The e-balance system provides the means to match both time varying local renewable production and consumption, what allows a better utilisation of the electricity grid, while improving the distribution efficiency. Dynamic distribution grids show a major DER penetration and resilience is of utmost importance because reliability and flexibility are goals for deploying future proof smart grids.
End-users of the e-balance system in Bronsbergen (Netherlands) can set in the graphical user interface their preferences to maximise self-consumption or allow its remote control by the DSO to balance the neighbourhood power flows. In this way, end-users can decide if they participate in the neighbourhood energy balancing, based on incentives, or shift their flexible loads (smart-appliances) to periods with higher electric energy production.
Dynamic Voltage Control of Low Voltage Power Grids with Distributed Generation: Overview
The dynamic voltage control algorithm developed by INOV is applicable in Low Voltage (LV) power distribution grids to which various micro and mini producers are connected – designated in the literature as distributed generators (DG). The objective of the algorithm is to maximize the DG production, but it may also enforce fairness among DG producers, taking into account the contracts limitations with power grid operators or regulators and while keeping voltage levels within the standard operational limits in all coupling points in the power grid. This algorithm was submitted as a patent (INOV-INESC: Nunes, M., 2015. Dynamic Control Method of Power Injected into the Power Grid by Distributed Generators. PCT/2015000040, Patent submitted in July 2015.)
An e-balance LV power grid with distributed photovoltaic (PV) generation is depicted in Figure 1. Each PV generation site comprises one or more PV panels, an inverter and a CMU that interfaces with the e-balance system. The voltage control processing module runs the algorithm at the LVGMU, computing power generation setpoints (SP) based on voltage (V), current (C) and power factor (PF) values measured at the PV coupling points and issued by the CMUs. The calculated setpoints are then issued to the CMUs in order to dynamically adapt the PV inverter parameters.
Figure 1: LV power grid architecture and data flow between the voltage control processing module and the CMUs at DG sites.
Although existing algorithms in the state-of-the-art allow the calculation of the SP and to send it to the DG for voltage regulation in the distribution grid, they do it by successive approximations and usually they require some knowledge of the grid topology and generally achieve approximate solutions. In the present invention, by contrast, knowledge of the network topology is not required and an accurate solution is obtained by optimizing an objective function subject to a set of constraints to the currents and voltages at the output of DG coupling points. As the proposed method relies on measurements performed only at the feeder where the DGs are connected, the number of measurements and equipment involved is reduced, resulting in a lightweight and scalable computational implementation, wherein the implementing complexity grows linearly with the number of applicable feeders. As there is no interference between the various feeders, the implementation to a larger grid is merely a replication of the same method to multiple feeders. Another advantage of the locality of the proposed method is that we can select only the feeders that have DG installed, which reduces the number of network equipment installed to an absolute minimum.
The pseudocode of the voltage control algorithm is listed as follows:
FOR each DG in the feeder DO
- Measure the Voltage (V), Current (I) and Power Factor (PF) at the output of DG;
- Based on that measurements calculate the power that is produced by each DG
- Send a setpoint (SP) to each DG with a slight decrease of the power that is currently injected in the grid by that DG.
- Measure again the Voltage (V), Current (I) and Power Factor (PF) at the output of each DG;
- Based on the measurement made, calculate the sensitivity matrix (A) of that feeder, which shows how each DG affect the voltage in all the DGs.
- Based on the sensitivity A matrix and in different constrains (maximum voltage level, maximum DG power, minimum DG power), calculate the power injected by each DG that optimize the global DG power. (SPoptimal)
FOR each DG in the feeder DO
- IF the power currently injected by the DG in the grid is different of the calculated SPoptimal for that DG THEN
- Send setpoint with the calculated SPoptimal value to that DG.
Repeat the algorithm periodically.