HK1180040B - Method and system for evaluating noise and excess current on a power line - Google Patents

Description

Method and system for evaluating noise and overcurrent on power line

COPYRIGHT NOTICE

Some of the material contained in this document is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, but does have the copyright rights whatsoever.

Background

1. Field of the invention

The present invention relates to power lines for power distribution, and more particularly, to an evaluation of noise and overcurrent on a power line for the purpose of identifying the location of a source of noise or overcurrent.

2. Description of the related Art

The approaches described in this section are approaches that could be pursued, not necessarily approaches that have been previously demonstrated or pursued. However, unless otherwise indicated, the approaches described in this section are not prior art to the claims in this application and are not admitted to be prior art by inclusion in this section.

Partial Discharge (PD) is a phenomenon that occurs in damaged insulation of a power cable, the cause of which may be, for example: aging, physical damage or exposure to excessive electric fields. PD can damage cables, connectors, lightning arresters and other high voltage equipment. A defective insulator for an overhead cable also causes noise generation and is accompanied by frequency and phase characteristics similar to PD. The PD generates short pulses, which are nanosecond or shorter in duration. PD pulses often occur in a particular phase of an Alternating Current (AC) power voltage and are approximately synchronized with the power frequency or twice the power frequency. PD is considered to be one of known noises such as line synchronization noise or line trigger noise. PD pulses have a continuous, broadband spectrum that typically includes a range of 1 khz to hundreds of mhz.

Us patent 7,532,012 discloses various techniques for capturing PD pulses and also discloses various parameters for evaluating waveforms to distinguish between waveforms related to line frequency synchronization phenomena, such as PD pulses, and external disturbances, known as "intrusions", which do not have line frequency periodicity. For waveforms discerned from a PD, these parameters have utility in further quantifying the intensity of the PD.

Damaged cables may also be subjected to very brief high current pulses, such as: an arc or other temporary short circuit condition across itself may be expected. It is necessary to identify the damaged cable before it completely fails, more particularly to find its specific location.

Disclosure of Invention

Techniques for evaluating parameters related to noise and over-current on a power line are disclosed. These techniques are particularly well suited to characterizing Partial Discharges (PDs) and to identifying the location of noise sources or sources of over-current.

One of these techniques is a method comprising (a) measuring a maximum amplitude of a first spectral component of a partial discharge pulse sensed on the power cable, (b) determining a phase of a power frequency signal on the power cable where the maximum amplitude of the first spectral component occurs, (c) measuring a maximum amplitude of a second spectral component of the partial discharge pulse of the phase, and (d) determining a location on the power cable where the partial discharge pulse occurs based on a relationship between the maximum amplitude of the first spectral component and the maximum amplitude of the second spectral component.

Another of these techniques includes (a) measuring peak amplitudes of spectral components of a PD pulse sensed on a power cable by multiple phases of one cycle period in a power signal on the power cable; (b) subtracting a background noise level from the peak amplitude to produce a resultant amplitude, and (c) adding the resultant amplitudes to produce a sum of PDs, which is indicative of the activity of PDs on the power cable.

Another of these techniques is a method comprising (a) measuring, at a first location of a power cable, a first magnitude of a first current exceeding a threshold; (b) measuring a second magnitude of the second current at a second location on the power cable that does not exceed the threshold, and (c) determining a location of the fault on the power cable based on a relationship between the first magnitude and the second magnitude.

The invention also discloses a system for performing the above methods, and a storage medium containing instructions for controlling a processor to perform the methods.

Further, there is provided a system comprising:

(i) a switch through which noise from the power line passes when the switch is closed; when the switch is open, noise cannot pass through the switch;

(ii) an amplifier, arranged after the switch, which generates an amplified output;

(iii) a channel having:

(a) a filter through which the amplified output spectral components pass within a particular frequency band to produce a filtered output; and

(b) a detector for detecting the filtered output values a plurality of times to generate a series of values; and

(iv) a processor:

(a) the processor determining a minimum value of the series of values when the switch is open, thereby obtaining a first baseline value;

(b) the processor determining a minimum value of the series of values when the switch is closed, thereby obtaining a second baseline value; and

(c) a difference between the second baseline value and the first baseline value is determined, resulting in an excess value representing an excess value of power line noise that is greater than the amplifier noise.

Drawings

Fig. 1 illustrates a portion of a power distribution system having a plurality of components for detecting Partial Discharges (PDs) on cable lines in the power distribution system.

Fig. 2 is a block diagram of a PD detector.

Figure 3 illustrates a portion of a power distribution system including a network of couplers and PD detectors to detect partial discharges at multiple locations within the power distribution system.

Fig. 4 is a graph of background noise for a single cycle of a power frequency signal for a single channel of a PD detector.

Fig. 5A is a graph of a signal at the channel output of a PD detector, where the signal includes a single PD pulse.

Fig. 5B is a graph of data points exceeding a threshold for the same signal as in fig. 5A.

Fig. 6A is a graph recording two PD pulses captured in the channel of a PD detector.

Fig. 6B is a graph of the same two PD pulses as in fig. 6A, captured in another channel of the PD detector.

Fig. 7 is a graph of sampled signals, where each sample value represents a separate time block, e.g., 4 phases, in a single channel or PD detector for a single cycle of the power frequency.

FIG. 8 is a state diagram of a state machine that controls transitions between a maximum hold mode and a single cycle mode in a PD detector.

Fig. 9 is a schematic diagram of a peak current recorder.

Common components or features are denoted by the same reference numerals throughout the several figures.

Detailed Description

In power line communication systems, the power frequency range is typically 50 to 60 hertz (Hz) and the data communication signal frequency is greater than 1 megahertz (MHz), typically in the range of 1 to 50 MHz. A data coupler for power line communications couples data communication signals between a power line and a communication device, such as a modem.

One example of such a data coupler is an inductive coupler that includes a magnetic core and a coil wound around a portion of the magnetic core. The core is made of a magnetic material and includes a hole therein. The inductive coupler operates like a transformer and is located on the power line such that the power line passes through the aperture and acts as a primary winding of the transformer, while the coil of the inductive coupler acts as a secondary winding of the transformer. Through which a data communication signal is coupled between the power line and the secondary coil. Which in turn is coupled to the communication device.

Another use for an inductive coupler is to place the inductive coupler around the phase or neutral line and induce high frequency energy generated by a Partial Discharge (PD). The synergy achieved by the combination of functions, including the continuous sensing of cable and insulator conditions, and data communication, is highly advantageous.

Capacitive couplers may also be used for partial discharge sensing and communication. However, high voltage capacitors are themselves susceptible to their intrinsic partial discharge, which is difficult to distinguish from partial discharges of cables or insulation. Thus, inductive couplers are more suitable for the task of the invention, although capacitive couplers can also be used for inducing partial discharges.

Fig. 1 illustrates a portion of an electrical distribution system 100, the electrical distribution system 100 having a plurality of components for detecting partial discharges on cables in the electrical distribution system 100. The power distribution system 100 includes a medium voltage underground cable, i.e., cable 105; distribution transformer 101, ground rod 118, an inductive coupler, i.e., coupler 120, and a Partial Discharge (PD) detector 130.

The coupler 120 includes a magnetic core (not shown) having a through hole (not shown). The coupler 120 functions as a transformer, and is located on the cable 105 such that the cable 105 passes through the through hole and functions as a primary coil of the coupler 120. Coupler 120 also includes a secondary coil having leads connected to PD detector 130 via cable 125. The cable 105 has a plurality of coaxial centerlines 110, the coaxial centerlines 110 are twisted together to form a strand 112, and the strand is connected to a ground rod 118 through the through hole.

The strands 112 pass through the vias resulting in no neutral current being induced into the secondary of the coupler. The net causes coupler 120 to induce currents in the phase lines of cable 105, including power frequency currents, as well as currents caused by partial discharges and "intrusions". The induced current is effective at the secondary coil of coupler 120 and thus also appears as a signal via cable 125.

As an alternative to the configuration of coupler 120 on cable 105, or when cable 105 does not include coaxial neutral 110, for example, as in the case of a multi-phase power cable, coupler 120 may be placed directly on phase conductor insulator 106. In this case, coupler 120 should preferably be enclosed within a strong and grounded conductive shield that is capable of directing the fault current to ground unless the insulation of the phase line has failed. Alternatively, the coupler 120 may also be placed over the strands 112.

Distribution transformer 101 is powered by cable 105 through elbow connector 107. Distribution transformer 101 has a neutral 115 connected to ground rod 118, and a secondary terminal 140. From the secondary terminal 140, the distribution transformer 101 supplies a low voltage having a power frequency. There is a substantially fixed phase relationship between the phase of the voltage (and current) on cable 105 and the phase of the low voltage on secondary terminal 140. This phase relationship changes slightly as the load on distribution transformer 101 changes.

PD detector 130 receives the induced current from coupler 120 through cable 125 and receives a low voltage having a power frequency from secondary terminal 140 through cable 145. This low voltage with power frequency provides a phase reference for PD detector 130. PD detector 130 processes the induced current from coupler 120 to detect a partial discharge in cable 105 and provides an output 135 that is connected to a communications link (not shown in fig. 1) allowing a continuous partial discharge monitoring data stream to be transmitted to a remote monitoring station (not shown in fig. 1).

Coupler 120 may also function as a power line communication data coupler. That is, cable 125 may also be connected to a communication device, such as a modem (not shown in fig. 1), and coupler 120 is used to couple data communication signals between cable 125 and the communication device.

Fig. 2 is a block diagram of PD detector 130. The PD detector 130 includes a switch 205, a peak current recorder 211, an amplifier 210, a microcontroller 240, a trigger circuit 270, and a group of components divided into five channels, i.e., channels CH1, CH2, CH3, CH4, and CH 5. Microcontroller 240 includes multiplexer 245, analog-to-digital converter (a/D)265, processor 250, and memory 255.

Processor 250 is provided with logic circuitry capable of responding to and executing instructions.

The memory 255 is a computer readable medium encoded with a computer program. In this regard, memory 255 stores data and instructions that may be read and executed by processor 250 for controlling the operation of processor 250. The memory 255 may be Random Access Memory (RAM), a hard drive, Read Only Memory (ROM), or a combination thereof. One component in memory 255 is a program module 260.

Program module 260 contains instructions that are read by processor 250 to cause processor 250 to perform the actions of the method used by PD detector 130. The term "module" as used herein is intended to denote the functional operation of an integrated assembly comprising an independent element or a plurality of sub-elements. Thus, program module 260 may be a single module or a plurality of modules that operate in cooperation with one another. Further, although program module 260 is described herein as being disposed in memory 255 and disposed in software, it may be disposed in any hardware (e.g., electronic circuitry), firmware, software, or combination thereof.

When program module 260 is indicated as having been loaded into memory 255, program module 260 may be provided on storage medium 275 for subsequent loading into memory 255. Storage medium 275 is also a computer readable medium encoded with a computer program and may be any conventional non-flash storage medium that stores program module 260 thereon in a tangible form. Storage medium 275 may be implemented in a medium such as a floppy disk, optical disk, magnetic tape, read only memory, optical storage medium, Universal Serial Bus (USB) flash drive, digital versatile disk, or hard disk drive. Alternatively, storage medium 275 may be provided on a random access memory, or other type of electronic storage device, located in a remote storage system (not shown) and coupled to memory 255 via a network (not shown).

Channel CH1 is composed of bandpass filter 215A, log detector 220A, peak detector 225A, and sample and hold (S/H) 230A. The configuration of each of the channels CH 2-CH 5 is similar to channel CH1, including bandpass filters 215B-215E, logarithmic detectors 220B-220E, peak detectors 225B-225E, and sample and hold units 230B-230E. In a preferred embodiment, the band pass filters 215A-215E are Surface Acoustic Wave (SAW) filters.

As shown, switch 205 is in an open state, but when it is closed, it couples signal 206 from cable 125 (see fig. 1) to amplifier 210. Signal 206 includes power frequency currents sensed through coupler 120 and currents caused by partial discharges and "intrusions". Amplifier 210 amplifies signal 206 and outputs signal 212. Therefore, signal 212 is amplified signal 206. The signal 212 is passed into each of the channels CH 1-CH 5.

Each bandpass filter 215A-215E is tuned to a different center frequency and has a wide bandwidth (e.g., 1 MHz). Thus, each of the channels CH1 through CH5, on a different frequency band, may "hear" the signal from the coupler 120. One or more of the bandpass filters (e.g., 215A) has a low center frequency for which the cable 105 does not significantly attenuate partial discharges, while the other bandpass filters (e.g., 215E) has a high center frequency for which attenuation per unit distance is significant. Thus, the band pass filters 215A-215E are preferably selected to avoid frequencies of known sources of intrusion, such as radio broadcasts.

Consider the channel CH 1. Bandpass filter 215A receives signal 212 and passes the frequency of signal 212 within the passband of bandpass filter 215A, resulting in signal 217A. The logarithmic detector 220A receives the signal 217A and converts it to a logarithmic form, referred to as signal 222A. Peak detector 225A, upon receiving signal 222A, detects its peak value and generates signal 227A. Sample and hold (S/H)230A samples and holds the peak of signal 227A to produce signal 235A.

The PD source may be close to PD detector 130 and produce a strong signal or the PD source may be far from coupler 120 and produce a weak PD signal that decays quickly as it propagates along cable 105. Thus, the amplitude of the PD signal can span a wide dynamic range. Accordingly, signals 206, 212, and 217A, may also cover a wide dynamic range. The logarithmic detector 220A is capable of processing signals 217A having a wide dynamic range. Nevertheless, for computational convenience, the parameters described herein may also be calculated from linearized, rather than logarithmized, amplitudes or other non-logarithmic functions that compress the amplitudes.

The channels CH 2-CH 5 operate similarly to channel CH1 and generate signals 235B-235E, respectively.

The trigger circuit 270 receives a low voltage having a power frequency, referred to as signal 269, via cable 145 and generates a power line synchronization signal 272. Alternatively, trigger circuit 270 may not receive signal 269, but may receive signal 212 and extract the power frequency component from signal 212. Regardless, the powerline synchronization signal 272 shows one pulse per cycle of the power frequency, such as one pulse per 60Hz cycle or one pulse per 50Hz cycle.

Microcontroller 240 receives signals 235A-235E and power line synchronization signal 272. Signals 235A-235E are input to multiplexer 245 and selectively passed from the output of multiplexer 245 to a/D265. A/D265 converts signals 235A-235E into digital signals and passes to memory 255. The operation of multiplexer 245, and the routing and selection of signals 235A-235E, will be described further below.

Microcontroller 240 controls switch 205 via control line 242 and simultaneously controls S/Hs 235A-235E. Signals 235A-235E are analog signals. The purpose of S/Hs 235A-235E is to maintain stable analog signal values of signals 235A-235E for a short period of time so that signals 235A-235E can be routed through multiplexer 245 and converted to an ordered digital signal by A/D265.

Consider signal 235A. The data comprising a batch of samples of signal 235A is digitized by a/D265. As described in detail below, assume that the batch of samples contains 90 values, each value representing a 4 degree phase interval, since a power frequency cycle period amounts to 360 degrees, such as: 90 degrees equals 360 degrees/4 degrees. Thus, each interval may be defined as a time block of 1/60/90 seconds, or 185.19 μ s (parts per million seconds) for a 60Hz power frequency, or 222.22 μ s (parts per million seconds) for a 50Hz power frequency. This time measurement is obtained from the powerline synchronization signal 272 by a timer (not shown) in the microcontroller 240. The first sample in the batch of samples of signal 235A, triggered by a logic transition in signal 272, appears after the positive-going cross-zero of signal 269, i.e., a low voltage with a power frequency. Each such data sample is proportional to the logarithm of the peak of the amplitude of signal 217A, i.e., the output of band pass filter 215A within each time block. These 90 sample values are sent to the memory 255. Thus, for a single power frequency cycle period, microcontroller 240 obtains 90 values from channel CH 1.

Microcontroller 240, and in particular processor 250, controls S/Hs 230A-230E and multiplexer 245 in accordance with program module 260, takes 90 values for each of channels CH1 through CH5, evaluates the values to characterize one of the PD pulses, and provides the evaluation via output 135. Microcontroller 240 may obtain these values in a single power frequency cycle period or, under certain circumstances, in multiple power frequency cycle periods, as described below.

Fig. 3 is a diagram of a power distribution system, namely: the system 300, a portion of which is illustrated, includes a network of couplers and PD detectors configured to detect partial discharges at a plurality of locations within the system 300. System 300 includes distribution transformers 303, 329 and 349, power cables 320, 340 and 355, couplers 302, 332 and 352, and PD detectors 304, 333 and 353. Distribution transformer 303, coupler 302 and PD detector 304 are arranged at 305. Distribution transformer 329, coupler 332 and PD detector 333 are arranged at 330. Distribution transformer 349, coupler 352 and PD detector 353 are arranged at 350. The system 330 also includes a monitoring station 365.

The primary distribution transformers 303, 329 and 349 are carried by cables 320, 340 and 355 in a series, which is powered by cable 355. Distribution transformer 329 takes power from cable 340 and delivers power downstream through cable 320.

Each PD detector 304, 333, and 353, operates similarly to PD detector 130, as described above, and provides results via outputs 310, 335, and 360, respectively.

The coupler 332 may be connected to a communication node (not shown) that is configured as a repeater. Such a node may be incorporated into PD detector 333. Similarly, coupler 302 may be connected to a communication node, which may be incorporated into PD detector 304, while coupler 352 may be connected to a communication node, which may be incorporated into PD detector 353.

The monitoring station 365 includes a processor 379, a user interface 375, and memory 380. Processor 370 is configured as logic circuitry that responds to and executes instructions. Memory 380 includes instructions in program module 385 that are readable by processor 370 and that, when read by processor 370, cause processor 370 to perform the actions of the method used by monitoring station 365. When program module 385 is shown loaded into memory 380, it can also be provided on storage medium 390 for subsequent loading into memory 380. Memory 380 may be used in any of the described embodiments for memory 255 and storage medium 390 may be used in any of the described embodiments for storage medium 275.

User interface 375 includes an input device such as a keyboard or voice recognition subsystem for enabling a user to communicate information and command selections to processor 370. User interface 375 also includes an output device such as a display or printer. A cursor control, such as a mouse, trackball, or joystick, allows the user to manipulate a cursor on the display to communicate other information and command selections to processor 370.

In system 300, since each of PD detectors 304, 333, and 353 are at different locations 305, 330, and 350, system 300 obtains an indication of the power line condition and detects each location 305, 330, and 350. The monitoring station 365 receives the outputs 310, 335 and 360 (fig. 3, coupled via the bubbles A, B and C) and, based thereon, determines whether the cable or device has been damaged. Through the user interface 375, the monitoring station 365 gives a report indicating where emergency repair or pre-maintenance is required.

Regardless of PD, an aging cable may also be subject to transient high current pulses, such as: the transient high current pulse ranges from 1 millisecond to 500 milliseconds, as may be expected to overcome an arc or other temporary short circuit condition of its own.

Referring again to fig. 2, such pulses will be measured by peak current recorder 211 (shown in detail in fig. 9). Peak current recorder 211 receives a voltage from coupler 120 through cable 125 that is proportional to the current in the phase line of cable 105. Peak current recorder 211 measures the maximum instantaneous current in cable 105 and provides the maximum instantaneous current as an input to multiplexer 245. The peak current recorder 211 is used during peak current measurement, which will be described in the following of the present invention.

In the following, we will consider a number of parameters for evaluating partial discharges.

Noise floor parameter

See also fig. 1 and 2

The bandpass frequencies of the bandpass filters 215A-215E are preferably selected to avoid as much as possible the selection of an effective communication or broadcast band, i.e., a significant source of intrusion. Nonetheless, the transmitter may be effective in the frequencies of the filter skirt, such as in the skirt of bandpass filter 215A, where the attenuation of the filter is insufficient to reduce the ingress amplitude to a negligible value.

The recording of data may be accomplished by a single cycle period of the power frequency or multiple cycle periods of the power frequency, wherein the cycle periods are not necessarily continuous. The recording of a single cycle period is designated herein as a "single cycle mode". The recording of a plurality of cycle periods, for example, 5 cycle periods, may be referred to as a "max hold", similar to a "max hold" acting on the spectrum analyzer, and therefore, the recording of a plurality of cycle periods is designated herein as a "max hold mode". In the "maximum hold mode," for each time block, the maximum amplitude for that time block is held among a number of samples recorded in a number of scans or cycles.

A technique for distinguishing line frequency synchronous partial discharge signals from other signals includes measuring the intrinsic noise floor of each channel output. For example, the intrinsic noise floor of channel CH1 represents the intrinsic noise generated by amplifier 210 that falls within the passband of bandpass filter 215A. This measurement is made when the switch 205 is in the open position, as shown in FIG. 2.

Processor 250 controls PD detector 130 and evaluates the noise floor according to instructions in program module 260.

Fig. 4 is a single channel for PD detector 130, namely: the inherent background noise of a single cycle period of the power frequency signal of channel CH1, namely: a plot of noise 405. Noise 405 represents the inherent noise of channel CH1, i.e., where switch 205 is open. The phase axis, the X-axis, represents the phase of the power frequency, ranging from 0 to 360 degrees representing a single cycle period, such as a 60 hertz (Hz) power frequency of 16.6 milliseconds (ms) per degree. Zero degrees represents the positive zero crossing phase of the voltage. The curve in fig. 4 represents 90 discrete data values, however, for clarity, a line is drawn connecting the 90 data values.

The bottom of the noise floor of the channel, referred to as the bottom line 410, is defined as the minimum of 90 samples. In fig. 4, the bottom line 410 is 36 decibels (dB), with 0dB representing a fixed power level determined by the system gain and detector characteristics. Since absolute signal levels are not used in this analysis, all signal levels refer to decibels (dB) above a fixed reference level.

Even if the coupler is not connected, i.e., switch 205 is open, the instantaneous output of each channel, e.g., signal 235A, floats above the value of the lowest bottom line 410. The amplitude of this floating depends mainly on the intrinsic noise of the channel and the bandwidth of its filter. For example, in FIG. 4, the noise 405 floats from a low point of 36dB to a high point of 43 dB. This float is referred to as the noise float 420 and has a value of 7dB, i.e. 43dB minus 36dB equals 7 dB. The threshold 430 is defined at a level beyond which intrinsic background noise is not expected. In fig. 4, the threshold 430 is at 44dB, which is slightly higher than the sum of the bottom line 410 and the noise float 420.

In the method described below, the noise floor is measured when the switch 205 is open, and the noise floor is also measured when the switch 205 is closed. The measurement performed when the switch 205 is open is referred to as the "initial baseline 410" and the measurement performed when the switch 205 is closed is referred to as the "current baseline 410".

In a preferred embodiment, the logarithmic detector is an envelope detector with a 0dB reference level set below the background noise level, in which case its output is unipolar, e.g., the output is generally positive. As such, the output of the log detector 220A is greater than or equal to zero, and thus, the signal 235A is also greater than or equal to zero. However, the noise floor is measured when switch 205 is closed, and if a PD pulse is present, it will further increase the amplitude of signal 235A in a positive direction within only a single time block. In most PD pulses, they occur at a time block that is different from the time block at which the lowest value 410 of the background noise occurs. As such, when switch 205 is closed, the measurement of baseline 410 should be approximately near the same low value regardless of the presence of PD pulses.

PD pulses do not typically occur at a particular phase interval of a power frequency phase. In other words, signals that are not synchronized to the power line often have a continuous carrier, thereby raising the bottom line 410. By periodically monitoring the value of baseline 410 to obtain the current baseline 410 value and comparing it to the initial baseline 410 value measured when switch 205 is open, microcontroller 240 can evaluate whether the signal on a particular channel contains a heavily intrusive signal during installation of PD detector 130 and subsequent monitoring periods.

The bottom line 410 may be different for each of the channels CH1 through CH5, and the bottom line 410 may also be different for a single channel of the single circulation mode and the maximum holding mode. The max hold mode can be sensitive to intrusions that are not synchronized with the power line frequency, especially when such intrusions are pulsed signals, rather than continuous signals.

When the switch 205 is open, an initial baseline 410 is measured for each of the channels CH 1-CH 5, and when the switch 205 is closed, a current baseline 410 is measured for each of the channels CH 1-CH 5.

For a given channel, if the current baseline 410 is higher than the initial baseline 410, this means some degree of intrusion into the channel. If the present baseline 410 is significantly higher than the initial baseline 410, e.g., greater than 5dB, the channel suffers from a large amount of intrusion and may be considered insensitive to the detection of partial discharges. Thus, to compensate for this higher bottom line 410, the pulse amplitude above the bottom line 410 may be increased. When the increase is of a considerable magnitude, for example 10dB, the channel will no longer be suitable for detecting partial discharges.

Thus, during the period when PD detector 130 is testing output and the internal noise level of PD detector 130 is of interest, processor 250 will evaluate bottom line 410. In the field configuration of PD detector 130, processor 250 evaluates baseline 410 to detect an intrusion on a particular channel, which if detected in large numbers will be removed and no longer qualified for local discharge measurement or localization.

Accordingly, the present invention provides a system comprising:

(i) a switch through which noise from the power line passes when it is closed, and through which noise cannot pass when it is open;

(ii) an amplifier, disposed downstream of the switch, that produces an amplified output;

(iii) a channel having:

(a) a filter having spectral components, within a particular frequency band, across which the amplified output is transmitted, thereby producing a filtered output; and

(b) a detector for detecting the filtered output values a plurality of times to generate a series of values; and (iv) a processor:

(a) the processor determining a minimum value of the series of values when the switch is open, thereby obtaining a first baseline value;

(b) the processor determining a minimum value of the series of values when the switch is closed, thereby obtaining a second baseline value; and

(c) a difference between the second baseline value and the first baseline value is determined, resulting in an excess value representing an excess value of power line noise that is greater than the amplifier noise.

Maximum peak amplitude

Fig. 5A is a graph of a signal at the output of a channel, where the signal includes signal partial discharge pulses, namely: PD pulse 540.

Fig. 5B is a graph of the data points in fig. 5A where the same signal exceeds threshold 430. Fig. 5B is derived from fig. 5A. Starting with fig. 5A, all data points are subtracted by the threshold 430, i.e., 44dB, and the negative value is set to zero. The PD pulse 540A then appears above the horizontal substrate as shown in fig. 5B.

Reference is also made to system 300 in fig. 3. For a given PD detector, no one or more PD pulses may occur during a single power frequency cycle.

Consider PD detector 333 at location 330. Assume that one of its channels is designated as channel M and the other of its channels is designated as channel N. Channels M and N have different center frequencies. Also considering the independent PD pulse occurring somewhere in system 300, the independent PD pulse may include some spectral energy that falls within the bandpass of channel M, and some spectral energy that falls within the bandpass of channel N.

Fig. 6A is a graph of two PD pulses recorded, which are captured in channel N of PD detector 333. These two pulses are designated as PD pulse 605 and PD pulse 610, respectively. PD pulse 605 occurs at a phase of about 80 degrees and PD pulse 610 occurs at a phase of about 265 degrees.

Fig. 6B is a plot of the same two PD pulse recordings as shown in fig. 6A, i.e., one pulse is approximately 80 degrees and the other pulse is approximately 265 degrees, captured in channel M of PD detector 330. In fig. 6B, these two pulses are designated as PD pulse 615 and PD pulse 620.

For clarity, each of these two PD pulses is captured in lane N and lane M. The first of these two PD pulses, occurring at about 80 degrees, is captured as PD pulse 605 in channel N and PD pulse 615 in channel M. The second of these two PD pulses, occurring at approximately 265 degrees, is captured as PD pulse 610 in lane N and as PD pulse 620 in lane M.

In channel N, fig. 6A, PD pulse 605 has a larger amplitude than PD pulse 610. In this case, the PD pulse 605 has a maximum amplitude, i.e., 24dB, referred to as Vpeak_NAnd recorded. The phase of this maximum amplitude sample, i.e. 80 degrees, is called Φ_ΝAnd is also recorded. Thus, the peak PD pulse of channel N is represented by the data pair (Φ)_Ν,Vpeak_N) And (4) showing. This same PD pulse is also the maximum amplitude sample in channel M (see fig. 6B). More specifically, PD pulse 615 has an amplitude of 10dB, measured by the data pair (Φ)_Μ,Vpeak_M) And (4) showing. Phi_ΜAnd phi_ΝEqual, meaning that two channels, M and N, record the same PD pulse in their respective frequency bands.

Now consider the PD detector 304 at location 305 and the PD detector 353 at location 350. Each of the three PD detectors 304, 333, and 353 includes channel M and channel N and records the maximum amplitude of the same phase, i.e., 80 degrees. By way of example, we assume that the center frequency of channel M is greater than the center frequency of channel N, and that their maximum peak amplitudes are as shown in Table 1.

TABLE 1 amplitude of the PDM maximum peak above background noise level (dB)

The data from each PD detector 304, 333, 353 is transmitted to the monitoring station 365 and the data in table 1 is processed in the monitoring station 365.

A Partial Discharge (PD) source in system 300 generates PD pulses whose maximum frequency component is attenuated to less than all PD pulses detected by PD detectors 304,333, and 353. Therefore, a first approximation of the PD source location performed at the monitoring station 365 is to (a) determine the maximum frequency channel for a set of neighboring PD detectors in the presence of a partial discharge; in this case, the set of PD detectors includes PD detectors 304, 333, and 353, and channel M is the channel with the largest frequency, (b) the largest peak is obtained from the largest frequency channel, in which case channel M has values of 5dB, 10dB, and 7dB, and (c) the PD detector with the largest of these values is selected as the PD detector closest to the PD source, in which case PD detector 333 at location 330 has the largest value, i.e., 10 dB. Therefore, the PD source is most likely near location 330.

Accordingly, the present invention provides a method comprising (a) measuring the maximum amplitude of a first spectral component of a partial discharge pulse induced on an electrical cable, (b) determining the phase of an electrical frequency signal on the electrical cable that produces the maximum amplitude of the first spectral component, (c) measuring the maximum amplitude of a second spectral component of the partial discharge pulse induced on the electrical cable, and (d) determining the location on the electrical cable at which the partial discharge pulse occurred based on the relationship between the maximum amplitude of the first spectral component and the maximum amplitude of the second spectral component.

Sum of Partial Discharge (PD)

The supplementary maximum peak amplitude parameter is a parameter indicating the discharge degree of the partial discharge, and is referred to herein as "PD sum" and is defined as: the sum of the amplitude values of the threshold 430 sampled over a single cycle period of a single channel is exceeded, see fig. 5B. In fig. 6A, the PD sum indicates the sum of the area under the PD pulse 605 and the area under the PD pulse 610. The PD sum will follow the increase in amplitude and duration of the two PD pulses.

Fig. 7 is a graph of a sampled signal, where each sample represents a separate time block, such as 4 degrees in phase, in a single channel of the PD detector for a single cycle period of the power frequency. For the first PD pulse, sample 705 is obtained at 80 degrees phase and sample 710 is obtained at 84 degrees phase. Thus, the first PD pulse has a duration longer than one time block. For the second PD pulse, sample 720 is obtained at a phase of 264 degrees and sample 715 is obtained at a phase of 268 degrees. Thus, the second PD pulse also has a duration longer than one time block. In FIG. 7, the PD sum is equal to the sum of the amplitude values of samples 705, 710, 715, and 720.

A PD at a given fault location typically produces zero, one, or at most two discharges during a given cycle period. However, the phase of the PD pulse may vary by some degrees in different cycles, for example, 86 degrees in the first cycle and 90 degrees in the second cycle. Maximum hold mode recordings over several cycle periods will capture multiple discharges occurring at different phase angles.

The PD summation is a more sensitive partial discharge indicator when using the maximum hold mode than the single cycle mode. This is because, in the single cycle mode, only the signal cycle of the power frequency is considered, and if a partial discharge does not occur during the single cycle, the partial discharge is not detected. However, even under intermittent partial discharge conditions, partial discharge occurs over only some cycle periods, not all cycle periods, and by using the maximum hold mode, the PD sum will not be zero.

As PD sources (e.g., cables, connectors, etc.) deteriorate, PD pulses will appear at lower instantaneous voltage levels, thereby creating a discharge profile over a wider phase range. Therefore, the sum of PD recorded in the max hold mode, initially, is a smaller value of the cable, and is discharged only at a voltage peak (e.g., 90 degrees or 270 degrees), but as the cable deteriorates, the phase spread of the discharge increases proportionally. Thus, the PD sum recorded in maximum hold mode is very sensitive to PD amplitude and phase spread.

During max hold mode recording, the PD sum may vary from a single digital value, such as an occasional single PD pulse, to thousands, such as a severe partial discharge in each cycle. One convenient way to express the PD sum in a user-friendly manner is to use "PD intensity", defined here as 20log (PD sum + 1). PD intensities typically range from 0 to about 70, and the PD intensity of the maximum hold mode is greater than the PD intensity recorded for the single cycle mode. By subtracting the compensation factor from the PD intensity measured in the maximum hold mode, the PD intensity measured in the maximum hold mode can be compared with the PD intensity measured in the single cycle mode.

For example, the PD strength in maximum hold mode may be 35, while the same signal is measured in single cycle mode for a PD strength of 21. By subtracting 14 from the maximum hold measurement, a value 21 is obtained which is equal to the value obtained in the single cycle mode.

Accordingly, the present invention provides a method comprising (a) measuring a peak amplitude of a spectral component of a Partial Discharge (PD) pulse sensed on a power cable by a plurality of phases of a cycle of a power signal on the power cable, (b) subtracting a background noise level from the peak amplitude to produce a composite amplitude, and (c) adding the composite amplitudes to produce a PD sum representing a magnitude of PD activity on the power cable.

Max hold-single cycle state machine

As described above, the recording of the PD pulses may be performed in the single cycle mode or the maximum hold mode. For power cables and devices that only begin to experience partial discharge, PD pulses do not occur during most cycles of the power frequency signal. Thus, a single cycle mode cannot detect any partial discharge, especially when the partial discharge is once every few minutes. Thus, initially, the advantage of using the maximum hold mode is that the highest signal point is detected over a number of cycle periods, e.g. 5 cycle periods.

As PD sources, such as power cables or devices, deteriorate, and partial discharges become more prevalent, the maximum hold mode may tend to show a large amount of discharge overlap recorded over different cycle periods. That is, PD pulses typically occur at the same phase in each of the multiple cycles and therefore overlap, so the PD sum of the max hold mode does not increase. However, as the PD source deteriorates further, the phase spread of these discharges generally increases. For example, the PD pulse initially occurs at 90 degrees and then extends to include a range from 86 degrees to 94 degrees. Because of this, most high amplitude spots tend to saturate the recording. In this case, a single cycle mode requires one switch. The PD detector preferably switches to max hold mode when the partial discharge is reduced thereafter.

Thus, if PD activity is low, the PD sum becomes smaller, and the max hold mode is more appropriate for recording the PD activity. If PD activity is high, the PD sum becomes large, and a single cycle mode is more appropriate for recording the PD activity. Automatic switching between the max hold mode and the single cycle mode increases the dynamic range of the PD detector and allows waveform analysis of the wave to be neither blanked nor saturated.

At some threshold level of the PD sum, a state machine is employed to "switch" between the two recording modes, wherein a hysteresis is introduced to prevent overshooting the transitions. To use comparable scales for both modes, the sum of the PDs measured in the max hold mode can be divided by a compensation factor.

Fig. 8 is a state transition diagram of a state machine that controls transitions between max hold mode and single cycle mode within PD detector 130 (shown in fig. 2). The state machine may be implemented within microcontroller 240, for example, by processor 250 operating in accordance with instructions in program modules 260.

Recording is performed in the maximum hold mode. At the start of the max hold mode recording, the PD sum is cleared to zero and recording is performed for a plurality of power frequency cycle periods, such as 5 cycle periods. After the recording, the PD sum is calculated from the data obtained during the recording. If the value of the PD sum divided by the compensation factor is greater than the value 500, the PD detector 130 switches to the single cycle mode. If the value of the PD sum divided by the compensation factor is less than or equal to 500, the PD detector 130 remains in the maximum hold state.

In the single cycle mode, at the start of recording, the PD sum is cleared to zero, and recording is performed for a single power frequency cycle. After recording, the PD sum is calculated from the data obtained during recording. If the PD sum is greater than or equal to 300, the PD detector 130 remains in the single-cycle mode. If the PD sum is less than 300, the PD detector 130 switches to the maximum hold mode.

Scale-linear or logarithmic of the raw data

The PD parameters as defined above, i.e. floor, noise float, maximum peak amplitude, PD sum and PD intensity, can be calculated both in single cycle mode and in maximum hold mode. The value of each data sample (e.g., sample of signal 235A) is proportional to the logarithm of the output (e.g., signal 217A) of the bandpass filter (e.g., bandpass filter 215A). Alternatively, these values may be directly proportional to the amplitude of the filter output (e.g., signal 217A).

Alarm criteria (in particular, alarm on rapid changes)

Referring again to fig. 3, a number of parameters as described above, namely, baseline, noise float, maximum peak amplitude, PD sum, and PD strength, are communicated to the monitoring station 365 to locate one or more points of failure on the feeder, and to take into account the schedule of routine maintenance as the cabling and equipment deteriorates, providing good protection against failure. By avoiding power outages and expensive replacement of large cables, which are so-called replacement-requiring cables, it is possible to continue to provide reliable service for many years.

It is also useful to arrange for line maintenance personnel to repair or replace cables and equipment at a significant expense each time, and therefore to define criteria for generating a true alarm. The alarm causes the manager to notice a particular location before powering down. One or more of the parameters described above, depending on whether an absolute threshold is exceeded, may generate an alarm, referred to as a "level alarm"; and another alarm, referred to as a "change alarm," may be generated independently based on whether the incremental threshold is exceeded.

At the stage of installing the PD detector, the Partial Discharge (PD) levels of the multiple channels are all stored in memory. A change alarm will be generated in response to an increase in PD parameters on one or more channels and a new PD level will be updated, storing the value in memory, well before the absolute threshold is reached, i.e. well before a level alarm occurs. The repeated occurrence of a change alarm within a short time may be understood to mean a rapidly aging cable and require attention from an administrator to the suspect location.

For example, if the PD intensity of any of channels 1 to 5 of PD detector 333 (shown in fig. 3) reaches or exceeds the value of 40, a level alarm will be sent to monitoring station 365. Separately, once the PD intensity increases significantly from the previously measured level (e.g., from 23 to 33, i.e., by 10), a change alert will be sent to the monitoring station 365.

Peak current measurement

As mentioned above, aged cables may also be subjected to very brief and large current pulses, such as: an arc or other temporary short circuit condition across itself may be expected and such a pulse may be measured by peak current recorder 211 (shown in fig. 2). Referring again to fig. 3, each of the PD detectors 304, 333, and 353 also includes a peak current recorder 211.

Each PD detector 304, 333, 353 will record the level of the current pulse using inductive couplers 302, 332, 352, PD detectors 304, 333, 353, and a data communication infrastructure connecting the PD detectors 304, 333, 353 to a monitoring station 365, such as by power line communication, wireless communication, or other medium communication. Once the current pulses exceed a certain level, such as the maximum rated current value of the cable, their respective microcontrollers 240 will send an alarm (via their outputs 135) to the monitoring station 365. The monitoring station 365 is in turn monitored, for example by a person, and the alarm sent to the monitoring station 365 is also evaluated to determine whether to perform power cable maintenance.

Considering a cable short-circuit situation, it will result in all consumers powered by the power cable being powered off. The peak current detector of each PD detector will measure the current of the cable at the PD detector location. At the upstream position of the short circuit, i.e. between the feeding point of the cable and the PD detector, a very high fault current will be measured, whereas downstream of the PD detector a normal load current will be measured until some protective device, such as a fuse, is disconnected from the cable from its feeding point. The type of current measured along the cable line may indicate the approximate location of the short circuit. However, if the PD detector is unable to pass this information before it loses power by itself due to a fault, this information may be lost.

In a preferred embodiment, each PD detector, such as PD detector 304, has its own back-up power source, such as a battery (not shown), which may ensure operation of PD detector 304 for a period of time sufficient for the PD detector to send an alarm to monitoring station 365. For multiple PD detectors 304, 333, and 353 located along the power line, the type of peak current will indicate the location of the short circuit fault prior to power outage. For example, a PD detector between the power input location and the fault location will measure the fault current, while a PD detector outside the fault range can only detect normal load current.

The present invention provides a method comprising (a) measuring a first magnitude of a first current at a first location on a power cable that exceeds a threshold, (b) measuring a second magnitude of a second current at a second location on the power cable that does not exceed the threshold, and (c) determining a location of a fault on the power cable based on a relationship between the first magnitude and the second magnitude. The threshold value may be set to a maximum rated current value of the power cable. The first and second currents may be continuous currents for a period of time, or they may be instantaneous currents for a period of time ranging from about 1 millisecond to about 500 milliseconds. Processor 370 performs the method in accordance with program module 385.

Fig. 9 is a schematic diagram of the peak current recorder 211 (see fig. 2). Peak current recorder 211 includes an amplifier and a low pass filter, i.e., an amplify-filter 905, to reject intrusions and partial discharges; a full wave rectifier 910 equally sensitive to either pole of the current peak; the signal is fed to an amplifier 915 of the input range of the a/D converter, and to a peak detector 920. The peak detector 920 has a capacitance C44 that acts as an analog memory and stores the most recently recorded maximum value, which is converted to a digital data memory.

In normal operation, the peak current value will be equal to the feeder current multiplied by the root mean square (rms) value of the crest factor, equal to the square root of two of the sinusoidal waveform. A typical cable rated current is 200 amps rms, with a normal peak current value of 282 amps. Any value that exceeds this value will sound an alarm to the monitoring station 365.

The techniques described herein are exemplary and should not be construed as any specific limitation on the present disclosure. It should be understood that various changes, combinations and modifications can be made by one skilled in the art. For example, the steps of the methods described herein are performed in any order, unless specified or dictated by the steps themselves. The present invention is intended to embrace all such alterations, modifications and variations that fall within the scope of the appended claims.

The terms "comprises" or "comprising" should be interpreted as specifying the presence of the stated features, integers, steps or components as referred to, but does not preclude the presence or addition of one or more other features, integers, steps or components, or groups thereof.

Claims (2)

1. A method of assessing noise and over-current on a power line, comprising:

measuring peak amplitudes of spectral components of Partial Discharge (PD) pulses sensed on a power cable at a plurality of phases of a cycle of a power signal on the power cable;

subtracting a background noise level from the peak amplitude to obtain a resultant amplitude;

adding the resulting amplitudes to obtain, for the first time, a first PD sum representing the PD activity on the power cable;

after a period of time, repeating the steps of measuring, subtracting, and adding, thereby producing a second PD sum for a second time; and

issuing an alarm if a difference between the second PD sum and the first PD sum is greater than a threshold.

2. A system for assessing noise and over-current on a power line, comprising:

means for measuring peak amplitudes of spectral components of partial discharge PD pulses sensed on a power cable at a plurality of phases of a cycle of a power signal on the power cable;

means for subtracting the background noise level from said peak amplitude to obtain a resultant amplitude;

means for summing the resulting amplitudes to obtain, for a first time, a first PD sum representing PD activity on the power cable;

means for repeating the steps of measuring, subtracting, and adding after a period of time, thereby producing a second PD sum for a second time; and

means for issuing an alarm if a difference between the second PD sum and the first PD sum is greater than a threshold.