Campus Facilities Load Shedding Design

The project aims to design a microgrid load shed algorithm to balance power on the University of Idaho island when Avista undergoes a blackout event. We will determine which campus buildings get power and the priority of the buildings. The loads will be grouped in a logical way to effectively reduce overloading on University of Idaho microgrid generation.

=Introduction=

Background
The overall UI microgrid scope includes the entire campus, with a total of 143 floors, excluding Northern farms. Currently, 12% of the campus load is expected to come from the Steam Plant turbines, of which 3% must provide electricity for the Steam Plant. The remaining 9% of annual power production has been allocated for our load shed product. Our project scope includes designing the load shed algorithm for the following campus buildings: McClure, CNR, GJL, BEL. The current microgrid generation considered for this project includes the steam turbines in the Steam Plant and solar panels to be placed on the roof of IRIC. We will aim to design a module load shed scheme that can be scaled when more generation sources are determined for the UI microgrid.

Problem Definition
1.The UI does not have a systematic control system to control microgrid loads.

2. A load shed algorithm is required to efficiently regulate microgrid power.

Project Goals
1. Load shed algorithm to allocate the Steam Plant and IRIC power generation efficiently and safely a. Load prioritization for McClure, CNR, BEL, and GJL b. Easily modifiable load-shed algorithm c. Determination of switchgear required to implement the system d. Load shed system to be modularized to allow for microgrid expansion 2. System testing and verification with Real Time Digital Power System Simulator (RTDS) and SEL equipment

=Microgrid Simulation Design=

Specifications
Our goal is to utilize SEL equipment to implement an efficient system to allocate power from the UI Steam Plant to local campus buildings during Avista blackouts. SEL equipment has functions such as automatic bus synchronism, trip/close commands to remote enabled breakers and so on.

System Overview
This figure illustrates the microgrid one-line for the University of Idaho. The generation in this microgrid includes the Steam Plant turbines and IRIC solar. The loads include many buildings that will not be included in the load-shed scheme as well as the four priority buildings. The last two relays in this system will include two tie relays on either side of the microgrid. These tie relays will be responsible for islanding the system as well as returning to the utility.

Alternatives
Our initial test design consisted of nine sel-751A relays and one SEL-RTAC. Due to the fact that we only have six SEL-751A relays, we need to reduce the original design of nine relays to six. Our first design was to remove the IRIC, GJL, and non-priority buildings so that six SEL-751A relays and non-priority buildings could be simulated in RTDS without hardware. The goal is to keep the eastern and western ties of Avista and make the whole system more visual. In this way, we can show the process of island, load reduction, power balance and grid reconnection. Our current design is to remove Avista western tie, IRIC and non-priority builds so that we can keep the four priority buildings including McClure, CNR, GJL, BEL, according to the requirements of the campus. The remaining two relays are then assigned to the steam plant and Avista eastern tie.

Communications


The 6 SEL relays in this network represent one western tie relay, one relay on the Steam Plant, four relays on the load shed buildings. These smart devices must communicate to the RTAC in order for power to be balanced and for the system to respond to utility events appropriately. We decided to use a primary and redundant IEC 61850 GOOSE and MMS communication network to pass information to/from the RTAC and relays. The redundant network helps limit the chance of information being dropped in the Ethernet-based communication scheme.

RTAC Algorithm
This figure shows the inputs and outputs from the RTAC. Ideally, this controller would be centrally located at the Steam Plant. The controller is to take information from the relays on the system and issue the appropriate commands to operate the microgrid system. The utility ties will communicate their breaker status to the RTAC as well as the power flow into the system. Additionally, the RTAC will send a maintenance mode to the tie relays so when the Steam Plant is not operational, the system is not islanded. The Steam Plant relays will communicate the power generation (real and reactive) as well as the frequency. Frequency is included here because if the frequency of the system begins to drop, that is a sign that there is too much load. The priority buildings will communicate their power consumption (real and reactive) as well as voltage. Like frequency, if the voltage begins to drop at the load, that is a sign the system is overloaded. The priority buildings will receive open/close commands from the RTAC. The last relay type included in our network is the non-priority building relays. These relays will send power consumption (real and reactive) and receive open/close commands from the RTAC. Our RTAC algorithm begins with islanding. The UI system will island when the voltage drops for a set amount of time on the Avista bus. This logic will be done in the tie relays and their breaker status will be shared with the controller so the RTAC can implement the appropriate decisions for balance power in the island. This represents the first box in the figure above. The controller will receive power from the Steam Plant and tie relays, and will use this information to calculate how much power will need to be shed when the tie breakers are opened. Once this event occurs, the calculated load shed required will be used to trip campus buildings based on priority. Once the microgrid is in island mode, the power must be constantly balanced within the system. Our team is using two schemes to balance power. The first is contingency load shedding. A contingent load shed scheme compares the power generation to the power consumption and adds or removes load based on this comparison. The second scheme is frequency/voltage-based load shedding. The frequency of the steam turbines and the voltage of the loads will be monitored in the microgrid system. If either the frequency or voltage begins to drop, load will be shed to restore the power balance. The last action the RTAC must implement is returning the microgrid system to the Avista grid. The tie relays implement most of this logic. Figure shows when the voltage returns to the Avista bus for a certain amount of time, the tie relays begin a synchronism check. This synch check verifies the voltage, frequency, and phase on the Avista bus and the UI microgrid match. Once the synch check is complete, the UI microgrid is connected back to the grid. The breaker status from the tie relays is then sent to the RTAC which then closes all of the UI microgrid campus buildings back into the grid.

=Simulations Results and Analysis=

RTDS Draft
A Real Time Digital Power System Simulator (RTDS) in Dr. Brian Johnson’s lab was utilized to test the load shedding design. The RTDS simulates a power system that is developed by the user. This figure shows the RTDS draft that was built by the team to replicate the University of Idaho microgrid. The RTDS equipment provides real time analog voltages and currents to the relays to represent this system. The RTDS allows our team to change the load and power consumption as well as the utility voltage. Since the voltage and currents are sent to the relays in real time, the control algorithm in the RTAC was verified.

Runtime HMI
Runtime is the human machine interface (HMI) with the RTDS equipment. This gigure shows the Runtime model that was built so our team can interact with the RTDS hardware and manipulate the power system in real time. The main functions we implemented in Runtime are frequency and voltage control of the steam plant and utility, open and close for all breakers, and load consumption control.

=Lessons Learned=

=Team Information=

=Additional Documentation= Timeline Schedule
 * Timeline Schedule

Meeting Agendas
 * Meeting Agendas

Meeting Minutes


 * Meeting Minutes

Presentations
 * Snapshot Poster1
 * Snapshot Poster2
 * Design Review1