PROTECTION AND CONTROL

Integration of Recording Relays Into Disturbance Analysis

In the past electromechanical protective relays world, the only way to monitor and assess relay operations was by way of stand alone fault recorders. These fault recorders sometimes had mechanical methods of providing prefault information by way of rotating drums or through some sort of storage medium such as looped magnetic tapes. While somewhat crude by today's standards, these early recorders provided much insight into what took place during faults and system disturbances. As modern microprocessor power evolved in the 1980s, stand-alone recorders were able to use the power of digital data processing to provide the analytical engineer with power system voltage and current waveforms. These waveforms could then be analyzed in the electronic medium before being printed on paper. This, of course, was a tremendous step forward. Development and use of power system analysis tools such as fault, stability and electromagnetic transient programs also became more widespread. During this time the emergence of microprocessor relays occurred.

What Is Driving Today's Needs For Fault And Disturbance Recording?
Today's electric power utility operations are being driven more and more by issues like deregulation, where more has to be achieved with less human resources in a shorter time period.

To address power system reliability issues, the North American Electric Reliability Council (NERC) has been charged with the responsibility to see that this is achieved. It is not in the scope of this article to look at the pros and cons of NERC, other than to say that there exists some expectations by this group for power utilities to demonstrate knowledge of what their systems look like and what their performance is like during a particular type of fault or system disturbance condition.

Time Frames Associated with Power System Operation
Various events happen within various time frames. Some of these events are categorized below:

TIME FRAME	EVENT	EXAMPLE
microseconds	switching surges	breaker restrikes
milliseconds	harmonics	variable speed drives
cycles	fault clearing	relay/breaker operation
seconds	load flow changes	governor/exciter response
minutes	system stability	reaction to dispatch actions
hours	load variation	new generation schedules
days	NERC reports	system disturbance
months/years	system modeling	fault, load flow and stability

Today's Microelectronics World in the Utility Field
While it is true that many pieces of equipment are still protected by electromechanical relays, it cannot be denied that the digital world has arrived in the power industry. Processors such as the Digital Signal Processor (DSP) have a tremendous ability to perform mathematical computations for data processing and for performing protective functions. Digital technologies today are able to provide both a protection as well as a recording function in one box as a robust and technically feasible solution (See Figure 1).

Recording relays have the ability to cover recording time frames from milliseconds to hours with a sampling speed of about 100 samples per cycle. As well, this sampling rate can be averaged down to cover the power swing time domain. This permits issues related to harmonics, faults, load flow and stability to be covered as well. There is no reason to believe that this direction will not continue as station integration of all IEDs produces technically sound and efficient solutions.

Benefits of Using Fault Recording Digital Relays
Some of the benefits of fault recording relays are:

• Cost, minimizing external connections and hardware.
• Voltage and current recorded waveforms are the same as those used by the relay's protection.
• Relay functions that operate are all part of the fault record itself. That is, fault record events are analyzed by the same software as the analog volts and amp channels.
• Relay settings at the time of the fault are easily attached to the fault record to help produce a more complete fault database.

Overlapping Recording Zones
For a fault that may occur at Bus D, the recording relay at Bus C will see the fault by using its overreaching distance functions zones 2 and 3 (21-2 and 21-3) see Figure 2). Normally these zones of protection would have time delays applied to them for backup tripping protection. These relays could however be programmed to collect fault recordings immediately upon the detection of a fault into zone 2 or 3 and in this manner make fault recordings available for faults that occur at adjacent Bus D. In a similar manner, zones 3 and 4 provide recordings for faults at Bus A, B or E.

2. What happens if the relay gets burdened with recording that slow down its protective function?
Make no mistake, a recording relay's function must be to provide protection first, and recording second. Intelligent recording relay design can use one processor such as a DSP to provide protection, then hand off recording to another on microprocessor, thus leaving protection unaffected for any system condition.

3. Data rates for relay functions are typically in the range of 4 to 16 samples per cycle. Where does good quality fault recording come from?
Again, processors such as DSPs can be made to sample analog voltage and current quantities at rates of say 100 samples per cycle, then average down this rate as required for relay algorithm needs Ñ the NERC standard for recording relays is 64 s/c. Advances in memory technologies have made data storage in the relays possible.

4. What about Time Synchronization?
Most relays as well as stand-alone disturbance fault recorders are able to accept time signals such as IRIG-B to ensure accurate time resolution between events. Failure to do this with data collected by other IEDs can lead to a chicken and egg situation. Fortunately, IRIG-B is getting better and less costly for station applications.

Other Disturbance Data Sources
The presence of and the ability to collect disturbance data from recording relays is of course not the only data source. Other data sources include:

• Stand alone fault and disturbance recorders
• SCADA and RTUs
• Other relays and IEDs such as meters, controllers and special protection devices.

Basic data management model questions that need to be asked:

• Who needs the information?
• How soon is the information required?
• What type of information is needed for each user?
• Where is the information needed?
• When is the information needed?

• Where should the data be stored?
• What is the fallback position if data is not received?
• What to do with redundant data?
• How can data be integrated into a useful form?
• How is the data formatted?

The data from the various IEDs may be sent in the same or different formats. In order to place the data into the Central Station in a meaningful fashion, this data needs to be converted to a standard format such as COMTRADE.

Once this is done, various tasks are performed on this data. Appropriate data is absorbed into the various tasks by different data processors. For example, fault analysis can be performed by an artificial neural network or by a combination of human and software interactions. The results of this activity can either go off to other areas or can be put back into another part of the Central Station database.

In some cases, such as the need for fault location, the information may be required sooner than the Central Station information route can provide.

In these cases, the data from the relays is provided locally and through connection with systems such as SCADA to provide the information to the right people at the right time. This does not preclude the possibility of also sending information such as fault information to the Central Station for further statistical analysis.

Central Station Format
In order to provide enough flexibility in the use of data collected by the Central Station, the Central Station must be structured to allow this to take place quickly and effectively. One Central Station possibility is presented in Figure 4.

This Central Station model allows the data collected from recording relays, recorders and other IEDs to be distributed to various user workstations where data is analyzed or used by other software programs. This model also goes some distance in addressing issues such as backup data storage. In all the models for common data use, time synchronization and the ability to generate standard data forms using standards such as COMTRADE to normalize the information is important. If this cannot be achieved, then the user is forced to use several Central Station programs that make it very difficult to correlate to a given fault or disturbance. This unfortunately still happens today for far too many post-disturbance analysis activities.

Conclusion
As the need to know exactly what happens on power systems is becoming more and more important, any and all sources of fault and disturbance data will be sought and used by power system analysts and system operators. All information available for a particular disturbance will therefore be actively used.

For all faults and disturbances, it is guaranteed that one form of protective relay or another will be involved and assessed. The ability of a recording relay to provide information about the power system during these disturbances is therefore of great value in the analysis process.