Fault Recorder Standardizes on DADiSP

Fault-recorder vendors standardize on popular analysis package

While we all benefit from the economies of scale with regard to PC hardware, such isn't always the case for software, especially programs that work with niche peripherals. A good example comes from the power industry, where a handful of manufacturers supply digital fault recorders (DFRs) to a limited number of customers. Many of these manufacturers are relatively small and don't have the resources to write sophisticated analysis software to accompany their hardware. Thus, after the industry leaders decided to export data in several file formats including one for Dadisp from DSP Development Corp (Cambridge, MA (800) 777-5151), virtually all the others followed suit and that format has become a de facto standard.

For users not familiar with a DFR, that device is a data-acq system that electric utilities and other producers of power use to monitor the performance of generation and transmission equipment under stressed conditions. One common form of stress is a lightning strike to a transmission line. The entire sequence of events from a strike to fault recognition by protective relays to fault clearing with circuit breakers is accomplished automatically in a span of 0.05 to 0.083 sec, so the process is far too fast for human intervention. During such a fault, a DFR triggers, saves a record of the power and current, and then it later transmits the information to central offices over a modem, where a utility engineer can perform post-event analysis to determine if the relays, circuit breakers and other equipment functioned properly.

Because of their limitations and because a half dozen firms are fighting for sales in a small and very competitive market, DFR manufacturers only provide bare-bones analysis software that can typically measure the peak value of a waveform or its effective rms value. For any other calculations engineers are on their own, and hence the need for such units to export data in a format that other analysis packages can read. The selection of Dadisp arose, says John Demcko, a senior consulting engineer at Arizona Public Service Co (Phoenix, AZ, jdemcko@apsc.com), because the product has established a good history and it seems to fit well the types of analysis that electrical utilities collect.

As an example of the type of work that John does within Dadisp, he gave an example from Cholla 4, a coal-fired plant. That facility uses wide-bandwidth transducers to provide field voltage and sometimes current to a DFR. Recently a question arose about the validity of analog dial readouts of field voltage in the control room. Within half an hour of importing the DFR data, he had set up the software to complete his verification analysis.

Field voltage is a complex waveform whose rms value you can't determine by inspection. Instead, you calculate the rms value of a periodic time-varying waveform V(t) as follows:


where t(o) is the starting time, T is the period.

The nearby screen shot shows the 6-window Dadisp setup John uses to implement this equation. The first window (W1, upper left) shows the field voltage record, and W2 allows him to apply a scaling factor, which wasn't necessary in this case. W3 displays the square of the time-series waveform, while W4 calculates the integral of the squared-voltage function. He decided to calculate the voltage over three cycles of the 180-Hz exciter-system frequency, leading to 3T = 0.01667 (as in W5). Finally, W6 shows the square root of the integrated waveform over three periods. The calculated rms value of generator field voltage is measured as 232.16V rms, which agrees within 3.3% with the 240V dc read from the control-room meter.

John admits that these calculations aren't terribly sophisticated, but he can set them up or modify them quickly within a Dadisp worksheet. As examples of more-sophisticated applications, he cites another case. Voltage dips during normal breaker-clearing times at the service entrance of an industrial customer's facility were suspected of creating problems with sensitive electronically controlled loads. A frequency-domain analysis using built-in windowing and FFTs revealed that the customer's loads had a very high harmonic content. More recently, John has used the software to investigate the slow degradation of windings on a generator.

Although pleased with the software's performance, John would like an easy way to print out a window that includes a cursor on a point in a waveform showing its value. He presently annotates the printed copy. For future enhancements, he'd like to automate the data-import process. Before he can import data into Dadisp, he must run a DFR-supplied utility that converts raw data and makes a Dadisp-compatible file. He feels it's almost certainly possible to automate the process with the macro language within Dadisp, but he hasn't found the spare time-he's only learned what's needed to get the job done now.

Copyright 1996 PEC Inc. at
Personal Engineering