GB2073980A - Speed signal determination - Google Patents

Abstract

A signal indicative of speed of displacement, for example in respect of steam turbines, has to be repeatedly obtained for control purposes typically in periods of 0.01 sec. with good resolution over substantially the whole speed range. Speed is derived from a count of time pulses generated in a period and a count of displacement pulses (derived from two relatively movable parts) equal to a registered number to determine the duration of the period. The duration of the next successive period is achieved by determining whether the count of time pulses is within, below or above a range set by lower and upper limits and respectively maintaining, increasing or decreasing said registered number accordingly. An overspeed facility is provided. Acceleration is determined using successive speed measurements. Linear speeds and accelerations can also be determined. <IMAGE>

Description

SPECIFICATION Speed signal determination The invention relates to obtaining a repeatedly updated signal indicative of speed of displacement particularly, though not exclusively, for the control of turbine power generators.

It has already been proposed in UK patent specify cation No. 1385452 to use digital techniques in a speed responsive system for controlling a prime mover such as a steam turbine. In that proposal it is recognised that the determination of a speed signal (a so-called "fine" speed signal) by monitoring the total time required for a predetermined number of motion displacement pulses to occur, whilst giving acceptable resolution in the speed determination in a band of relatively high speeds, cannot be used at lower speeds owing to the increase in time that is required for the predetermined number of displacement pulses to occur.

Accordingly, in that proposal speed is also determined by monitoring the number of displacement pulses occurring in a predetermined time (i.e. 0.1 seconds) to give a so-called "coarse" speed signal. If the "coarse" speed signal is below a pre-determined switch-over speed value (of say 1600 revolutions per minute, RPM), the "coarse" signal is adopted for control purposes, as the speed signal. Alternatively, if the "coarse" speed signal is equal to or above the switch-over value, the "fine" speed signal is adopted as the speed signal.

The resolution of the "fine" speed signal at synchronous speed, i.e. 3600, is t1 RPM (it varies slightly depending on the RPM at which it is calculated) and the resolution of the "coarse" speed signal is + 5 RPM over the whole range.

According to the invention, a method of obtaining a repeatedly updated signal indicative of speed of displacement comprises generating displacement pulses and counting a number of the same equal to a registered number of pulses to determine a period during which high frequency time pulses are counted and deriving a speed signal from the count of time pulses taken with said registered number of pulses, determining whether the count of time pulses is within, below or above a range set by lower and upper limits, and respectively maintaining, increasing or decreasing said registered number two determine the next successive period.

Apparatus, also according to the invention, for obtaining a repeatedly updated signal indicative of speed displacement comprises a generator for generating displacement pulses, a programmable register, a generator for generating high frequency time pulses, a counter for counting said time pulses for a period the duration of which is determined by the time taken to count a number of said displacement pulses equal to a number registered in the register and means for processing the outputs from the register and the counter to produce a signal indicative of said speed of displacement and control means operable according to whether the count of time pulses is within, below or above a range set by lower and upper limits to programme said register to respctively maintain, increase or decrease said number registered in the register.

Preferably, the method and apparatus include an overspeed facility.

A method and apparatus will now be described by way of example to illustrate the invention with reference to the accompanying drawings, of which: Figure 1 is a schematic block diagram of a module containing equipment for producing signals by which speed and/or acceleration can be determined by processing means; and Figure 2 is a flow diagram showing the steps involved in processing the displacement pulse signals and the time pulse signals.

The particular system described with reference to the drawings is adapted for use in measuring the speed and acceleration of a steam turbine used for power generation. At the synchronous speed of the turbine, e.g. 3000 RPM, it is essential that the speed and acceleration are continuously updated to allow real-time control of the turbine to avoid possible shut-down of the electricity supply network. Consequently, there is a requirement for very small sampling periods to be used, i.e. sampling periods of the order of 0.01 seconds, in order to allow early detection of faults and, consequently, early corrective steps to be taken.

Referring now to Figure 1, a toothed wheel having sixty-four teeth (not shown) is mounted on a turbine shaft and a probe (not shown) is mounted adjacent to the wheel to provide two displacement pulses per tooth to a module 10. A signal proportional to the period required to produce a number of displacement pulses equal to a registered number is passed from the module 10 to a processor (not shown) where speed, acceleration and other calculations are performed.

The wheel may have any suitable number of teeth but advantageously has 2" teeth, which conveniently allows the registered number to be halved or doubled between the limits of one and sixty-four as discussed more fully below.

The probe, which typically is a permanent magnet reluctance probe, provides a sinusoidal signal which is fed through an isolation transformer 12 to a zero-crossing detector 14 with hysteresis. The resultant square wave from the detector 14 is passed directly to one input 18 of a two-input exclusive OR gate 16 and via a delay network to the other input 20.

The output of the exclusive OR gate 16 consists of a small displacement pulse of finite length (owing to the effect of the delay network 22) coincident with each change of state of the input square wave.

The output of the OR gate 16 is fed to a singlepulse detection circuit 24 (consisting of TR1, R1 and Cl).

The output of the OR gate 16 is also fed to a programmable displacement pulse register 26 consisting of a 4-bit counter 28, a binary-to-decimal converter 30 used as a decoder and receiving a 3-bit code from the counter 28 and an 8-bit binary down counter 32. The pulses from the OR gate 16 cause the down counter 32 to count down to zero during a measuring period from a registered number which is one of the numbers 1, 2,4,8, 16,32 or 64. The counter 28 is programmed by a tooth count control unit 34 which instructs the counter 28 when, and in which sense, to count. The register 26 instructs the conrol unit 34 when the minimum number of pulses (i.e. one), or the maximum number of pulses (i.e. 64) are loaded into the down counter 32.The counter 28 counts numbers 0 to 6, being the powers to the base 2 of the numbers 1,2,4,8, 16, 32and 64, respectively, presenting the numbers of displacement pulses to be loaded into the counter 32 (the decoder 30 decoding the output of the counter 28 accordingly) representing the effective number of teeth to which the probe responds in any given revolution. At synchronous speed, the shaft makes 0.5 revolution in the required sampling period of 0.01 seconds.

Consequently, in this particular application, the maximum number of displacement pulses used to determine the measuring period is sixty-four; A system clock 36, of the quartz crystal type of pulse generator, provides time pulses at a frequency of 1 MHz. The system clock 36 feeds time pulses to both a system control unit 38 (to allow clocking of various items for resetting etc.) and, through an AND gate operable by the system control unit 38, to a 17-bit time pulse counter 40. The counter 40 is linked to a 16-bit shift register 42. The counter also supplies signals to the tooth control 34 when the time pulse count reaches lower and upper limits. The limits in this instance are 213 and 214, respectively. If the time pulse count goes over-range, i.e. above 216, then the counter 40 sets a latch 44.

A block of output latches 46 is also provided. The latches 46 receive signals from the detection circuit 24 to indicate that the system is working; from the counter 28 of the displacement pulse register 26; and from the over-range latch 44. The latches 46 feed signals representing NUMBER READY (i.e. time pulse count), INTERRUPT, POWER OF TOOTH COUNT and OVER-RANGE to the processor (not shown).

The apparatus also has an overspeed facility provided buy a programmable 16-bit down counter 48 which can operate a relay 50 to shut down the turbine in the event of overspeed. The output from the down counter 48 passes to an AND gate together with the output from the counter 28. The down counter 48 is pre-programmed with a number representing overspeed and is decremented by time pulses from the system clock 36 but fed through the time pulse counter 40 to divide the pulses by eight.

Provided the counter 48 decrements to zero, a zero output to the AND gate ensures that a latch 52 does not trip a relay 50.

The operation of the system will now be described.

At the start of a cycle the down counter 32 is loaded with a number which is dependent on the number of displacement pulses counted in the previous cycle, as explained below; at the same time the time pulse counter 40 is reset. Each pulse from the OR gate 16 decrements the down counter 32, until it reaches zero at which time the system clock pulses to the time pulse counter 40 are stopped. That counter 40 now contains a count proportional to the period needed to count a number of displacement pulses equal to the number registered in the counter 32.

The tooth count control unit 34 tries to ensure that the number in the time pulse counter 40 at the end of each measuring cycle is between two limits, in this case 213 and 214. It does this by retaining or changing the number registered in the down counter 32 at the start of a cycle according to the number of time pulses recorded by the counter 40 during the previous cycle.The unit 34 operates as follows: a) if the counter 40 peviously recorded between 213 and 214 time pulses, the number aleady registered in counter 28 is retained, the unit 34 inhibiting any count action by the counter 28 and the same number as previously registered is re-loaded into the counter 32 when it is re-set; b) if the previous number of time pulses was less than 213 and the number registered in the counter 32 was less than sixty-four then the control unit 34 operates to cause the counter 28 to count upwards to double the number registered in the counter 32 to be decremented in the next cycle;; c) if the previous number of time pulses was greater than 214 and the number registered in the counter 32 was greater than one, then the control unit 34 operates to cause the counter 28 to count downwards to halve the number registered in the counter 32 to be decremented in the next cycle.

The overspeed counter 48 is pre-programmed by having a number loaded into it representing a maximum acceptable speed of rotation of the turbine shaft. In each cycle, provided the shaft speed is below that set overspeed, the counter 48 always decrements to zero and so the AND gate receives a zero output from the counter. Thus, regardless of its other input, the AND gate passes a zero output to the latch 52 and no overspeed action is initiated.

If the shaft speed reaches or exceeds the set overspeed value, the time pulses received by the counter 48 are insufficient to decrement it to zero which causes a positive output to pass to the AND gate. If the output from the counter 28 is at its maximum of six, representing a maximum number of sixty-four displacement pulses to be counted, the AND gate receives a positive output from the counter 28 and the gate then gives a positive output to the latch 52.

The latch 52 is thus enabled (see paragraph (d) below) to cause the relay 50 to operate the overspeed contact to initiate appropriate control of the turbine.

The control unit 38 provides pulses for timing the sequence of operation of the system. The sequence is as follows: a) the down counter 32 decrements to zero at the end of a measuring period; b) on the reset rising edge of the system clock 36 the system clock time pulses to the time pulse counter 40 are inhibited; c) a predetermined number of system clock pulses are counted to achieve a delay during which time the time pulse counter 40 output stabilises; d) a pulse OS is issued which is normally ineffective but which enables the overspeed latch 52 to function as described above in the event of over speed; e) a pulse Q6 is issued which commands the re-loading of the overspeed counter 48 ready for the next period;; f) a pulse Q7 is issued which commands the re-loading of the shift register 42 associated with the time pulse counter 40 with the count recorded during the previous period, and clocks the latches 46 to record the conditions of NUMBER READY, POW ER OF TOOTH COUNT and OVERRAN GO during the previous period; g) a pulse Q8 is issued which resets the time pulse counter 40 and the overrange latch and clocks the counter 28 which counts upwards or downwards or is inactive depending on the outputs from the tooth count control unit 34 as described above; h) a pulse Q6 is issued which commands loading of the down counter 32 ready for the next period and resets the tooth count control unit 34.The loading of the down counter 32 with a new number inhibits the zero output to the system control 38; i) on the next rising edge of the system clock 36, time pulses pass to the down counter 40 and the system control unit 36 resets.

Figure 2 is now referred to.

The processor (not shown) is used to calculate speed from the number recorded by the time pulse counter 40. The programme steps are as follows: a) following clocking of the latches 46 by the pulse Q7 as described above, the NUMBER READY/INTER RUPT signal is used to interrupt the processor and initiate the speed programme; b) the POWER OF TOOTH COUNT, OVERRAN GO and SERIAL OUTPUT are read by the processor (not shown). (i) If the OVERRANGE is active the speed is set to zero, a strobe pulse is issued to clear the latches 44 and the rest of the programme is bypassed; (ii) If the OVERRANGE is not active the processor issues fifteen strobe pulse reading the SERIAL OUTPUT after each pulse.The first strobe pulse clears the latches 46. in all, the processor obtains 16 bits from the shift register 42 which are equivalent to those recorded by the time pulse counter 40 during the previous period; c) before using this number to calculate speed, a correction equivalent to the number of time pulses lost while the system control unit 38 was in operation is added to the data (in this instance 17 time pulses). The result is directly proportional to the time taken for the pevious period.If desired, this could be avoided by duplicating the apparatus, so that one set may start as soon as the previous period had finished; d) the processor (not shown) executes a division using a constant as the numerator and the timedependent data as the denominator to obtain a rate for the previous period; and e) the POWER OF TOOTH COUNT is used to scale the rate nd provide an output proportionl to speed.

Successive outputs from step (e) can be used to calculate the acceleration of the shaft, which is used for control purposes, such as for detecting loss of load and feedback for damping the controls for speed changing, for example.

The number of teeth effective to produce displacement pulses depends upon the nominal rotating speed of the shaft and the desired rate at which updates of the speed computation are to be carried out; this leads for example in a system used with a power generator as described above to one update per half-revolution at synchronous speed (3000 RPM). If no mechanical inaccuracy were present, then predictably each computation would, under constant-speed conditions, produce a result identical with its immediate predecessor. However, if a small tooth pitch error is present, two successive readings of speed will not be the same; the readings obtained from one half of the wheel will be consistantly greater than those obtained from the other half of the wheel, whereas the true speed will be mid-way between the two readings.

To overcome this error at synchronous speed, a "running average" of the speed measurements is obtained from one half of the wheel and another running average from the other half. The difference between these two running averages is equivalent to twice the error caused by tooth-pitch errors.

One half the difference between the two running average is added to the last speed reading which is consistantly low or is substracted from the last speed reading which is consistantly high. The resultant speed output and, consequently, the acceleration output, is independent of tooth-pitch errors. The same two teeth (one in each half-revolution) are used to derive the indication of speed so long as synchronous speed is maintained.

At speeds other than synchronous, e.g. during start-up of the turbine, the acceleration is not a relevant factor and the correction for tooth-pitch errors (which is complicated by the use of a smaller number of displacement pulses for determining the measuring period) need not be used.

In other applications where longer sampling periods may be allowable, it may be possible to use a complete rotation of the shaft to set the duration of the measuring period. In that instance, the toothpitch error would not affect the speed signal.

The speed signal is accurate to a least 0.01% (i.e. 1 in 10000 time pulses) and this accuracy is substantially maintained over the whole range of operation since the tooth count control unit 34 attempts to mantain the time pulse count within the limits 213 and 214. Clearly, even at 1 in 213 (213 = 8192) a very high accuracy is maintained. That accuracy obtained at 3600 RPM would be equivalent to + 0.432 RPM at 3600 RPM as compared to the + 1 RPM at 3600 RPM quoted in UK patent specification No.

1385452.

The lower and upper limits for the time pulses can be chosen to give the desired accuracy of measurement, or the accuracy can be varied by varying the time pulse frequency; a frequency of 1.2 MHz would provide 0.01% accuracy at 3600 RPM would be equivalent to +0.36 RPM.

When the relatively moving parts are relatively moving at a substantially constant speed and the number of displacement pulses determining successive periods in constant, for example when the turbine is operating at about synchronous speed (i.e.

3000 RPM), the accuracy of the speed signal can be improved, if desired. In that instance, the processor stores successive time pulse counts (i.e. the outputs from the shift register 42) for a number of periods, say five, after which a speed signal is derived from the sum of the time pulse counts in the five periods.

The accuracy of the speed signal is about 0.002% (i.e.

1 in 50000 time pulses) for a five-period time pulse total. Clearly, the accuracy around synchronous speed can be varied considerably depending on the number of periods for which the time pulse counts are stored. However, the up-date frequency also varies depending on the number of periods selected.

If necessary, the processor could operate to provide a speed signal derived from the time pulse count of each period and to provide, less frequently, a speed signal of greater accuracy derived from a number of time pulse counts.

The overspeed downcounter48, in receiving time pulses divided by eight, is not as accurate as the counter 40. However, if greater accuracy is required on the overspeed facility, it is necssary to use a counter 48 having a larger capacity and to feed it directly, via the AND gate, by the system clock 36.

The above description has referred to the measurement of both speed and acceleration but it will be apparent that either speed or acceleration alone could be determined.

In the particular system described, the lower limit of the range is 48 RPM since at that speed the displacement of the shaft during the 0.01 second measuring period is equal to the displacement required for the next displacement pulse to be generated. However, it will be apparent that by the selection of a different maximum time for the measuring period, speeds lower than 48 RPM could be measured.

When the speed of the turbine shaft falls to below 48 RPM,the processor receives an OVERRANGE signal and gives a speed signal of zero. It is a requirement, however, that the turbine is kept turning at a minimum speed typically 10 RPM which is known as the "barring" speed. To ensure it is known that the turbine is still turning the detection circuit 24 feeds its output pulses to the latch 52 to reset the latch 52. Consequently, if the pulsefrequency falls too low the latch is not reset and the relay 50 is tripped. That also happens if the displacement pulses fail.

While the method and apparatus have been described with reference to a rotating shaft, it will be appreciated that they are equally applicable to determining linear speed provided the application permits displacement pulses to be generated. For example, it may be desirable to monitor the speed of a conveyor system in which instance the displacement pulses may be generated by a photocell detecting markings on the conveyor.

In a modification (not shown) only one pulse may be generated per tooth if preferred instead of two.

Claims (8)

1. A method of obtaining a repeatedly updated signal indicative of speed of displacement comprising generating displacement pulses and counting a number of the same equal to a registered number of pulses to determine a period during which high frequency time pulses are counted and deriving a speed signal from the count of time pulses taken with said registered number of pulses, determining whether the count of time pulses is within, below or above a range set by lower and upper limits, and respectively maintaining, increasing or decreasing said registered number to determine the next successive period.

2. A method according to claim 1, in which said registered number is increased or decreased by doubling or halving the registered number previously used.

3. A method according to claim 1 or claim 2, in which a count of time pulses below said lower limit determined when said registered number is a max- imum is used to initiate generation of an overspeed signal.

4. A method according to claim 1,2 or 3, in which when the speed and the number of displacement pulses determining successive periods are constant, the total of time pulses counted in each of at least two successive periods is stored and summed and said signal is derived from said summed totals.

5. Apparatus for obtaining a repeatedly updated signal inidicative of speed displacement comprising a generator for generating displacement pulses, a programmable register, a generator for generating high frequency time pulses, a counter for counting said time pulses for a period the duration of which is determined by the time taken to count a number of said displacement pulses equal to a number registered in the register and means for processing the outputs from the register and the counter to produce a signal indicative of said speed of displacement and control means operable according to whether the count of time pulses is within, below or above a range set by lower and upper limits to programme said register to respectively maintain, increase or decrease said number registered in the register.

6. Apparatus according to claim 5, in which a second, programmable, counter for counting time pulses in each period is programmable with a predetermined number less than said lower limit and, if the number of time pulses counted in a period is less than said predetermined number, has an output usable to initiate generation of an overspeed - - signal if said registered number is a maximum.

7. A method according to claim 1, substantially as hereinbefore described with reference to the accompanying drawings.

8. Apparatus according to claim 5, substantially as hereinbefore described with reference to the accompanying drawings.