected to renewable energy power plants. It is based on the similarity of time-frequency characteristics. First, compressed sensing is used to compress the high-frequency sampling signal, transmit low-frequency compressed data and reduce the digital quantity of protection communication. Then wavelet transform is used to extract the time-frequency characteristics of the fault current, and principal component analysis (PCA) is used to eliminate the redundant information of the time-frequency characteristics to obtain the time-frequency characteristic matrix. Finally, Canberra distance is used to measure the similarity of time-frequency characteristic matrices, and a high-speed protection principle based on time-frequency characteristic similarity is proposed. The proposed protection can reliably identify internal and external faults within 5 ms, and has strong ability to withstand fault resistance and system noise. It is also applicable to the scenario of weak output of renewable energy resources. The effectiveness of the proposed protection is verified by a ‘hardware in the loop’ experiment and field recorded data.

This work is supported by the National Natural Science Foundation of China (No. 51725702 and No. 52061635102).