Federated Class-Incremental Learning with New-Class Augmented Self-Distillation