The development of the synchronous generator in the late 19^th century was a catalyst for the energy revolution we experienced in the 20^th century. Charles Fortescue's paper demonstrating that unbalanced phasors could be expressed as a symmetrical set of balanced phasors was the match that lit the fire of this energy revolution. This paper is regarded as the one of the most important papers written in the 20^th century and it has laid the foundation for how every single utility in the world performs fault analysis. The underlying assumptions in this analysis are 1) faulted system is linear, which means sources can be represented by a representative Thevenin model, 2) load currents can be neglected compared with fault currents. However, times are changing, and so must our methods of fault analysis.

            Over the past 30 years the price of fossil fuels, climate change awareness, and efficiency of non-conventional methods of generation such as wind and solar have all increased drastically. This, paired with progressive policy-making using tax breaks and renewable quotas, has begun another revolution in the power industry. Wind and solar are growing at an accelerating rate and this growth is causing waves in the utility industry. These resources use inverters to create AC waveforms on the grid. The primary problem with the proliferation of inverter-based resources is that almost all of them limit the amount of current they can output during a fault scenario to protect their internal components such as MOSFETs and IGBTs. In addition, most inverters connecting solar generators and Type IV wind turbine generators block negative sequence currents. This means an inverter-based resource (IBR) cannot be modeled as a linear source. Due to the low fault contribution the practice of neglecting load currents in fault analysis also comes under scrutiny.
            Thus, as IBRs reach higher rates of penetration (and in the case of certain microgrids, 100% penetration) traditional ways of carrying out fault analysis and standard protection schemes will prove to be incapable of achieving their performance objectives. This research will focus on developing new ways to perform fault analysis by using an iterative method to accommodate the behavior of nonlinear sources. The approach will be based on recommendations developed by Working Group C24 of the IEEE Power System Relaying and Control Committee (PSRCC) [1]. The approach uses the output characteristics of IBRs over a range of terminal voltages provided by the manufacturer, which allows for a control-agnostic modeling. By using a standard IEEE system and methodically replacing synchronous generators with IBRs the new approach to perform fault analysis will be demonstrated that can hopefully be scaled up to be used in field. Results will be validated with the same system simulated in electromagnetic transient program (EMTP), using PSCAD software.

[1] "Modification of Commercial Fault Calculation Programs for Wind Turbine Generators", draft report developed by Working Group C24, IEEE Power System Relaying and Control Committee.

Technological advances, federal and state policy, and public opinion supported the exponential increase in PV penetration which has already changed the landscape of distribution planning. While the high PV penetration steepens the ramp rates for the generation and transmission planners, the power producing hours for PV do little to nothing to reduce the demand for the distribution planners. This is especially true for utilities that are winter peaking - as this peak likely occurs in the early morning before the sun has risen. While the high penetration of PV may not affect the capacity needs for distribution planners, utilities have already experienced: back feed through the substation transformers, excessive equipment operations, and difficulties with protection and control.

As distribution planners continue to deal with these problems presented by the increasing penetration of PV, they must now also consider the difficulties that will be faced as battery energy storage system (BESS) penetration starts with unforeseeable adoption rates. While utility owned BESS might postpone distribution upgrades, customer owned and operated BESS might expedite the need for distribution upgrades. This research seeks to present the possible outcomes from a variety of penetration scenarios to guide utilities in their planning as they shift their distribution planning paradigm to address the inevitable adoption of behind the meter BESS.

The IEEE 123 bus system and EPRI Ckt5 systems were used for this analysis. Since historical data is unavailable for these systems, a study was performed to determine whether the random number generator used for assigning load shapes would negatively affect the results of this research. It was determined that the effect of the random number generator was negligible.

A dataset created by another student in conjunction with this research was used. There are 100 sets of curves, each set containing: (1) base load, (2) base load + PV, (3) base load + PV + Time of Use (TOU) BESS, and (4) base load + PV + Maximum Demand (MD) BESS. For each circuit, every load was assigned one of these curve sets. The base case was simulated by applying (1) to every load. The solar penetration level was increased by replacing the penetration percentage of (1) with (2). The battery penetration level was increased by replacing the penetration percentage of (2) with (3) or (4). That is, 50% PV penetration with 50% BESS penetration represents 50% (1), 25% (2), and 25% (3) or (4).

As is expected, the increasing solar penetration produces the duck curve. Due to the choice of intervals used for the TOU pricing scheme, the increasing battery penetration worsens the duck curve. With minimal effect on reducing the peak, the extreme case of 100% battery penetration with 100% solar penetration has the worst back feed and the most variability in the head of feeder demand. For the MD pricing scheme, the increasing battery penetration level flattens the demand curve. The morning and evening peaks are reduced and the ramps are smoothed out. The valley in the middle of the solar day does not disappear because there is not sufficient capacity in the battery to charge extra in hopes of reducing the peaks further.

The increasing solar penetration injects more variability in the voltage from hour to hour while reducing the voltage variability throughout the circuit for a given hour. The hour to hour variability is seen in the increased rigidness of the voltage profiles. The reduced voltage variability throughout the circuit for a given hour is seen by the tightened bandwidth - that is, the difference between the minimum and maximum voltage during the solar hours is smaller for high solar penetration levels. The introduction of the BESS worsens the rigidness but counters the tightening of the bandwidth. Both the increases in the solar penetration and battery penetration work to push the voltage outside of the acceptable limits.

Clemson, in conjunction with TECO Westinghouse, has been awarded funding by the DOE to create and test a high speed induction machine (HSIM) and drive. This machine runs at 15,000 rpm. This system is replacing a low speed induction machine and gearbox that typically steps up to a higher speed. Efficiency increases by taking out the gearbox. For testing purposes, the HSIM is coupled with a gearbox to a low speed induction machine (LSIM) to act as a varying load. There is a power amplifying unit (PAU) that controls the LSIM.

   The Speedgoat is a real-time target machine that is being utilized for communication between the both of the machines and their drives. Previously, the Speedgoat had not been used at the facility. Therefore, research has been done on communication protocols and how to implement on the Speedgoat. Initially, loopback test was conducted on digital inputs/outputs, IP/TCP, UDP to determine capabilities. Modbus communication will be the next topic as the PAUs use this protocol. Modbus will be used to send commands to both drives to stop testing due to faults. Currently, models in MATLAB/Simulink are being created to do HIL testing to simulate the faults and response of the Speedgoat.

protection strategy based on local measurements for multi-terminal dc grids

    Researcher- Munim Bin Gani, PHD student

    Advisor - Sukumar Brahma

    Funding- Idaho National Laboratories, Department of Energy

Recent advancement in renewable energy generation technology is enabling the conventional ac grid system to shift towards dc, or hybrid grid. Since majority of renewable technologies produces dc output, which is being converted to ac through inverters, it is being conceived that such dc sources be connected to a dc grid at low voltage levels to feed various electronic loads that inherently operate on dc, and chargers for rechargeable devices including electric vehicles (EVs). In addition to reducing conversion losses, this approach can benefit from the known operational simplicity of dc circuits.

Although this idea is conceived at microgrid-level at low voltages, dc is also gaining popularity in high voltage transmission systems. With the advent of voltage source converter (VSC), HVDC systems have achieved much better controllability compared to the old line-commuted systems. Theoretically, there's no stability constraints or length-restrictions on power transfer over dc lines. These technological advances and operational advantages have prompted huge investments in HVDC grid, with China leading the field, having installed unconventionally long dc lines and formed a multi-terminal dc grid [1]. However, successful implementation of multi-terminal DC grids is slow due to lack of proper protection strategies. The difficulty of breaking dc currents is well-known but is being addressed by power-electronic devices [2]. The still unresolved issues in protection of dc systems are speed and selectivity. With the inception of fault, the current in dc system rapidly spikes, and goes on increasing until it reaches a steady state value which is generally much higher than the tolerance limit of the inverter circuitry, or the breaking capacity of the associated circuit breaker. Thus, fault current has to be interrupted in the transient stage, before it exceeds these limiting values. The required breaking time is shown to be a couple of milliseconds in HVDC [1] and a fraction of a millisecond in LVDC grids [3]. Such requirements are unprecedented. The protection system must reliably detect the fault (reliability), locate the faulted section (selectivity) and isolate the fault in as low as 50-100 microseconds (speed), depending on the system topology [3]. For this reason, main protection should be based on local measurements rather than remote communication, as communication latencies may not be able to match the speed requirements of protection for all topologies. In the existing HVDC systems, whenever a fault is detected, the whole system is de-energized due to lack of selectivity, which is a major drawback.

Due to speed requirements, the proposed methods of fault detection in different scholarly articles are based on traveling wave for HVDC systems. For LVDC grids, the methods are still based on overcurrent or undervoltage. However, selectivity is yet to be achieved for both LVDC and HVDC systems. With traveling wave based protection, the threshold values that have to be set for selective detection are highly system-dependent. This makes the settings vulnerable to temporary and permanent changes in system topology and creates obstacles towards further expansion of the system. A lot of simulations are required to determine the thresholds, which is time-consuming, and carries a risk of error under field conditions.

This project aims to completely resolve these issues and develop a protection scheme based on local measurements that lends itself to all topologies at high and low voltages, using the physics underpinning the system transients. An ambitious project, if successful, it will contribute towards the successful realization of multi-terminal and meshed dc grids.

 [1] Sukumar Brahma, "Protection of Distribution System Islands Fed by Inverter-Interfaced Sources", Proc. IEEE PES PowerTech 2019, Milan, Italy.

 [2] Sukumar Brahma, Nataraj Pragallapati, and Mukesh Nagpal, "Protection of Islanded Microgrid Fed by Inverters", Proc. IEEE PES General Meeting 2018, August 2018, Portland, USA.

flexible protection scheme for microgrids

    Researcher- Trupal Patel, MS student

     Advisor- Sukumar Brahma

    Funding- Sandia National Laboratories, CUEPRA