Utility Solar and BESS Plant Controller

Video

Acknowledgements

Project Advisor: Robin Steele

Primary Client Contact: Clifton Oertli

Power Systems Advisor: Carson Bates

SCADA Advisor: Keith Gubat 

Advisor: Kris Moen

Elevator Pitch

The purpose of this project was to design a power plant controller (PPC) that integrates a battery energy storage system (BESS) with a photovoltaic (PV) power plant. A major issue with renewable energy is the difficulty in providing consistent power while balancing load deviations throughout the day. Battery storage is a method to mitigate the duck curve phenomenon in which the energy produced via solar generation peaks when grid demand is the lowest. As seen in the graph above, energy consumption drops at about 8 am and spikes around 6 pm. However, solar energy has peak production during the hours of 10 am to 2 pm. Our controller facilitates the charge cycle of a power plant’s batteries to provide the amounts of active and reactive power at the plant’s point of interconnection (POI) to the grid. Ultimately, the PPC assists grid voltage and frequency stability utilizing synthetic inertia and a battery energy storage system to ease the societal transition from conventional fossil fuel generators to highly variable renewable energy generation methods.

Our PPC has been implemented onto an industry-standard programmable logic controller (PLC). The PPC program has been designed to be scalable to PV-BESS systems of varying sizes, as the program is written in per unit.

The input parameters include a meter that monitors Point-of-Interconnection system variables such as voltage and frequency, while grid operators have the ability to dispatch complex power commands. These values were incorporated into a closed-loop feedback

Design Approach

The team was provided with a list of specifications from the client outlined in Table 1 below.

The team initially created a one-line diagram of the photovoltaic (PV) battery energy storage system (BESS) plant. The team’s one line diagram can be seen below. The team initially chose a DC-coupled (direct current) design, meaning the power would be converted to AC (alternating current) at the point of interconnection (POI) to the grid.  After consulting an industry expert, the team changed the design to be an AC-coupled design, meaning that each DC device in the system (i.e. each battery and PV string) has an inverter to convert the DC into AC, so that they can all be joined together at a common AC bus.  AC-coupled systems are the industry standard and utilizing inverters at each device allows us to send individual active and reactive power commands to each device.  This is crucial to the design of our system, as batteries may have different charges at different times and the amount of power generated by the PV strings varies without the day. 

Next, the team developed the controller logic.  The team based the design of the controller off of the Western Electricity Coordinating Council (WECC) Renewable Energy Plant Control Model A (REPC_A), which is a model for plant-level active and reactive power control.  The objective of power plants is to provide enough power to the grid to maintain the grid voltage and frequency.  Our point of interconnection (POI) agreement as specified by our client is 13.2kV @ 60Hz.  A drop in voltage on the grid is associated with a lack of reactive power generation, whereas a rise of voltage on the grid is associated with too much reactive power being on the grid.  Similarly, an under-frequency condition on the grid is associated with a lack of active power generation (such as what happened in Texas in February).  Using the REPC_A model, we are able to incorporate logic into our controller to generate active and reactive power commands for our plant to respond to different voltage and frequency conditions, with the goal of maintaining the values as set forth by the POI agreement.  

This graph shows the grid frequency plotted over time during the Texas blackout. As seen in this timeline of frequency levels through the event, Non-winterized natural gas equipment and locked up wind turbines forced the Energy Reliability Council of Texas (ERCOT) to shed thousands of megawatts in load in order to prevent the frequency from dipping under 59.4Hz for over 9 minutes, which would have caused irreparable damage and a total blackout.

The team then implemented the designed logic onto a programmable logic controller (PLC), which are the industry standard for creating controllers.  Once the logic was created and implemented onto the controller, the team conducted a factory acceptance test (FAT) to test a variety of cases that could happen.  As the team went through the cases, modifications had to be made to the logic to get the system to respond as intended.  Similarly, trial and error was needed to tune the PID controllers.  Naturally, there were some bugs in our controller code that had to be fixed.

Design Solution

First, the team designed a 15MW AC-coupled photovoltaic (PV) plant with a battery energy storage system (BESS). The client specified the plant to consist of five 3MW PV strings, each with a 3.5MVAR inverter to convert the direct current (DC) generated by the PV strings to alternating current (AC). A transformer is provided for each inverter to bring the voltage up to 13.2kV, which is what the client specified the voltage to be at the plant’s point of interconnection (POI) with the electrical grid. For the battery energy storage system, the client specified four 2.5MW / 2.5MWh batteries, for a total storage capacity of 10MWh. Bi-directional inverters were used to convert AC on the bus to DC to charge the batteries and convert DC from the batteries to AC to discharge the batteries. 

At the point of interconnection, a meter shall be installed to measure the voltage, frequency, active power, and reactive power. Additionally, a capacitor bank has been specified to compensate for the static reactive power losses due to the transformers in the system.

Next, the team designed a controller to send active power and reactive power commands to the inverters, in order to maintain the voltage and frequency at the POI as specified by the POI agreement (13.2kV @ 60Hz). The controller has additional logic to charge and discharge the batteries when needed and open/close the capacitor bank switch. 

The logic of the controller we designed has been implemented onto a programmable logical controller (PLC). PLCs are the industry standard, and we have used TCP/IP communication protocols to allow the PLC to communicate with other devices. The PLC is programmed in ladder logic, and the controller executes the rungs sequentially. To make the PLC scalable to systems of different sizes, the program has been written with all electrical values in per unit, and User Defined Types (UDTs) have been utilized for equipment like the batteries and inverters. These pieces of equipment all have the same parameters, and UDTs are essentially the equivalent of objects in other languages, allowing us to abstract all of the parameters associated with an inverter into one PLC tag. Additionally, we have separated the logic into different “routines” so that the logic is organized into separate tasks. For example, one routine is the distribution of the active power command to the PV inverters, while another one is the distribution of the active power command to the battery inverters.

We created a human-machine interface (HMI) to allow an operator to monitor system values and change parameters. The HMI is essentially a graphical interface for the PLC. We used Ignition, which is an HMI platform, and hosted it on a laptop. The laptop is connected to the PLC via ethernet cable to retrieve the tags. On the main HMI screen, we have a simple one-line diagram of the system, to monitor inverter values, battery charge levels, and the POI measurements. We also have a setpoint screen to change the major setpoints of the system, such as deadband values and the base power commands. Finally, we have a trend screen, which graphs critical values such as the POI voltage and battery states of charge over time. 

More details of the specifics of our solution can be found in the following video.

Next Steps

The next steps for this project include the following: developing more testing scenarios for the controller, analyzing the impedance reflections, programming the recloser controller, and writing an IEEE paper. The first item on the next steps is to include more testing scenarios for the controller which would make it more robust and ready for field deployment. The next item to be completed is the impedance reflections to determine if the reflections are more costly than having 1 large cable size for the entire system. The programming of the recloser is dependent on the location of the system and the grid sequence impedances. From this, the programming of the entire system would be ready for deployment. The team has also considered the possibility of utilizing excess active power on the grid during over-voltage conditions to charge the lithium-ion batteries, and if the batteries are capable of accepting more power. Rather than forcing upon the power plant controller to not respond to an over-frequency condition, this inclusion of functionality would increase operational capabilities of the power plant controller. The last step would be to compose an IEEE paper that would enable others in the power industry to quickly adapt their scenarios to the work that has been performed by the Team.

Meet the Team