The fast expansion of solar implementations has dramatically changed electricity generation, but also the attack surface in smart grids. Existing cyber security tools are not able to defend against (and thereby mitigate) sophisticated and adaptive attacks which arise in PV grids, such as false data injection, DoS (denial-of-service), an inverter hack. This paper presents an ICDF which leverages the emerging AI, edge computing and blockchain technologies to offer robust security protection in DSNs. The machine learning-based intrusion detection system, which recognises anomalous communication behavior, combined with blockchain-enabled data integrity modules that protect the transactions and control commands among DERs. A hybridised threat response model which utilises deep-reinforcement learning to adapt protection levels based on the current state of system risk indicators. Simulation results show that the ICDF is effective in improving detection accuracy, reducing false alarm rate and mitigating response time as opposed to conventional rule-based methods. This study lays the foundation for autonomous and adaptive cybersecurity architectures targeting next-generation solar energy systems, showcasing reliability, privacy, and operability sustainability in the context of fast-paced evolving energy internet.