Implementation and Performance Evaluation of a Solar-Powered Closed-Loop Recirculating Hydro-Gravitational Energy Storage System with Reinforcement Learning Control

Abstract

This repository contains dataset files, mathematical model implementations, and code artifacts for a Solar-Powered Closed-Loop Recirculating Water Gravity Battery System. The system integrates a differentiable digital twin framework for microgrid energy management, deep reinforcement learning algorithms (including Proximal Policy Optimization) for dynamic dispatch and battery energy storage system scheduling, and integrated cyber-physical security monitoring protocols for smart grid optimization. Key Features: Hydro-Gravitational Energy Storage Modeling: Formulations and numerical simulations for closed-loop, solar-rechargeable water gravity battery systems. Reinforcement Learning Dispatch Control: Implementation of PPO-based adaptive control and Mixed-Integer Linear Programming (MILP) optimization models for dynamic load management and energy storage scheduling. Cyber-Physical Power System Security: Algorithms for false data injection attack (FDIA) detection, Extended Kalman Filtering (EKF), and Cumulative Sum (CUSUM) verification protocols for resilient microgrid operation. Performance Analysis: Experimental datasets, simulation scripts, and output logs evaluating hydraulic efficiency, non-dimensional dynamics, and load balancing across operational constraints