Energy storage technologies, coupled with advanced power electronics systems, are key to maintaining electrical system reliability when faced with increasing installation of renewable energy resources. Spain has set a target of more than 40% of its electricity generation to come from renewable energy by 2020. Endesa, the country’s largest utility, is a leader of the Storage Technologies of Reliable Energy (STORE) project, which consists of an intensive test matrix on a number of European island networks.
Located off the northwest coast of mainland Africa, the Canary Islands — comprised of seven isles — was selected as a site for the STORE project. Traditionally, the stability of the Canary Islands’ network has been affected by unmanaged technologies, such as renewables, and the cost of conventional energy generation is higher than on the mainland.
The STORE project sets out to present a greater understanding of energy storage integration under the conditions inherent in isolated electrical systems. Furthermore, it will provide an analytical basis intended to create greater interest in the future development of storage technologies.
Three energy storage technologies have been installed in Endesa’s generation plants on the Canary Islands:
- In La Aldea de San Nicolás on the island of Gran Canaria, a plant contains lithium-ion electrochemical storage batteries (1 MW/3 MWh), supplied by Saft and installed to provide in-service capacity as a generating unit. This allows demand management by providing inertia and active power to the system, voltage regulation and participation in the secondary regulation provided by the automatic generation control system.
- In Alajeró on the island of La Gomera, inertial storage equipment has been installed in the form of a flywheel (0.5 MW/18 MW), supplied by ABB. This provides inertia and active power for the primary regulation, in addition to continuously stabilizing frequency on the island.
- On the island of La Palma, a diesel power station is equipped with ultracapacitor technology (4 MW/20 MW), supplied by Ingeteam and Endesa, to support the frequency stability of the system. The technology is also installed to validate its ability to avoid loss of electrical supply from unforeseen breakdowns, thereby providing the system with improved robustness and quality of supply.
The equipment installed on La Gomera and La Palma uses energy storage technologies with very fast response; therefore, it is suitable for avoiding unexpected events in small power systems subject to partial loss of power or a total blackout. Meanwhile, the Gran Canaria plant has a larger storage capacity and can function as a group of manageable generation, so plant loading can be scheduled on a daily basis.
All three energy storage technologies can be controlled locally and remotely for the command and supervision of all equipment, recording of trends and communication with load dispatching and maintenance.
The lithium-ion electrochemical storage system consists of three container-type lithium-ion batteries, each with 17 racks (756 Vcc, 90 A maximum). They are mounted in an independent container with the power electronics — a direct current-direct current (DC-DC) converter and a direct current-alternating current (DC-AC) inverter — and a transformer to inject and draw power from the grid.
The system can operate in either automatic mode, following the selected control strategy, or manual mode. The system can select between an active power control (using a set point or a set point with frequency tracking) or reactive power control (using a set point or set point tracking with voltage or cosine phi), or a mixture. In automatic mode, it is possible to select peak-shaving or load-leveling functions.
The peak-shaving function is intended to limit the peak power on the distribution line and to prevent the entry of peaking generation units. Some strategies require an estimation of the demand, which is calculated by a method of the current daily average. These strategies employ a user-configurable period for the battery charge and discharge periods, managed by a selected algorithm:
- Low-capacity final. The aim of this algorithm is to allow a full discharge in each period so, at the end of the period, the battery is discharged to the minimum value specified by the manufacturer. A discharge limit, called the rolling threshold, is calculated above this minimum value to limit battery discharge. The initial threshold-based limit is determined by iteration until a certain performance is achieved, configurable in the SCADA system.
- Final low capacity with dynamic update. This algorithm has the same objective as the low-capacity final algorithm, whereby the demand estimation is performed at the beginning of each iteration. This way provides greater accuracy in energy management and the optimization of the battery discharge.
- Power limit. This algorithm establishes a demand limit above which it begins to discharge the battery. This control strategy has a high level of uncertainty as it is not designed to fully discharge the battery, so it is dependent on the knowledge of the operator to set this limit.
- Variable threshold. This algorithm uses the low-capacity final and power limit methods to determine which one is most applicable. The algorithm calculates the energy used by each method and calculates the percentage of the configured variable threshold. It dynamically completes the iteration necessary to obtain the threshold discharge for optimal performance.
The load-leveling function uses a calendar to record the charging and discharging characteristics of the batteries every 15-minute period in a week. For each period, it is then possible to configure a set point for an active or reactive power factor.
The flywheel energy storage system combines a low-speed flywheel (3600 rpm at 100% capacity) with two solid-state insulated-gate bipolar transistor (IGBT)-based inverters, resulting in a high-performance grid-stabilizing device. The system is able to sink or source energy up to its nominal power rating, and is capable of responding to power system changes in approximately 5 msec. The system can operate in manual mode with active power control (set point, proportional response or inertial response) or reactive power control (set point or tracking with voltage), or a combination of modes.
The flywheel’s main purpose is to stabilize power systems to increase renewable energy penetration and quality of supply:
- Flywheel charge control. The charge level at which the flywheel normally operates can be set between full and empty from the control system. The normal charge level is set to ensure sufficient energy to carry out the frequency support for smoothing the power system load fluctuations, plus a margin to cover the loss of generation plant such as a diesel generator.
- Step-load response. The flywheel has a very fast response time for both charging and discharging. In case of loss of plant within a power system (for example, a generator has tripped off-line), a step in the system load usually appears that results in a large frequency deviation. The flywheel is capable of reacting to this step change in load by discharging up to its nominal power rating within milliseconds.
The ultracapacitor system (STORE-U) consists of two equal units of power converter (2000 kW for 5 sec) and its associated ultracapacity. The power converter consists of an inverter AC-DC. This is required for the static synchronous compensator (STATCOM), the energy storage system (ESS) module and the converter (DC-DC), which only starts in ESS mode.
An operator configures the ultracapacitor system to activate at a certain frequency, including its derivative (δ) for a specific amount of time. Ultracapacitor operations will occur if the frequency drops below the defined limit (FLIMIT) with a δ(DFLIMIT). System power injection starts if these two conditions occur simultaneously. Both units can operate independently in either the STATCOM (voltage control/voltage power) mode or the ESS (primary response/active power) mode.
On-Site Test Results
Endesa has conducted different tests to validate the grid response to the storage facilities installed. These tests depend on the type of storage equipment (namely, batteries, flywheels or ultracapacitors) and the anticipated response to load leveling, the ancillary services of inertia.
Lithium-ion batteries were used to test the capability to provide load-shifting services, and voltage and frequency regulation. The algorithms were tested in real-time conditions by recording the electrical demand on the Gran Canaria-Anzofe overhead power line at the source substation. Based on the demand information, the battery control system calculates the output required in real time.
The La Gomera island grid is defined by low inertia and weak short-circuit levels. The flywheel allows the supply of fast energy to the system to increase the frequency response of the generators. This equipment was able to continuously filter the frequency variations in such a way the maximum variation in frequency was decreased.
Three tests were performed to determine the benefits of the flywheel in terms of the frequency response following the loss of a single generator supplying 500 kW of the island’s 8000-kW peak demand. For each test, the value of the proportional gain (Kp) response was changed from Kp = 0 kW/Hz to Kp = 1000 kW/Hz. There was a significant improvement in the behavior of the grid frequency as the value of Kp was increased. However, the final value of Kp is a compromise between the improvement in grid frequency and system stability. Additionally, the tests confirmed the need to check the behavior of the energy stored in the flywheel. When it reaches the lower limit, it activates a soft disconnection by means of a tail discharge, reducing the power exponentially because a high value of Kp could provoke a sudden disconnection when the power injection into the grid system is completely dissipated.
La Palma island grid has similar electrical characteristics as the island of La Gomera, namely, a low short-circuit level, although the peak demand on the grid is four times greater.
The purpose of the ultracapacitors is the injection of a high-power output during transient events until the load imbalance between the generation and the demand is rectified. The energy supplied by the ultracapacitors is designed to restore the drop in frequency faster than the generators, minimizing the possibility of load shedding. The tests undertaken demonstrated the effect of increasing the Kp of the ultracapacitors from Kp = 0 pu MW/pu Hz to Kp = 100 pu MW/pu Hz, which resulted in faster grid frequency recovery.
As a result of Endesa’s project in the Canary Islands, the utility has validated several of the different storage solutions present in the market designed to provide a more efficient and reliable electrical system in isolated grids installed on islands. The technologies that have been tested in an extensive experimental matrix proved their capability to operate when integrated into the generation portfolio of the system.
Thanks to the STORE project, the application of these technologies has been demonstrated to provide additional generation services, improving the quality, efficiency and safety of the grid, while increasing the incorporation of new sources of unmanageable and distributed generation.
Pablo Fontela ([email protected]) holds a master’s degree in thermal and fluids engineering from the Carlos III University, Madrid, Spain. As a researcher and head of hydrogen hybrid systems in BESEL, a hydrogen technologies research company, he was involved in projects related to hybrid vehicle power trains using fuel cells. Since 2007, he has worked in Endesa Generation’s R&D unit, leading national and international R&D projects in the energy sector.
Alberto Barrado ([email protected]) was awarded a MSEE degree from the Universidad Pontificia de Comillas (UPCO), Madrid, Spain. As a researcher at the Institute for Research in Technology of UPCO, he was involved in research projects related to power system stability (weak island systems). In 2000, he joined Endesa’s grid studies department, working on projects linked to increasing generation on islands.
Jorge Martinez ([email protected]) received a master’s degree in industrial engineering (with an electronic specialty) from the Universidad Pontificia de Comillas and has worked on the automation of substations, diesel and hydraulic plants. Since 2009, he has worked in Endesa’s instrumentation and control department, working on automation projects in Endesa’s generation plants.
Juan Carlos Ballesteros ([email protected]) received his bachelor’s and master’s degrees in mining and a Ph.D. degree from the Universidad Politécnica of Madrid, Spain. Since 2006, Ballesteros has been head of the R&D subdirection in Endesa, providing technical leadership to the Endesa Generación R&D team on a wide variety of projects.