This paper addresses the challenge of detecting cyber attacks, particularly false data injection attacks (FDIAs), in unbalanced power distribution grids with distributed energy resources (DERs). The proposed method is an unsupervised adversarial autoencoder (AAE) model that leverages long short-term memory (LSTM) to capture temporal dependencies in time-series measurements and generative adversarial networks (GANs) to reconstruct input data. The AAE model is designed to detect FDIAs without relying on labeled data, making it suitable for real-world applications. The effectiveness of the AAE model is evaluated on IEEE 13-bus and 123-bus systems using historical meteorological and load data. The results show that the AAE model outperforms other unsupervised learning methods in detecting FDIAs, demonstrating its superior performance in unbalanced power distribution grids. The model's ability to handle nonlinear characteristics and capture spatiotemporal correlations makes it robust against various types of cyber attacks.This paper addresses the challenge of detecting cyber attacks, particularly false data injection attacks (FDIAs), in unbalanced power distribution grids with distributed energy resources (DERs). The proposed method is an unsupervised adversarial autoencoder (AAE) model that leverages long short-term memory (LSTM) to capture temporal dependencies in time-series measurements and generative adversarial networks (GANs) to reconstruct input data. The AAE model is designed to detect FDIAs without relying on labeled data, making it suitable for real-world applications. The effectiveness of the AAE model is evaluated on IEEE 13-bus and 123-bus systems using historical meteorological and load data. The results show that the AAE model outperforms other unsupervised learning methods in detecting FDIAs, demonstrating its superior performance in unbalanced power distribution grids. The model's ability to handle nonlinear characteristics and capture spatiotemporal correlations makes it robust against various types of cyber attacks.