MOD-033-1 Guideline1

Guidelines for Validation ofPowerflow and Dynamic Cases for MOD-033-1

155 North 400 West, Suite 200

Salt Lake City, Utah 84103-1114

Western Electricity Coordinating Council

MOD-033-1 Guideline1

Table of Contents

1Introduction

1.1Purpose

1.2Definitions

2Overview of MOD-033-1

3Discussion

4Monitoring Equipment

5Methodology

5.1Selection of Events

5.2Data Acquisition

5.3Real-time WSM Validation (Peak RC best practice)

5.4Steady-state Model Validation (using the WECC power flow base case)

5.5Dynamic Model Validation (WECC base case)

5.6Review of Steady State and Dynamic Model

5.7Comparison

6References

7Approval

8Appendix A

9Appendix B

1Introduction

The NERC Steady-State and Dynamic System Model Validation Standard, MOD-033-1, was created to establish consistent validation requirements to facilitate the collection of accurate data and building of planning models to analyze the reliability of the interconnected transmission system. One of the requirements in this standard is that each Planning Coordinator shallimplement a documented data validation process. WECC created theSystem Model Validation Task Force (SMVTF) under the WECC Modeling and Validation Work Group (MVWG) to facilitate the MOD-033-1 validation process and to enhance the model validation.

1.1Purpose

This guideline is meant to provide basic information on how WECC staff and Peak RCprepare the studycases for model validation and performs steady-state and dynamic system model validation. The cases may be leveraged for use by its members, as applicable, to meet the MOD-033-1 compliance.

This guideline also includes a description of monitoring equipment requirements,how members can use these cases for model validation, and provide some guidance on identifying unacceptable differences and how to resolve those differences. This is intended to provide some consistency for members across WECC.

1.2Definitions

Term or Acronym / Definition
SMVTF / System Model Validation Task Force
RC / Reliability Coordinator (Peak RC)
SE / State Estimator
PC / Planning Coordinator
WSM / West-wide System Model

2Overview of MOD-033-1

The NERC MOD-033-1 standard becomes effective on July1, 2017. Here is a brief overview of MOD-033-1: requirements (for the complete requirements, see the NERC MOD-033-1 standard):

R1.EachPC shall implement a documented data validation process that includes the following attributes:

1.1 Comparison of the performance of the PC’s system in a planning power flow model against actual system behavior represented by State Estimator (SE) case or other Real-time data sources;

1.2 Comparison of the performance of the PC’s system in a planning dynamic model against actual system response;

1.3 Guidelines the PC will use to determine unacceptable differences in performance under 1.1 and 1.2;

1.4 Guidelines to resolve unacceptable differences identified under 1.3;

R2: Each RC and TOP shall provide actual systemdatanecessary to the PC to perform validation under Requirement R1 within 30 calendar days of a writtenrequest.

Validation of the planning power flow and dynamic models are to be performed at least once every 24 months.

3Discussion

The focus of MOD-033-1 is comparison of the performance of the PC’s portion of the existing system for steady-state and dynamic responsefor a local event. Additionally, it is specified in the standard that a dynamic local event could also be a subset of a larger disturbance involving large areas of the grid. NERC’s main emphasis is the utilization of local disturbances for the evaluation of the model; however, there are numerous advantages in the use oflarge disturbance events within WECC, if available and relevant to the PC’s validation of its system model.Following arguments emphasize importance and advantages of using large interconnection wide disturbance instead a local one:

  • Dynamicsystem model validation requires full knowledgeof which units are on-line within theinterconnection during the event used for validation.
  • Interconnectiondynamic response including frequency response (initial and primary) depends on all units within a system and the planner must know which units are on line in the remote parts of the system in order to validate the simulated frequency response.;
  • After adisturbance occurrence, real and reactive power flows on major transmission lines and paths are directly affected by the response of many generating unit. If the status and outputs of on-line generating units in areas remote from the actual disturbance are not known, theresults of the event simulation maysignificantly differ from the actual measurements.
  • Correct dynamic modeling of generators is essential to properly simulate its behavior. Dynamic oscillatory behavior of the systemdepends on available rotational masses and excitation system response of the generators within the system. For example, after a system disturbance, the generatingunits in California may oscillate against generating units in Alberta. Hence,unless the generating units in Alberta are represented accurately,the resultant simulation response may significantly differ fromthe measurements.
  • Usage of large interconnection events for system validation has the benefit that all affected PCs within WECC may be able touse the same power flow basecase and dynamic model. Using the system common case may significantlyreduce time needed for base case preparation and potentially couldenhance the model validation process. In the case that a system event does not cause a significant impact on a PC footprint, the PC may choose to evaluate local events that could provide better validation of its power flow and dynamic models.
  • The adjustment of the WECC base case utilizing exclusively local events may lead to an over-tuning of the models in an attempt to match the simulation to the field measurements. For instance, if all PCs modify the modeling parameters within their area based only on localevents, the sum result may lead to the incorrect validation of the WECC-wide system model, particularly the dynamic response.

4Monitoring Equipment

Sources of actual measurement data that can be used for power flow model validation include:

  • EMS/State Estimator (SE)
  • SCADA/PI Historian
  • West-wide Model (WSM) case (from WECC EMS)
  • Phasor Measurement Units (PMUs)

This is not a complete list of possible sources. As long as the data providedis a time synchronized snapshot of the system (includes voltage, real and reactive power flow, status, settings, etc.), it could be used for power flow model validation.

Sources of actual measurement data that can be used for dynamic model validation include:

  • PMUs
  • DFRs (Digital Fault Recorder)
  • Relays

This is not a complete list of possible sources. As long as the data provided is time synchronized and can provide the data with at least 30 samples per second of the positive sequence data (including voltage, real and reactive power flow, frequency, phase angle), it could be used for dynamic model validation. It is better if the sampling rate is 30-60 samples per second, especially if new equipment is being installed. The input sampling rate is typically much faster (typically in kHz) than the data in the output file.

If continuous monitoring (which is preferred) is not available, a suggested starting point for trigger settings are as follows:

  • Voltage < 0.9 pu (note: may be in reference to either normal operating voltage (e.g. 117 kV) or nominal voltage (e.g. 115 kV), depending on whether it’s relevant and applicable to the PC’s existing system)
  • Voltage change +/- 5-10%
  • Frequency < 59.9 Hz or > 60.1 Hz
  • Event recordings to include a minimum of 2 second pre-event and 5 seconds post-event conditions. The ideal recommended recording times, if possible, should be 10 second of the pre-event conditions and 60 seconds of the post-event conditions.

These may be adjusted further based on the PC’s knowledge of its own system and engineering judgement.

Monitoring equipment for dynamic local events are located based on what is appropriate for each PC’s existing system. The number of devices will vary depending on the entity. The following considerations for locations of dynamic monitoring devices include:

  • At and/or near generation facilities
  • At major transmission facilities
  • At major load centers
  • At major interconnection points
  • Bulk Electric System (BES) buses with large reactive power devices

Most non-PMU recording devices will provide data as a point-on-wave quantity, at multiple samples per cycle. To effectively perform model validation, those recordings will need to be converted to RMS quantities in post-processing.

The high sampling rates necessary to capture the dynamic behavior of the system imposes a burden on the storage capacity of the recording devices; specifically, DFRs, relays, and PQ meters. For this reason, PCs will need to implement manual or automated systems to avoid event data over-writing.

Multiple software tools exist that can automatically poll DFRs/relays for new events, usually stored in COMTRADE format, and downloaded to a more permanent location. These tools can be installed within a substation (on a hardened PC, for example), which requires manual retrieval by someone at the station. Alternatively, if the communication system allows, it’s possible to install a central retrieval unit to poll field devices and download event records to a central location for storage and analysis.

5Methodology

There are essentially two components of planning model validation for the MOD-033-1 standard: steady-state (R1.1.1) anddynamics(R1.1.2). Both require adjustments of the WECC power flow basecase, either to the selected date and time or to pre-contingency event conditions. While the steady-state and dynamic system models can be validated separately, it may be more logical and efficient to use the same event and power flow case for validation of both (R1.1.1 and R.1.1.2). Note that an accurate steady-state model is needed for dynamic validation but asteady-state model validationdoes not require having a system event. Both of these topics are tightly linked but will be discussed separately.

This guideline is not meant to be perceived as the only acceptable document to be used for model validation. Additionalinformation is provided in references [1,2].

5.1Selection of Events

The first step inMOD-033-1 model validation is to select an event against which system response will be validated. Large system event occurs infrequentlyand unplanned and the one to be used for model validationmust be selected carefully.

An example of a useful large event for system model validation is the loss of PDCI and associated RAS that include multiple generation tripping. Thistype of eventand corresponding impacts are observed and recordedwidelywithin the interconnection. Other useful events are transmission line faults followed bythe loss of a large amountof generation. SMVTF will choose and prepare at least two WECC-wide system event cases annually. If time permits, and should other interesting event(s) occur, SMVTF may decide to prepare additional study cases. Another potential case that may be selected is a case for steady-state validation only, such as heavy winter peak or heavy summer peak scenario or choosing disturbance that eventually occurred during system conditions close to heavy winter or peak summer conditions.

Some events may not be suitable for MOD-033-1 validation purpose. These events may include,for example, asymmetric events that include highly unbalanced flows such as single pole reclosing or an event that occurred at the top of the hour when generating units are ramping up or down. In study simulation, during initialization process, we assume that all generating units are static with fixed outputs, but over the course of the simulation progress, where thetimeframe typically lasts 60 to 120 seconds, some of the units may ramp up or down and they would need additional modeling efforts to simulate. Such effort is unlikely to add value to model validation.

Fig.1 Example of path flow decreasing due to generation ramping down. In this case the units that ramp down need to be scaled down during the simulation.

5.2Data Acquisition

Once a system event is selected, the data (i.e., from state estimator, SCADA and PMU) for the system and the time duration being simulatedshould be acquired. The data can be requested through the Reliability Coordinator (RC) and/or TOPs (per MOD-033-1 R2). The RC in WECC is Peak Reliability (Peak RC). Peak RC can provide a snapshot from their State Estimator (SE) data prior to and immediately after the event.

ThePeak RC archives SE cases every 5 minutes. For that reason there is a need to verify if there were additional switching actions within the time of the saved SE case and the time of the event (theoretically this time interval can be a maximum of up to 5 minutes).

ThePeak RC receives over 130,000 real-time measurements that are mapped to the West-wide System Model (WSM). The measurements include analog data (MW, MVARs, kV) and statusof the equipment. These measurements allow the model to be adjusted to match real-time conditions. Thesemeasurements are received every 10 seconds via ICCP links.

Event sequence should be requested from the TOP(s) in the area that the event occurred. The TOPs would have the most accurate information for the event.

Peak RC will use PMU data available from different part of BES to validate general dynamic response. Individual PCs will use their own data for validation of their portion of the system dynamic response

5.3Real-time WSM Validation (Peak RC best practice)[1]

To enhance the process and data accuracy, Peak RC staff will simulate the selected event using WSM directly and compare the simulation results to available PMU data. The following are two reasonsfor this process:

To validate WSM itself for Peak Reliability;
  • To make sure that WSM snapshot provided for model validation represents event accurately since this snapshot is used as modeling inputs to the WECC power flow base case for pre-contingency operational conditions;

The process of WSM validation is relatively quick since the WSM case is a representation of the pre-event conditions adjusted by real-time SCADA measurements. Only event sequence needs to be investigated, prepared and then included in the study. Since the WSM case uses the same dynamic model as the WECC power flow case, all modeling issues found in this process will be reported to the WECC staff.

5.4Steady-state Model Validation (using the WECC power flow basecase)

AfterWSM validation is performed,the WECC power flow case needs to be adjustedand prepared based on the pre-contingency operating conditions. This process will be performed by the WECC staff and it may require a few weeks to prepare the power flow study case. This is the most time consuming part of the process.

For the preparation of the event base case,theWECC steady state planning model must be modified with generation dispatch, topology and load changes based on the real-time data noted above in order to achieve a close match to actual system condition for the selected time. Reference [1] from the NERC MWG document provides more details on this process. There could be some limitations on the part where the WECC staff can modify the power flow study case for the overall WECC-wide footprint. Additionalrefinementof the power flow study case may be required by the PC for their own planning area.

The main intent for validating a steady-state power flow model is to compare the pre-disturbancemeasurement (e.g. bus voltages, real and reactive power flow on system elements and paths, generation dispatch, phase shifter settings, LTC tap positions,etc.) to the powerflow solution from the WECC study case that is adjustedto the pre-disturbance operating conditions. The desired outcome would be a close match of the results obtained between the power flow simulation and the real-time measured data.

If the results are not a good match, based on engineering judgement, it is necessary to investigate the cause(s) of the discrepancies.Therefore, it is recommended to adjust voltages by allowing LTC taps, SVCs and generators to adjust automatically based on measured conditions (AVR selection for power flow solution) and then to compare simulated tap positions and MVArvalues to the actual values. Thisprocess is helpful to pinpoint the issues and to correct transformer tap positions, controlled points and transformer impedances. Note that there will be differences between the WECC power flow case and WSM (State-estimator snapshot) study case. The difference is mainly due to mismatches that are introduced to the SE power flow case during the process of state estimations. The mismatches appear as small MW and MVAr loads that are added to the study case.

Process described above is used for events selected by SMVTF. If PC(s)decide to use other event than one chosen by SMVTF then they willneed to adjust basecase by themselves. Below is brief procedure on how basecase is adjusted:

In order to create an event case using a WECC base case the following will need to be adjusted in the WECC base case, generation dispatch, voltage, load dispatch, transmission line status and adjusting the flows on the Phase shifting transformers to match the WSM case.

In mapping the WSM[2] to the WECC base case start on the outer rim of the WECC footprint such as Alberta. Set the generation dispatch from the WSM into the WECC base case. This can be done by running a script that will use the WSM generator ID that will match the same ID in the WECC base cases. Next adjust the load dispatch to match the net MW interchange in or out of Alberta. The load dispatch is typically a little different due to different losses between the WSM and WECC base cases. Repeat these steps for all areas in the WECC base cases.