With the rapid development of clean energy

photovoltaic (PV) power plants have gained increasing attention. However

the inherent intermittency of PV generation requires the integration of energy storage system (ESS) to smooth power output and enhance grid stability. Coordinating multiple PV–ESS plants is essential to maintain system reliability

balance stochastic renewable outputs with real-time load demands

and leverage time-varying electricity prices for economic benefits. In this paper

a learning-based joint bidding framework is proposed to maximise the aggregated profit of PV–ESS plants. First

a multi-PV–ESS model is built to emulate the coordinated operation of PV and ESS units in the power grid

aiming to maximise PV power revenues while considering penalty payments for power shortages

real-time load demands and dynamic power prices. Then

the joint bidding operations of PV–ESS plants are formulated as a Markov decision process

and a deep reinforcement learning algorithm is developed to learn optimal bidding strategies that adapt to load dynamics and price fluctuations. Extensive case studies on distribution systems of different scales

including the IEEE 33-bus and 69-bus systems

are conducted to demonstrate the effectiveness of the proposed method.