Voltage Detection of Utility Service Disturbances

The invention detects utility service disturbances that may affect critical loads, such as sensitive electronic equipment. More specifically, the invention relates to detecting power outages, momentary interruptions, voltage sags and swells and other disturbances of commercial AC power supplied by utilities.

Utility power disturbances and momentary power interruptions degrade power quality and can adversely affect the operation of sensitive electronic equipment. Even momentary interruptions of just a few cycles can result in unscheduled controller resets, shutdown of a factory assembly lines, loss of computer data, etc. Thus, loss of electric power for only a short time can be detrimental by factory downtime, product damage and reduced productivity, etc.

The solution is to provide a system that continuously monitors power quality for power outages, momentary interruptions, voltage sags and swells, and switches sensitive loads from the utility service to an alternate standby power source when a power disturbance is detected. It is desirable that the switching take place as quickly as possible so sensitive electronics in the critical load are not affected.

Prior art methods of detecting voltage fluctuations and momentary interruptions include the template method and the three-phase to two-phase transformation methods. These methods can be inaccurate and are not as robust as the present invention.

The invention provides a practical method of quickly and accurately detecting utility service disturbances. It uses a modified RMS processing technique, and is especially well-suited for use in uninterruptable power supplies.

More specifically, the method involves the use of an isolation transformer to provide an analog signal representing line voltage in the main electrical power supply line. Digital samples are taken to generate a plurality of instantaneous voltage magnitude values (V.sub.i) at a preselected sampling rate. The preferred sampling rate is approximately 40 cycles per half-cycle of the main power supply waveform. The instantaneous voltage magnitude values are each squared in real time to create respective instantaneous squared voltage values (Vi).sup.2 for each sampling period. A running average of the (Vi).sup.2 value is generated by accumulating the instantaneous squared voltage values (Vi).sup.2 for a preselected period of time which is referred herein as an accumulation window of time. The running average, referred to herein as the running voltage value RVV, is adjusted in real time by subtracting the instantaneous squared voltage value (V.sub.i-41).sup.2 for the sampling period immediately prior to the start of the accumulation window and by adding the instantaneous squared voltage value (V.sub.1).sup.2 for the most recent sampling period. The running voltage value is then compared to a preselected under-voltage threshold value K.sub.min, and also preferably to a preselected over-voltage threshold value K.sub.max. If the running voltage value sags to or below the preselected under-voltage threshold value K.sub.min or swells to or exceeds the preselected over-voltage threshold value K.sub.max, a disturbance detection signal is generated to switch the system to the standby electrical AC power supply.

The preselected under-voltage threshold value K.sub.min and the preselected over-voltage threshold value K.sub.max are selected in accordance with appropriate tolerance levels for the quality of power supplied. Preferably, the threshold values are determined by multiplying the number of samples per accumulation window times the square of the normalized voltage tolerance.

In order to promote system response, it is desirable that the accumulation window of time be as short as reasonable. In order to eliminate AC ripple effects in the running voltage value, it is desirable that the running voltage value be generated by accumulating the instantaneous squared voltage values (V.sub.i).sup.2 over integer multiples for the half-cycle of the main power supply waveform (i.e. the accumulation window approximates an integer multiple of the half-cycle for the main power supply waveform). Otherwise, the preselected under-voltage threshold value K.sub.min and the preselected over-voltage threshold value K.sub.max may have to be adjusted to account for AC ripple in the running voltage value. Therefore, it is preferred that the accumulation window be defined as one half-cycle for the main power supply waveform.

The invention can be carried out in a system in which the main electrical AC power supply is single-phase, and can also be carried out in a system where the main electrical power supply is three-phase, as long as each phase is treated independently from the other phases.

In the preferred embodiment of the invention, the analog isolated voltage signal from the isolation transformer is scaled and rectified to generate a half wave rectified sinusoidal waveform having a preselected peak voltage (e.g. 0 to +4 volts). An A/D converter located in a system microprocessor converts the analog half wave rectified sinusoidal waveform into digital values representing the instantaneous voltage magnitude values V.sub.i at the respective sampling period which are used by the microprocessor to generate the running voltage value. The scaling and rectifying of the analog isolated voltage signal is desirable so that reasonable resolution can be achieved by an 8-bit A/D converter. Scaling and rectifying make this possible because the A/D converter need not be concerned about negative values and 0.

The invention provides effective protection against utility service disturbances by accurately and quickly detecting under-voltage and over-voltage conditions. The invention does not require synchronization, and is robust in the presence of harmonics, phase variations and frequency variations. In addition, the preferred embodiment of the invention operates in real time without requiring substantial processing power (e.g. an 8-bit microprocessor is useable).

Various other features, objects and advantages of the invention will be made apparent from the following drawings and detailed description.

FIG. 1A is an illustration comparing a normal AC voltage waveform to both a sagging AC voltage waveform and a swelled AC voltage waveform.

FIG. 1B is an illustration of an AC voltage waveform during a power outage.

FIG. 2 is a block diagram showing an uninterruptable power supply (UPS) for detecting utility service disturbances and transferring power from utility service to a standby power source after detecting a power disturbance.

FIG. 3 is an illustration of the AC output power waveform fed to a critical load upon switching to a standby AC power source.

FIG. 4 is a flowchart illustrating a voltage detection technique in accordance with the invention.

FIG. 5 is a schematic drawing showing a control board for the UPS shown in FIG. 2.

In FIG. 1A, the waveform labelled with reference numeral 10 illustrates one cycle of a normal AC voltage input signal 10 from a utility power line (i.e. 120 VAC, 60 Hz sine wave). The waveform labelled with reference numeral 11A illustrates one cycle of a swelled voltage input signal 11A, and the waveform labelled with reference numeral 11B illustrates one cycle of a sagging AC voltage input signal 11B. FIG. 1B illustrates the AC input signal of FIG. 1A during a power outage. In FIG. 1B, voltage is lost in the first half-cycle prior to reaching the peak of the sinusoidal waveform, and the voltage falls to zero. The voltage input signal 12 does not typically drop to zero immediately because the circuit maintains residual voltage for a short period of time.

Referring to FIG. 2, an uninterruptable power supply (UPS) 18 in accordance with the invention monitors, detects and transfers power from a main utility service 19 to an alternate standby power source 32 when a utility power disturbance is detected (e.g. when a power outage, a momentary interruption, or significant voltage sag or swell is detected). The UPS 18 is able to switch sensitive electrical loads 21 (i.e. critical load) from the utility service 19 to the standby power system 32 in approximately one-quarter cycle.

The UPS 18 includes a static switch assembly 20, an isolation transformer 34 as well as the standby power system 32. Commercial AC power enters the UPS 18 through line 19 into the static switch assembly 20. The static switch assembly 20 contains an input circuit breaker 22, a static switch 24 and an output circuit breaker 28. The static switch 22 includes a set of SCRs (silicon controlled rectifiers, not shown) as is well known in the art. Under normal operating conditions, input circuit breaker 22 and output circuit breaker 28 are closed and the SCRs within static switch 24 are on. Therefore, under normal operating conditions, electrical AC power is supplied from utility power line 19 through the static switch assembly 20 to the critical load 21. The static switch assembly 20 also has a static switch bypass circuit breaker 26 that is normally open under normal operating conditions but can be actuated to override the UPS 18.

The input circuit breaker 22, the static switch 24, the output circuit breaker 28 and the bypass circuit breaker 26 are all monitored and controlled by a system control 30 for the standby power system 32. The standby power system 32 contains the system control 30, battery modules 40, an inverter bridge 38, a battery charger 36, and associated circuitry. The battery charger 36 converts AC power in line 35 from a bidirectional isolation transformer 34 to DC power. The isolation transformer 34 protects the UPS 32 from common mode and normal mode transients. DC power flows from the battery charger 36 to battery modules 40 in order to charge batteries 40. The inverter bridge 38 converts DC power from the battery module 40 into AC power when a power disturbance is detected, thus maintaining power to the critical load 21. AC power from the inverter bridge 38 is transmitted to an isolation transformer 34 which steps up the voltage of the AC power and transmits AC power to the static switch assembly 20 in the line between the static switch 24 and the output circuit breaker 28 (i.e. tap location 27).

The isolation transformer 34 generates an isolated voltage signal in line 35 in response to the voltage in the utility AC power line at tap location 27. The isolated voltage signal 35 is used to provide AC power to the battery charger 36, and is also used to provide analog input through line 35A to the system controller 30. If the system controller 30 detects a utility service disturbance in the isolated voltage signal, the static switch 24 SCRs are turned off and the inverter bridge 38 is turned on to switch the load to the standby power system 32.

The waveform 16 in FIG. 3 illustrates the output signal waveform to the critical load 21 upon switching to the standby power system 32. The voltage disturbance is detected at point 16A of FIG. 3. The SCRs of static switch 24 must be commutated to be turned off. This is accomplished by bringing the load voltage above the utility voltage for a short period of time, approximately 1 millisecond, to commutate the SCRs. Thus, after a slight electronic delay (100 milliseconds), the voltage rises and overshoots the utility voltage for a period of about one millisecond (1 ms) to a peak voltage at point 16B in FIG. 3 to turn off the stator switch 26. Once the SCRs of static switch 24 have been commutated (i.e. turned off), the voltage returns to normal as shown at point 16C in FIG. 3. It takes approximately one millisecond for the voltage to drop down from the peak value 16B to the normal operating value 16C of the utility voltage. The time period from detection until transfer over to the standby power system 32 from the utility takes approximately 2 milliseconds. Thus, if the commercial power fails, the inverter bridge 38 turns on and begins supplying AC power to the critical load 21 after a delay of approximately 2 milliseconds (i.e. approximately 1/4 of the half-cycle for the main power voltage supply waveform).

FIG. 4 is a flowchart illustrating the preferred voltage disturbance detection technique implemented by the system controller 30. In FIG. 4, the disturbance detection technique is illustrated for detecting a disturbance in a main electrical AC power supply that is single-phase. The technique can also be used to detect disturbances when the main electrical power supply is three-phase by performing each of the steps described in FIG. 4 independently for each phase.

In FIG. 4, block 100 depicts an analog voltage signal input which is generated by the isolation transformer 34 in response to the voltage in the main electrical AC power line 27. The system controller 30 samples the analog voltage signal input at a preselected sampling rate (e.g. 40 cycles per waveform half-cycle) to generate digital data representing the instantaneous voltage magnitude for each respective sampling period, V.sub.i, as depicted by block 102. Each of the instantaneous voltage magnitude values V.sub.i are squared in real time to create respective instantaneous squared voltage values (V.sub.i).sup.2, block 104. The respective instantaneous squared voltage values (V.sub.i).sup.2 are accumulated, block 106, to generate a running voltage value (RVV), which is a running average of the instantaneous squared voltage values (V.sub.i).sup.2 for a preselected number of sampling periods (e.g. 40): ##EQU1## The preselected number of sampling periods (e.g. 40) is referred herein as an accumulation window of time. It is preferred that the accumulation window for determining the running voltage value RVV be generated by accumulating the instantaneous squared voltage values, block 104, over integer multiples for the half-cycle of the main power supply waveform, voltage tap 27. Although the invention does not require selecting the accumulation window in accordance with integer multiples of the half-cycle, AC ripples will appear in the running voltage value, block 106, unless it is selected in this manner. In order to promote system response, it is preferred that the accumulation window for the running voltage value RVV be one (1) half-cycle of the main power supply waveform. Using an accumulation window of two half-cycles, or even three half-cycles may be desirable, but system response will be compromised.

The running voltage value RVV is adjusted in real time during each sampling period by subtracting the instantaneous squared voltage value (V.sub.i-41).sup.2 for the sampling period immediately prior to the start of the accumulation window for the running voltage value RVV, block 108, and by adding the instantaneous squared voltage value for the most recent sampling period (V.sub.i).sup.2 to the running voltage value RVV, block 110. The running voltage value RVV is compared to a preselected under-voltage threshold value K.sub.min, block 112. It is preferred that the running voltage value RVV be compared to the under-voltage threshold value K.sub.min every sampling period, however, in some circumstances it may be desirable to compare intermittently. If the running voltage value RVV sags to or below the under-voltage threshold value K.sub.min, the system controller 30 generates a disturbance detection signal, block 114. In a similar manner, the running voltage value RVV is compared to an over-voltage threshold value K.sub.max, block 116. If the running voltage value RVV swells to or exceeds the over-voltage threshold value K.sub.max, the system controller 30 generates a disturbance detection signal, block 114.

The under-voltage threshold value K.sub.min, block 112, and the over-voltage threshold value K.sub.max, block 116, are preselected and are chosen primarily in accordance with the operating tolerance of the critical load 21, FIG. 2. In particular, the threshold values K.sub.min and K.sub.max can be determined by multiplying the number of sampling periods per accumulation window (e.g. 40) times the square of a normalized voltage tolerance level (e.g. K.sub.min approximately equal to (0.9).sup.2 ; K.sub.max approximately equal to (1.1).sup.2). It may be desirable to slightly adjust the threshold values K.sub.min and K.sub.max because of AC ripples in the running voltage value RVV. In practical systems, AC ripples are normally present in the running voltage value RVV, block 106, because it is difficult to exactly match the accumulation window to a precise integer multiple of the half-cycle for the main power supply waveform.

Referring now to FIG. 2 in addition to FIG. 4, at system start-up, the system controller 30 does not turn on the static switch 24 until the system controller 30 has run the method described in FIG. 4 for at least one full accumulation window of time.

FIG. 5 illustrates a schematic of a control board 42 within the system control module 30 of the UPS 32. Input power from the utility enters the control board 32 through phase A input 44, phase B input 46, and phase C input 48. All three phases of the utility power enter through inputs 44, 46, 48 and into voltage scaling circuitry 50 where the input voltage waveforms 44, 46, 48 are scaled down to a 4-volt peak sine wave. The 4-volt peak sine wave input voltages enter analog absolute value circuitry 52 and a rectifier 54. The analog absolute value circuitry 52 and the rectifier 54 converts the 4-volt peak sine wave input power voltages to a voltage waveform that consists entirely of positive values, i.e. 0 to +4 volt half-wave rectified sine waves. The 0 to +4 volt half wave rectified sine waveform enters microprocessor 56 through A/D inputs 58, 60, 62. The microprocessor is preferably an Intel 87C51GB. The microprocessor is an 8-bit microprocessor which performs the steps illustrated in FIG. 4. The software code for instructing the microprocessor is embedded in firmware within 32K EPROM 64. An assembly language listing of the subroutines used to instruct the microprocessor 56 is set forth in Appendix A.

Once a utility service disturbance is detected by the control board 42, a low level digital output signal is transmitted from microprocessor 56 as a disturbance detection signal 66 to signify the presence of a disturbance in the main power supply line 27. Microprocessor 56 also includes six parallel outputs for signalling over-voltages or under-voltages on each of phases A, B and C of the power input signals 44, 46, 48. The phase A over-voltage is in line 68; the phase A under-voltage is in line 70; the phase B over-voltage is in line 72; the phase B under-voltage is in line 74; the phase C over-voltage is in line 76; and the phase C under-voltage is in line 78. These lines 68, 70, 72, 74, 76 and 78 are electrically connected to six separate LEDs to warn an operator of a disturbance on phases A, B or C. The output on lines 68, 70, 72, 74, 76, 78 is also transmitted to a parallel to serial converter 80. The parallel to serial converter 80 transforms the parallel inputs 68, 70, 72, 74, 76, 78 to a single serial output 92 which is sent to a master control.

A 12 MHz crystal oscillator 82 supplies a clock signal to the microprocessor 56. A precision +5 volt DC reference 84 is supplied to microprocessor 56 for reference. A DC power supply monitor 86 monitors DC power supplied to the control board 42, and is electrically connected to microprocessor 56. Input of the data and software code into microprocessor 56 is accomplished by address latch 88 which controls the flow of data and code into and out of microprocessor 56 through data bus 90.

As mentioned previously, the analog voltage waveform for phases A, B and C are input into the control board 42 and are transmitted sequentially to the voltage scaling circuitry 50, the absolute value circuitry 52 and the rectifying circuitry 54 to result in a 0 to +4 volt half wave rectified sine wave. The 0 to +4 volt half wave rectified sine waveform is transmitted to an A/D converter in microprocessor 56. Because the microprocessor 56 does not need to account for negative voltage amplitudes or zero crossings within the 0 to 4 volt half wave rectified sine wave, reasonable resolution can be achieved using an 8-bit microprocessor.

The invention has been shown in the drawings and described herein as a voltage detector for detecting utility service disturbances in conjunction with an uninterruptable power supply, but the method of detection may be useful in other applications not involving uninterruptable power supplies. For instance, a disturbance detector in accordance with the invention may be used in other voltage monitoring applications.

While the invention has been described with reference to a preferred embodiment, those skilled in the art will appreciate that certain substitutions, alterations and omissions may be made without departing from the spirit thereof. Accordingly, the foregoing description is meant to be exemplary only, and should not be deemed limitative on the scope of the invention set forth with following claims.