Photovoltaic inverter agent

What is a smart solar PV inverter?

In , a hybrid architecture of both centralized and distributed control with the coordination of solar PVs and demand response is proposed. Smart solar PV inverters can provide fast and continuous active and reactive power control with low operational costs.

What is k th solar PV inverter?

The k th solar PV inverter has a maximum apparent power capability S k. The active power output is set at the available solar PV production potential. The reactive power output is limited by the inverter rating. If the inverter is not oversized, then it cannot provide reactive power compensation when p k t g = S k.

Why is intermittent solar photovoltaic (PV) voltage fluctuating?

Higher penetration of intermittent solar photovoltaic (PV) systems in the distribution grid results in frequent voltage fluctuations. The conventional voltage regulating devices operating on a slow-timescale need to be supplemented with the fast-operating smart inverters with adjustable reactive power setpoints.

How to determine the reactive power of smart inverters?

In the fast-timescale, a deep deterministic policy gradient (DDPG)-based algorithm is employed to determine the setpoints of the reactive power of smart inverters . We design a communication scheme for the DRL agents in two different timescales to exchange information and learn the control policy concurrently.

Why is solar PV a problem in a distribution network?

Specifically, high solar PV penetration in the distribution network creates serious operation challenges such as over-voltages and increased line losses . Moreover, the intermittent nature of solar energy can cause fast and large voltage fluctuations in the distribution grid .

How do PV inverters schedule OLTCs and capacitors?

In the fast timescale, PV inverters’ scheduling is modeled as Markov games and solved by a multi-agent soft actor-critic (MASAC) algorithm. In the slow timescale, OLTCs and capacitors are controlled by the soft actor-critic (SAC) algorithm. The agents in two different timescales are coordinated by the reward signal.