Distributed Generation: Things You Don’t Want to Miss!

by K. Dartawan, R. Austria

What is Distributed Generation (DG)? Unlike big generation stations connected directly to the utility’s transmission grid, DG is typically smaller, about 10 MW or less connected to the distribution network or customer side. The DG could be fueled by renewable sources such as photovoltaic (solar), wind, bio mass or could be non-renewable energy such as diesel or gas.

What Opportunity Does DG Bring? Locating generation near where the customers are reduces electrical losses and loading on distribution and transmission lines. For example: a utility in the West Coast has installed and owns photovoltaic systems on many residential and customer sides. This helps customers reduce electric cost while the utility could delay or defer capital investments in new wires. In addition, tax incentives provide an additional benefit to both customer and utility. There is indeed an opportunity in DG and it is not surprising that DG is fast becoming a viable option and gaining wider acceptance. Surely nobody wants to miss such opportunity!

How Do We Study the Impact of DG? While DG offers many advantages, a comprehensive study of their impact is still required especially when the DG is relatively big as compared to distribution capacity. Because this area is still evolving, experience in the field is limited with many lessons still to be learned. This is evidenced by the need for development of industry standards. Engineers involved in this area typically do not want to miss a thing in their analyses because this could mean failure or expensive tune-up during operations.

So what kind of analyses is needed for DG interconnection to the grid? This Blog attempts to jot them down some of the issues with brief descriptions. More detailed discussion of each subject and how to mitigate each issue will be available in Pterra’s blogs or from IEEE conference papers authored by Pterra Staff. In addition, industry standard IEEE 1547 also provides guidelines on this matter.

Issues of DG Interconnection

Islanding. Islanding occurs when the DG continues to serve load after disconnection of the utility source by either manual or automatic means (e.g., operation of an upstream breaker or fuse). Islanding thus can occur intentionally or unintentionally. Intentional islanding is planned in advance and the system, as well as the equipment, must be designed to handle the situation. For example, the DG should be able to control voltage and frequency in the islanded portion within acceptable limits. For unintentional islanding, the industry standard requires the DG interconnection system to detect the islanding and cease to energize within two seconds.

Unintentional islanding could lead to the following risks:

  • Safety issue for both utility workers and the public
  • Out of phase reclosing which could damage utility equipment, customer load, and even the DG itself
  • Overvoltage due to neutral shift and resonance
  • Increase of restoration time; thus reducing reliability

Grounding and Ground-Fault Overvoltage The grounding scheme of the DG interconnection should not cause overvoltages that exceed the rating of the equipment connected to the grid nor disrupt the coordination of the ground fault protection on the grid. Unfortunately, these two issues should be addressed with a compromise approach (a trade-off approach). Attempts to limit overvoltage issue would increase the impact of loss-of-sensitivity on the grid’s relay coordination and vice-versa. With DG full-on and full-off, the study should determine the change in maximum ground fault resistance that will be detected by the protective devices, and estimate whether the change is within acceptable levels.

Short-Circuit. All generating sources generally increase short-circuit current on the system. The increase of short circuit current should not violate the rating of interrupting devices. Conventional synchronous machines such as diesel or gas would contribute significantly higher (6 €“ 10 times of their full load current) and decay much longer than inverter based DG (1 €“ 2 times of their full load current). Thus it is expected that installing inverter based DG on distribution side would have a small impact from short circuit point of view. Generally other issues, such as thermal overload, would hit first before the short circuit issue comes up. Although the impact of short circuit current is significantly less for inverter based DG, this type of DG has greater impact on power quality (flicker, harmonics, and voltage regulation) than the conventional synchronous machine, as discussed below.

Protective Relaying. Protective Relaying on distribution lines is generally designed for a radial system with only one direction of flow, i.e. from the bulk electric grid to the load. This is no longer valid when DG is connected to the customers or load side, the flow could go from both directions either going from the load side to the grid or from the grid to the load side. If not handled properly, this could create many issues such as loss of coordination, sympathetic tripping, and loss of sensitivity. All these effects reduce reliability of the distribution system. Customers could lose power for an isolated fault on the distribution system or healthy feeder trip where they normally would stay online if no DG were present.

Steady State Power Flow (Thermal, Voltage Performance, and Power Factor). Steady state power flow is tested with DG full-on and DG full-off. With DG full on and worst case contingency, lines/transformers should be within their contingency thermal rating. The contingency scenario in distribution is typically one feeder out; and some loads are switched to the feeder under the study. Similarly for voltage, with DG full-on and full-off, the voltage performance on some key buses/substations should be within the defined criteria. For DG with variable output (photovoltaic and wind), the simulation with load and generation profiles should be performed to assess the number of extra tap change and capacitor switching operation over certain periods. Power factor impact can be minor because the utility generally require the DG to be operated at unity power factor.

Voltage Flicker. Most people experience light flickering. This phenomenon is caused by a small magnitude change (few percent of nominal) but rapid fluctuation in system voltage that can result in observable changes (flickering) in light output. DG equipment may cause voltage flicker through the starting or stopping of generators, normal and abnormal output fluctuations of generators, and dynamic interactions with other loads or other DG units on the system. Because voltage flicker is mostly a problem when the human eye observes it, usually it is considered to be a problem of perception. Thus, to determine the appropriate levels of flicker that can be allowed on the system, flicker sensitivity curves are used such as the one developed by General Electric, shown in the figure. The figure indicates that the most sensitive frequency range for voltage flicker is approximately 5-10 Hz. In essence, this means that the human eye is more susceptible to voltage fluctuations in this 5-10 Hz range (5 €“ 10 dips per second). In this region, the eye can detect a recurring voltage fluctuation of about 0.25 percent. As the frequency of flicker increases or decreases away from this range, the human eye generally becomes more tolerant to luminance fluctuations. In order to limit voltage flicker to a level where it could not be detected by the eye, voltage fluctuation should not be greater than the indicated threshold of perception curve. The flicker at load points could be assessed using a short circuit study or using the maximum voltage change from steady state power flow cases (DG full-on and full-off cases).

Harmonics. Harmonics are a steady state distortion of the fundamental frequency (60 Hz) which could potentially cause overheating of transformers and rotating equipment, mis-operation of electronic devices, failed capacitor banks, wasted capacity €“ inefficient distribution of power. Inverter based DG such as photovoltaic units produce a dc output that needs to be converted into ac for the interconnection to the electric-grid system. A pulsed-width-modulation (PWM) inverter power supply is often used to transform the dc to ac. With the AC output being made-up of small width steps, the AC output will have a harmonic content. Depending on the technology (such as how smoothing is done in the inverter), the harmonic content will vary among DG manufacturers. Because inverter based DG is a harmonic source, it could cause voltage and current distortion. The distortion could be very high especially when circuit configuration has resonance point near the frequency point where harmonic source is the greatest. The amount of allowable harmonics in the system is specified by the following two standards:

  • IEEE-519- €œIEEE Recommended Practices and Requirements for Harmonic Control in Electrical Power Systems€
  • IEEE-1547 €“ €œIEEE Standard for Interconnecting Distributed Resources with Electric Power Systems€.

System Stability and Transient Study For DG that is very €œbig€ as compared to the electric grid, stability analyses and transient study may be needed. Generally the cost and the efforts required to complete such studies would be very significant. For example, for stability analyses the model and databases in both transmission and distribution sides should be merged. The process is typically tedious and very time consuming since the database and software tools used are different. Most of the time, the stability database or model for distribution system is not readily available; thus, modeling from scratch is needed. Generally previous experience with existing system, type of DGs and careful engineering judgment mainly decides the need for such studies.

Summary This Blog has summarized issues that could potentially be faced when interconnecting distributed generation. This could be used as a checklist for study requirements. Further discussion is probably required to identify scenarios where more detailed studies (such as stability and transient study) are required. More detailed discussion of each issue will be available on Pterra’s blog in the future.