IE45251B1

IE45251B1 - Electrical energy meters

Info

Publication number: IE45251B1
Application number: IE140577A
Authority: IE
Original assignee: Heliowatt Werke
Priority date: 1976-07-07
Filing date: 1977-07-06
Publication date: 1982-07-14
Also published as: CH623416A5; IT1080625B; FR2357907B1; BE855872A; DE2630959A1; AT368815B; IE45251L; NL7704500A; DE2630959C2; LU77696A1; GB1575148A; FR2357907A1; ATA291277A; DK205777A

Abstract

This meter exhibits a measuring mechanism which contains a multiplexer (M), an analog/digital converter (AD) and a computer (MC). The multiplexer samples current and voltage variables at time intervals and its inputs are connected to zero potential and reference potential for correcting calibration errors and for reducing the number of clock cycles of a sampling cycle. The values supplied to the computer (MC) in digital form from the multiplexer (M) and the analog/digital converter (AD) are combined with stored correction values before further processing and stored for calculating the instantaneous power since the associated input variables are sampled successively, and the sampling error is compensated for with the aid of linear interpolation or extrapolation. The instantaneous values obtained in this manner are combined with a reference frequency and supplied to memories with indicating devices.

Description

This invention relates to electrical energy meters (usually kilowatt-hour meters)having a static cr solid, state measuring unit, and is particularly cut not exclusively concerned with a meter for three-phase supply systems comprising Voltage dividers and current transformers for level matching, a current supply, and an indicating device.

Known kilowatt-hour meters having static (solid state) measuring units as at present employed operate with analog signal processing methods, preferably based on the time-division method with subsequent current-frequency conversion. Such methods are unsuitable for an economical three-phase current meter, because with analog signal processing, accurate and hence costly component elements must be employed, which can additionally falsify measurement results by reason of their inconsistency in respect of tims.

Energy supply organisations expect that a kilowatt-hour meter can be left in the supply system for many years without being checked and that such an appliance should remain operable for twenty to thirty years.

The invention is concerned with the provision of a generally improved meter.

According to the present invention, there is provided an electrical energy meter for metering an a.c. supply having one or more phases, the meter having a solid-state measuring unit which comprises: a multiplexer having inputs for receiving current- and - 3 - *82£ίϊ voltage-proportional signals for the or each phase, a zero potential signal, and a reference potential signal, and arranged to output said signals successively; an analog-digital converter connected to receive the multi5 plsxer output and arranged to output the same in digital form? and a digital computer arranged to control operation of the multiplexer, to receive the output cf the converter, and to perform tha f: ϊowing operational calculation and storage of a zero-point error value from the zero-potential signal when received by the computer; calculation cad storage cf a e·*·libration error value from the reference potential signal when received by the computer; compensation of ea<'h received said current- and voltage15 proportional signal for said error values; fcr the or each phase, storage of a succession of values of at least one cf said current- and voltage-proportional signals for that phase, e.··? accessively sampled by the multiplexer, and interpolation :/1 thin or extrapolation from said succession of values to compensate for relative time displacement between sampling instants of the current- and voltage-proportional signals; fcr the or each phase, calculating from the error value compensated currant and voltage-proportional signals the instantaneous power for that phase; and providing an output signal representative of energy consumption from said supply.

The computer say typically be a minicomputer.

For a better understanding of the invention and to show how the same may be carried into effect,reference will now be made, by way of example, to the accompanying diagrammatic drawings, 48351 - 4 in which: Figure 1 illustrates a meter employing a two-channel multiplex system; Figure 2 illustrates a mater employing a single-channel multiplex system; Figure 3 illustrates a meter employing a single-channel multiplex system in whieh all the input quantities are present as current signals; Figure 4 are waveform diagrams illustrating signal sampling possibilities for calculating instantaneous power; and Figure 5 illustrates a sampling sequence of a voltage and of a current for the formation of an instantaneous power.

Since a digital computer can only process digital values, the three voltage-current pairs of the three phases of a threephase supply system must be converted into corresponding digital values. It must here be. borne in mind that, for determining the active Dower, the time relationship of the voltage ’ . and of the z current of a phase must be maintained. For conforming to preset error limits, at most one error of a few minutes of angle must occur, this corresponds to a time error of a few microseconds at 50 Hz.

If two analog-ditital converters and one two-channel multiplexer are employed, a time error can be avoided at additional cost (Figure 1).

Advantageously, however, there are employed an analogdigital converter and a single-channel multiplexer initially having six inputs. In this case, for example, the six quantities to be measured are applied to the analog-digital converter in the following sequence by the multiplexer and thereafter converted into digital values and applied to a digital computer for S 2 5 i further processings VWW» Between the sampling instant, for example of the voltage ϋ, and the sampling instant of the associated current I , there R * R occurs owing to the multiplexing a time offset whose magnitude depends upon the multiplexing frequency. isiliCS tscxCil £jticL ntity to be measured is sampled by the multiplexing., its value is determined substantially from the harmonics which have still to be monitored. Assuming that harmonics higher than the tenth play no part, a sampling frequency of 1 kHz would be sufficient. Taking 2kHz, at least 16 kHz=60 us are required, including zeroising and calibration cycles. In this case, a time shift of 60 us lies between the sampling of a voltage anti that of the associated current? in this way, the measurement of the active power is already inadmissible falsified, If the multiplexer, the analog-digital converter and the computer are so that the multiplexing frequency can be raised to such extant that the residual time error is negligible, no further measures need be adopted. For the computer, the multiplication of two numbers in a few microseconds occctitutes a problem, and this requirement can be met substantially only by parallel multiplication.

This means that a fast computer must be employed, which does not involve any increase in cost of the equipment. In the following, therefore, there will be described a number of methods by which the time error occurring with slow sampling can be substantially compensated for, so that it is possible to employ for the processing of the measured value a computer having a substantially lower output capacity. 453S1 The first method involves the use of analog delay elements which are looped into the supply conductors leading to the multiplexer at the three msasurands subsequently sampled in each instance. Starting from a numerical value of 60 (is as assumed in the preceding example, such a delay is difficult to establish and to maintain constant over long periods of time.

At higher multiplexing frequencies, when the time error is in any case already small, however, this method appears to be entirely applicable for substantially compensating for this residual error.

It is known from German Patent Application P 2537549,9 that the time error which unavoidably occurs in single-channel multiplexing can be profitably used to compensate for signal delays or angle errors v/hich have arisen elsewhere in the electronic meter, by ensuring that the sign and the magnitude of the tv/o delays cancel one another out. It is true, however, that the magnitude fixes the multiplexing frequency which might in some circumstances be inconsistent with other requirements.

Advantageously, herefore, the computer can be arranged always to first store a number of sampled values before evaluating them. These stored values can be used in multifarious ways for interpolation and extrapolation calculations. in the simplest case, it is possible to provide a possibility of accurate interpolation by a variation of the sampling sequences In this case, the multiplexer requires no more inputs, i.e. eight inputs including the zero-point and calibration correction, but the number of timing pulses par multiplexer cycle is increased. The computer can then calculate by linear 4SSSI interpolation from the two sampled voltage values on either side of the associated current value in respect of time the missing voltage value and can then multiply the latter by the respective current value for ascertaining the instantaneous powar.

A further sampling sequence is: & ·: us“as“Is”UwVIT from which the computer calculates by extrapolation the voltage value assaeiatted in point of tims with the current value.

In order to avoid the increase in the number of timing pulses per multiplexer cycle, it is alternatively possible to effect an interpolation which is more extensive in respect of time. In this case, like multiplexer cycles no longer succeed one another, but the three voltages: Wt are first alternately sampled, and the three currents: T _T _T “R S T are sampled in next passage through the multiplexer. The computer then interpolates the voltage value associated with the current value in respect of time from the two voltage values which are derived from the preceding and succeeding multiplexer cycles. With a sufficiently high multiplexing frequency, the interpolation error which occurs remains sufficiently small.

In principle, it is immaterial whether the voltage values or the current values are ascertained by interpolation. In the method comprising double sampling cf the voltages or of the currents, however, it is desirable to sample the currents more frequently, because the supply system voltage is generally of very good sinusoidal form, while the current may be greatly distorted, for example by thyristor-controlled loads, so that it is more accurately determined by the more frequent sampling.

It is also possible in the interpolation to effect the mean-value formation on the plane of the instantaneous power, that is to say, for example, with the sampling sequence U—I—U the first voltage value may also first be multiplied by the current value, and then the second, and the averaging can be effected by way of the two products.

It is shown in Figure 5 how - without regard to the multiplexing of the three phases - a voltage u(t) and an associated current i(t) are sampled. The nuiriber of sampling operations will be denoted by n. The individual sampled values are multiplied by one another as follows: pui(fc) ’ ·*%-!) - V By introducing the functional values, there is generally 15 obtainad: N2-1 pui(t,n = 1 n [u sin(fflQ v T )] [ f sin(ffior T . f . 27Γ )] n y=0 [ό 8ίη(ωθΚ TM)] [fsin(£Doy· T^ . f - 27Γ )]J Χ=·1 The result for the relative measurement error of the power is: 2° frel=/cos W17 The error is dependent neither upon the drive nor upon the phase displacement between voltage and current, but only upon the nuiriber of sampled values per period. The period and the nuiriber of sampled values per period, however, are constants; 45351 therefore, the measurement error can be included in the calibration once and for all and is thus eliminated. For a measurement error of 1%, at least n=44 samples per period (20 ms) are obtained? this leads to a sampling frequency of 1.1 kHz. It is immaterial whether the sampled values of voltage and current can be acquired in one and the same multiplexer cycle or in alternate multiplexer cycles; the only determining factor for the measurement error is the time spacing.

With alt.-.-nxtc cycles, the number of timing pulses in a sampling cyc Is· is reduced to four, -che sequence having, for example, the following appearance: ϋ -U -IL-zero — I_-I -I -reference voltage — ϋ -ϋ -U^-zero — fs M Ji i\ J. K, X IR~1S' et0‘ Theoretically, it is possible to store intermediately in analog form at the instant of the sampling of one quantity,for example of a voltage, the associated value of the other quantity, for example of a current, so that a pair associated in respect of time is available for the formation of the product. In practice, this ·.: difficult because, in kilowatt-hour meters, tha measuring error remains constant in relation to the momentary measured value ever the whole measuring range, so that even very small values must be accurately determined. Even the intrusion cf switching pulses would falsify considerably the Stored values if an analog store were employed.

Since the problems arising with single-channel multiplexing can thus be solved, tha principal remaining source of error is ths analog-digital convertor, because the subsequent digital further processing by ths computer may be regarded as free from error. in each multiplexer cycle — or, for example, in each second multiplexer cycle in the case of alternate multiplexer cycles - there is contained a zeroising phase in which the input of the analog-digital converter is connectad to zero potential by way of the multiplexer. V7hen the analog-digital converter applies to the computer a digital value other than zero, the computer detects the deviation and stores the value, in order thereafter to correct therewith each further digital value supplied by the analog-digital converter, before the evaluation. Owing to the rapid succession of zeroising phases, even relatively rapidaaso-point migrations of the analogdigital converter are eliminated. Consequently, there is no zero-point error at any time in the whole lifetime of the equipment.

In each multiplexer cycle “ or in each second one in the case of alternating cycles — the input of the analog-tligital converter is applied by way of the multiplexer to a reference voltage whose desired value is contained in the store of the computer. When the analog-digital converter supplies a digital value differing from this desired value, the computer establishes the difference, stores the correcting factor and thereafter corrects each supplied value after having taken into account the zero-point error. In this way, the calibration ot pitch error is eliminated throughout the useful life of the appliance.

As residual error, there remains the linearity deviation of the analog-digital converter conversion curve, but this can always be kept sufficiently small because it depends substantially only upon the precision of the resistance divider network.

Owing to the requirement for constant relative measuring error throughout the measuring range, it is undesirable to operate with a linear analog-digital converter conversion curve which requires a more costly analog-digital converter, the accuracy of which would not be fully utilised in the case Of large measured values. Therefore, a non-linearity gradated characteristic curve ia preferred, which is cf such kind that the relative measurement error is equal at each point of the curve. Tha logarithmic gradation, which is suitable per se for this purpose, is advantageously replaced by binary gradation, which is easier to apply. Of course, any desired analog-digital cotiverter conversion curve can be taken into account by a suitable program of the computer. All the circuitry parts of the electronic energy meter which have hitherto been referred to tiie multiplexer &, the microcomputar «4C and the analog-digital converter ?.D — can be monolithically integrated, including the resistors cf the analog-digital converter.

In the case of a three-phase current meter, there are employed, for example, voltage dividers and current transformers, for example of tr.e type having a ferrite shell-type core, with or without electronic error compensation. The multiplexer and the analog-digital converter may selectively process all input signals in the form of voltages or of currents, it being more readily possible fcr currents to be passed through the multiplexer; the comp? rt· tor of the analog-digital convertor also preferably compares currents instead of voltages.

In any case, the six input quantities must be sealed down in order that they may be processed by the electronic system. In the manufacture of such a meter, therefore, the voltage divider ratios and the load resistors at the current transformers, for example, must be adjusted to their desired values. Such adjusting operations require costly adjustable component parts and are particularly labour-intensive.

This dis?. ':vnntsge can be obviated by virtue of the fact that, in principle, all adjusting operations are eliminated, being replaced by corresponding programming in a store, For this purpose, tha actual values of the voltage divider ratios, the current transformation ratios (including the load resistances) and the reference voltage are measured in a preferably fully automatic testing operation in each instrument, and corresponding correction values are written, into a store accessible to the computer. There is desirably employed for this purpose a programmable read-only store (PROM), which is contained in the computer switching circuit and is automatically programmed by the testing device. If a repeatedly programmable read-only store (RePROM) or electrically alterable read-only store (EAROM) is employed, subsequent recalibrations of the meter can again be effected with tha same testing apparatus by programming-in fresh values.

Electronically error-compensated current transformers are known and permit very accurate current transformations with small ferrite shell-type cores. However, owing to the necessary auxiliary amplifier, they generally have at the output an offset unidirectional voltage. In the calculation of the instantaneous pswer, no error is set up in this way provided that such a unidirectional voltage component is present only at one input of the multiplier; this is a pre-requisite if the meter is fed by voltage and current transformers, as has hitherto usually been the case. In the case of employment with direct connection, however, it is quite possible in the case of relatively long supply lines and thyristor-controlled loads for a undirectional component of the voltage to be set up which, together with the unidirectional offset of an electronically errorcompensated current transformer, would give a measurement error. It is therefore desirable to omit the electronic error compensation. Current transformers having ferrite cores exhibit relatively large errors. When the method described here is applied, the automatic testing device is capable of measuring 452oi the two error curves of each individual one of the three current transformers in a meter over the whole control range and to store them in the programmable read-only store (PROM). The computer will then weight each individual measured current value supplied by tne analog-digital converter with the associated error curves. A pre-requisite for the practical application of this method is that the error curves of such a current transformer should remain within given tolerances over a relatively Jong time and even after any possible overloading.

In a further development of the invention, temperaturedependent measurement errors are substantially compensated for in like manner by virtue of the fact that the known and generally only slightly specimen-dependent temperature coefficients of the component elements which ars aot affected by the automatic zero15 point and calibration correction are also stored in the programmable read-only store (PROM), while there is also provided a suitable temperature sensor whose output signal is applied to the computer at sufficient intervals of time by way of the multiplexer and the ma log-digital converter. When particular requirements he.'·.; - ™ be met in regard to measuring accuracy, it is of course also possible to measure automatically in the case of each individual specimen the temperature curve of the reference voltage source which determines the calibration, and to store it in the programmable read-only store (PROM).

When the meter is in service for a long period of time, the ageing-induced drifts of the component elements not affected by the zero-point and calibration correction are of importance. There are principally concerned here the input voltage dividers, the current transformers and the reference voltage source. In well-controlled manufacture, the ageing rates are known and could also be stored in the non-alterable read-only store (ROM) or programmable read-only store (FROM) of the computer. Since energy maters operate in most cases with a mains supply for the electronic system, there can readily be derived from the mains frequency a time base which is Set to zero in the aforesaid fully automatic testing operation. In the calculation of the energy, the computer takes account of the momentary state cf the time base and the ageing rates for substantially compensating for influences due to ageing, and, upon re-checking of the meter, new correction values or curves arising from the ageing which has meanxvhile occurred in the component elements are written into the respective store.

Xn the case of an energy meter—more especially a threephase current meter—fche analog-digital converter must process a very large dynamic range. Since its input is continuously changed over to other input quantities by the multiplexer, its input signal may jump from one sampling to the next, for example from the maximum negative value to the maximum positive value. These large signal jumps impose high requirements on the comparator of the analog-digital converter. The comparator requires a particular recovery time after each overshoot. The exact monitoring of a signal value suddenly changing from a maximum value to one close to zero is critical. A known monitoring analog-digital converter avoids this problem in that the comparator input signal corresponds in each instance only to the difference between two samples.

Xn a further development of the invention, the principle of the monitoring analog-digital converter is also applicable.

At e&ch sampling of an input quantity, the sampled value is any case stored, as described, in the computer for the purpose of interpolation. After each sampling, the analog-digital converter converts the analog value into a digital value which is taken up by the computer. Thereafter, the analog-digital 45201 converter is again free and can be preset by the computer, before the coifimsneement of the next sampling, to the last sampled value of the input quantity which is now to be sampled. The comparator again receives only the difference between two successive sampled values. The presetting of the analog-digital converter and the application of the input quantity to be measured must be so adapted to one another in respect of time that a comparator overshoot is kept to a minimum.

As alrec.dy mentioned, it is advantageous to supply the input quantities as currents and tc provide a current comparator.

As compared with a conventional Ferraris meter, the measuring method described here is susceptible to disturbing pulses owing to its rapidity. In order to make the electronic energy computer behave similarly to the Ferraris meter, the computer is advantageously sc programmed that the jumps between two samplings are limited in their extent and relatively large jumps are evaluated as spurious' pulses. In this case, either such sampled values recognised as spurious must be entirely suppressed, or tha part which exceeds the programmed maximum value is subtracted. In addition, a sliding mean-value formation is possible. In this case, the averaging takes place before the calculation of the instantaneous power, or even after it.

By appropriate programming of the computer, it can be ensured that the often undesirable accuracy of the measurement is reduced in respect of the determination of harmonics of the mains frequency and made similar to that of a Ferraris meter.

For this purpose, Fourier transformation (FFT) is carried out, for example by the computer advantageously to take into account only a fundamental wave. After the calculation of the instantane· ous power of each phase by multiplication, a digital word is available as result, the values of the three phases also having to be added in the case of a three-phase current meter. For determining the energy, an integration over the time is also necessary. — The energy consumption is detected in an accumulator register and displayed on an indicating device. For this purpose, the known counting and storing cyclometer-typa registers having stepby-step motor drive may be employed. Alternatively, storing electronic indicating devices, for example eleotrochromic or electrolytic displays, may ba used. Normal non-storing electronic indication is also possible if a non-volatile store, for example a MNOS store, is available, whieh store is refreshed sufficiently frequently to ensure that data is maintained. Owing to the limited number of write cycles which such a store can perform before becoming unserviceable, the numerical value which must be saved must be written therein in the event of a detected imminent mains failure.

With each kind of digital indication, a jump in the lowest digital position corresponds to a particular amount of consumed energy. These counting pulses must be obtained, for example, by the computer counting off, after determination of the whole instantaneous power, a number of periods of a reference frequency corresponding to this instantaneous power; those counted-off pulses must then be applied to the display unit either directly or after corresponding frequency division. Alternatively, it is possible to change over in each instance, with that value of the instantaneous power which is present as the digital word, a programmable frequency divider which continuously divides the pulses of the reference frequency source by the programmed factor. Preferably, the mains frequency is used as the reference frequency; it is alternatively possible to derive the timing frequency of the computer from the supply frequency by way of a phase control loop (PAL) or to use a quarts element. The computer is caused by a corresponding programme to count off particular times; the programmable timer contained in many 45351 - 17 computers may also be used with advantage.

For testing purposes, it is advantageous to employ a particular testing output of the computer, at which the energy consumption is available with a shorter measuring time than at the output intended for the indicating device.

Economical computers operate with relatively low timing frequencies and therefore require long times, fcr example, for multiplications. order to save time in the multiplication, use must again X made of the principle of the monitoring analogdigital converter, wherein the full values of voltage and current are not multiplied by one another each time, but only the difference values in relation tc the preceding sampled values, in accordance with the schemes If the multiplication time is too long for such an application in the case of serial processing, a separate parallel multiplication unit can always be employed without any change in the described measuring principle.

A parfeicuirr problem with electronic energy meters is the safety against disturbance. Conventional Ferraris meters withstand exceptionally high peak voltages at the inputs, and they often withstand peaks of more than lo kV. The current input proposed as advantageous - even on the voltage side - is also found to be more favourable in this respect, because practically no voltages arise on the equipment side, so that very effective protective measures can be applied.

In an energy meter comprising a computer, special precautions must be taken against penetration of disturbances into the electronic system, because a disturbing pulse in a computer can lead not only to incorrect data, but also to an incorrect instruction. In this respect also, we have found that the single-chip microcomputer proposed in German Patent Application P 253754S.9 has proved superior, because the effects of disturbances are minimised if the whole computer is integrated on a single semiconductor crystal. Disturbances then only affect the inputs and outputs of the computer, disturbances at the inputs merely producing periodically incorrect input data, but not causing any incorrect performance of the programme.

There is illustrated in Figure 1 the less advantageous, because more costly, solution comprising a two-channel multiplexer M and two analog-digital converters AD^, AD^» Each multiplexer M has five inputs, since each of fhe two multiplexer channels and each of the two analog-digital converters AD,, AD^ must be corrected in regard to their zero-point and calibration errors. A microcomputer MC controls both the multiplexer M and the analog-digital converters AD^, AD^· For the purpose of energy calculation, it also receives the reference frequency f^^ The cumulative energy consumption is finally displayed on an indicating device. In this figure, as in the succeeding figures, it is assumed that the microcomputer Iffi also comprises the necessary stores. This applies also to the frequently mentioned electrically programmable store for receiving the correction curves or values. For example, there may be employed a metalnitride-oxide semiconductor store (MNOS). Of course, it is also possible to assemble the microcomputer I4E from a number of integrated switching circuits; -this does not in any way change the idea of the invention.

Also, it is possible in accordance with the prior art to accommodate ths analog-digital converter or converters AD^, AD^ together with the associated resistors, as well as the multiplexer or multiplexers M, together with the computer on a single switching circuit. The whole electronic three-phase current meter then consists substantially of three current transformers, one resistance arrangement (for example thin-film or thick-film - 19 48331 technique), the major switching circuit, the display means and the current supply system. As described, it does not comprise any component elements that can be balanced or adjusted.

There is illustrated in Figure 2 the more advantageous solution comprising a signal-channel multiplexer M and only one analog-digital converter AD, wherein all the input signals of the multiplexer ti, are present in the form of voltages. The illustrated of the connections is not essentially identical with ihe sequence in which the input, signals from the multiplexer M are sampled. The computer passes the address of the desired input in each instance through a control line to the multiplexer M? the sequence is dependent upon the programme chosen.

There is illustrated in Figure 3 the circuit arrangement in the case where all ohe input signals of the multiplexer M are present in the form of current signals.

There are shown in Figure 4 different possible ways of sampling the signals and calculating the instantaneous power in each phase. Figures 4a, 4b and 4c here relate to the measurement of only one phase, while Figures 4d, 4e and 4£ apply to the measurement in the case of three phases.

The invention is also applicable with any desired number of phases, for example even· co a so-called three-conductor meter, in which only two measuring channels are required.

Figure 4a illustrates the sampling and the interpolation in the case cf the voltage and current measurement in a singlephase system: the indices denote the number of the respective sampling; the voltage and current inputs are alternately sampled. A first sub-product (instantaneous power) p^ is obtained, for example, by interpolating from the two voltage samples uq and u2 - 20 a voltage value which is thereafter multiplied by the sampled current value i,. A further sub-product p_ is obtained, for l 3 example, by interpolating from the two voltage values u2 and u^ a value which is thereafter multiplied by the current value i^The total power P is obtained for a particular number of samples n as: v=—(Pxtp +...), n in which n multiplications are performed.

Voltages and currents are completely equivalent and can thus be interchanged.

In Figure 4b, there is illustrated another product formation for the same case of alternate sampling of input voltage and input current; sub-products p^ are here formed from the sampled values present, the voltage and the current . being alternately interpolated. The power P is here calculated as : P-—{ρ^ρ^ρ^.. -) n There is illustrated in Figure an example of the case where the averaging takes place only after fche formation of the sub-products pn· The total power P is obtained as in case 4b, but only one store is here required for any one sampled value, because no interpolation takes place, or no mean value is formed before the multiplication.

The following applies to all three methods: P^_(Uoil+ilU2+U2i3+i3V· · *} n 45231 S xn Figure 4e. and the sampling of the The following sampling sequence has -I -U -1 -U -I , For that part of the measurement which is of greatest practical importance iii a three-phase current supply system, the sampling of the phase R is first shown in Figure 4d, the sampling of the phase phase T in Figure 4f. here been assumed: UJ The arrows shown with the sub-products p indicated how n these sub-products have arisen.

A large number of sampling frequencies is possible; some typical ones for the method of the invention are, for example: υ„-Ι -υ ~ϋ -1 -ϋ,,-ϋ -I ~U RRRSSSTTT ι-υ-υ-ι -ϊ, -i R S T R S T J -I„-U -I -U-Im R S T R S T As mentioned, the computer determines which input signal the multiplexer M is to pass in each instance,so that, depending upon the programme, any desired sampling sequences and any kind of evaluation of the samples are possible in the same circuit arrangement.

In the case of the energy meter comprising a computer as described here, the sampling according to Figure 4a is particularly desirable, because a computer can relatively rapidly add, but takes a comparatively very long time for a multiplication; according to Figure 4a, however, only n/2 multiplications are carried out.

In Figures 4a to 4f, the sampling of the zero-point or reference potential has been omitted for the sake of clarity. Depending upon the drift to be expected in the zero point or the calibration, it is merely necessary that the zero-point or - 22 calibration potential should be sampled at sufficient intervals of time and that corresponding correction values should be formed and stored. It is therefore by no means necessary for the aforesaid other two quantities to be simultaneously sensed in each cycle for the sampling of the input quantities.

In the above described meters, in the determination of the deviations of those component elements which enter the measurement results from desired values, the multiplexer and the analog-digital converter may be employed simultaneously.

Then, suitable input signals are applied to the meter and actual values are taken from the output of the analog digital converter or from a suitable output of the computer in digital form, so that all the curves or correction values thus measured contain the individual deviations both of the multiplexer and of the analog-digital converter from the desired conversion characteristic curve between 0% and 100% of the measurement range, whereby, with an appropriate number cf measuring points per correction curve, each deviation of the conversion characteristic curve from its desired form (linear, logarithmic, binary, etc) is compensated to within any desired small residual error.

The multiplexing frequency of the above described meters may be synchronised with mains frequency, or it may free wheel in relation to mains frequency or be additionally frequency modulated. Alternatively, the multiplexing frequency may be derived from a timing frequency of the computer by way of a program thereof. A time setting required for carrying out integration operations, may be produced in a manner known per se by a program loop or a timer included in the computer, where a timing frequency of the computer is used as a reference frequency. & Κ 2 5 - 23 The above described maters may be arranged to operate as maximum demand meters, or as excess demand meters. Their output signals may represent, in addition to energy consumption, a reference frequency, a meter test, or a meter identification code. An electrically programmable store of the meter may contain a meter identification which consists of a serial number and optionally a data code and a test code. The meters may be provided with a protected connection for a testing instrument, with the aid of wnich the meter can be tested. The meters may be arranged to perform metering operations other than simple energy computation. For example, in a meter, the computer may be programmed to receive instructions from a centralised control, and be so programmed that it effects by digital filtering methods, not only decoding of the centralised control instructions, but also the filtering out of centralised control pulses after appropriate level matching and signal processing. Then, programming of the computer as a centralised control receiver may be contained in a non-volatile store. In the performance of tasks of a centraiised-control receiver, the computer may output control signals particular to a respective energy consumer.

Claims

1. CIAIMSs1. An electrical energy meter for metering an a.c. supply having one or more phases, the meter having a solidstate measuring unit which comprises t a multiplexer having inputs for receiving current- and voltage-proportional signals for the or each phase, a zero potential signal, and a reference potential signal, and arranged to output said signals successively; an analog-digital converter connected to receive the multiplexer output and arranged to output the same in digital form; and a digital computer arranged to control operation of the multiplexer, to receive the output of the converter, and to perform the following operations: calculation and storage of a zero-point error value from the zero-potential signal when received by the computer; calculation and storage of a calibration error value from the reference potential signal when received by the computer; compensation of each received said current- and voltageproportional signal for said error values; for the or each phase, storage of a succession of values of at least one of said current- and voltage-proportional signals for that phase, as successively sampled by the multiplexer, and interpolation within or extrapolation from said succession of values to compensate for relative time displacement between sampling instants of the current- and voltageproportional signals; for the or each phase, calculating from the error value compensated current- and voltage-proportional signals the 452Si - 25 instantaneous power for that phase? and providing an output signal representative of energy consumption from said supply.

2. A mater according to claim 1, for metering a singlephase supply.

3. A meter .recording to claim 1, for metering a polyphase supply, wherein the computer is arranged to sum the instantaneous powers of the ' ~snsctive phases.

4. A metsr according to claim 3, for use in metering a three-phase supply.

5. A metar according to claim 4, wherein said multiplexer is a single-channel multiplexer having eight inputs, three of which are for receiving the phase current- proportional signals, another three of which are for receiving the phase voltageproportional signals, and the remaining two of which are for receiving said zero potential and reference potential signals.

6. A meter according to any one of the preceding claims, including indicator means for indicating said energy consumption.

7. A meter according to claim 6, wherein said indicator means has a non-volatile store associated therewith.

8. « metar according to any one of the preceding claims, wherein said interpolation or extrapolation is linear.

9. A meter according to any one of the preceding claims, wherein the computer is arranged to effect said time displacement compensation before the calculation of the phase instantaneous power (s5.

10. A meter according to any one of claims 1 to 8, wherein the computer is arranged to effect said time displacement compensation after the calculation of the phase instantaneous power(s). -Se

11. A meter according to any one of the preceding claims, wherein said output signal includes a reference frequency.

12. A meter according tosny one of the preceding claims, wherein the deviations of the actual values from the desired values of those parts of the meter which are not affected by compensation for said zero-point and calibration error values and of a source of said reference potential are determined and corresponding correction values are stored in an electrically programmable store and taken into account by the computer in the calculation of the energy consumption.

13. A meter according to any one of the preceding claims, wherein a timing frequency of the computer is derived from a quartz element or by way of a phase control loop from the mains frequency.

14. A meter according to claim 13, wherein the timing frequency of the computer is used as a reference frequency for integration operation, by the computer.

15. A meter according to any one of the preceding claims, wherein the multiplexing frequency is synchronised with the mains frequency. IS. A meter according to any one of claims 1 to 14, wherein the multiplexing frequency freewheels in relation to the mains frequency or is additionally frequency-modulated.

16. 17. A meter according to any one of claims 1 to 14, wherein the multiplexing frequency is derived from the timing frequency of the computer by way of a program thereof.

17. 18. A meter according to claim 4 or to any one of claims 5 to 17 as appendant thereto, wherein said multiplexer is a twochannel multiplexer each channel having five inputs, and is arranged to feed two said analog-digital converters, wherein one *82 SI - 27 of each set of five inputs is for receiving said aero potential signal, a second one of each set of five inputs is for receiving said reference potential signal, the phase voltage-proportional signals are applied, in use, to the remaining three inputs of one set, and the phase current-proportional signals are applied, in use, to the remaining three inputs of the other set.

18. 19. A meter according to any one of the preceding claims, wherein, for carrying out integration of instantaneous power values by proraam control, the computer is arranged to count off a number of periods of a reference frequency, corresponding to the respective instantaneous power value, and apply them to a storage device.

19. 20. A meter according to any one of the preceding claims, wherein, in use, instantaneous power values calculated hy the computer are used to set a programmable frequency divider or multiplexer, which divides or multiplies pulses of a reference frequency, and in that the counting pulses thus obtained are applied to a storage device.

20. 21. A meter according to any one of the preceding claims, wherein a timing frequency of the computer is used as a reference frequency, and in that a time setting required for integration operations is produced in a manner known per se by a program loop or a timer included in the computer.

21. 22. A meter according to any one of the preceding claims, including means for compensating for time delay in sampling the values of voltage- and cv,r rent-proportional signals of each phase, by means of an analog delay element.

22. 23. A meter according to claim 2, or to any one of claims 6 to 17 and 19 to 22, as appendant thereto, wherein the multiplexer is a single-channel multiplexer and the computer is arranged to receive sampled values of said current- and voltageproportional signals alternately and to always store the two 4 53 51 - 28 values last sampled, and to calculate, at each sampling, from the value just sampled and from the earlier sampled value of the same quantity which is available in the store, as interpolated value, and to multiply the latter value by the value of the 5 other quantity available in the store, thereby to determine the instantaneous power, whereby interpolation values for the voltage- and current-proportional signals are alternately calculated and multiplied by the last stored value of the other quantity.

23. 24. A meter according to claim 2, or to any one of claims 10 6 to 17 and 19 to 22, as appendant thereto, the multiplexer is a single-channel multiplexer and the computer is arranged to receive sampled values of said current- and voltage-proportional signals alternately and to calculate the mean value from two successive sampled values of one quantity in each instance by 15 interpolation and multiply this means value in each instance by that sampled value of the other quantity which exists in the time interval between them.

24. 25. A meter according to claim 2, or to any one of claims 6 to 17 and 19 to 22, as appendant thereto, wherein the multi20 plexer is a single-channel multiplexer and the computer is arranged to receive sampled values of said current- and voltage-proportional signals alternately and to continuously multiply by one another in each instance two sampled values of the two quantities which 25 succeed one another in tints, and sum the sub-products.

25. 26. A meter according to claim 4 or to any one of claims 5 to 17 and 19 to 22, as appendant thereto, wherein the multiplexer is a single-channel multiplexer which is arranged to sample the current-proportional signals (I , I , I ) and the R S T 30 voltage-proportional (U R , U g , U^) signals in the order: U -I -U -U -I„-U -U-Ϊ -U RRRSSSTTT 4BSS1

26. 27. A meter according to claim 4 or to any one of claims 5 to 17 and 19 to 22, as appendant thereto, wherein the multiplexer is a single-channel multiplexer which is arranged to sample the current-proportional signals (I R , Ιθ, 1^,) and voltage 5 proportional signals ill , U , IT ) in the order: R S T tt -τ -υ -I -U -I -V I -υ -I -U -X -U -I -ϋ -I -ϋ -I R R R S S S Τ Τ T R R R S S S T Τ T

27. 28. A meter according to claim 4 or to any one of claims 5 to 17 and 19 to 22, as appendant thereto, wherein the multiplexer is a sin “le-channal multiplexor which is arranged to 10 sample the current-proportional signals (I R , Ig, 1^,) and voltage proportional signals (U , U, U ) in the order: R ο T YVVWt

28. 29. A meter according to claim 4 or to any one of claims 5 to 17 and 19 to 22, as appendant thereto, wherein the multx15 plexer is a single-channel multiplexer which is arranged to sample the current-proportional signals (I R , Ig, 1^) and voltage proportional signals (ϋ ρ , TJ , U.J in the order: U -I„-U m -I -U -I R S T R S T

29. 30. A metc.i' according to any one of the preceding claims, 20 including means for supplying all input signals to the or each analog-digital converter in the form of voltages.

30. 31. A meter according to any one of claims 1 to 29, including means for supplying all in out signals to the or each analog-digital converter in the form of currents. 25

31. 32. A meter according to any one of the preceding claims, wherein the conversion characteristic curve of the or each analog-digital converter is so gradated that any measuring error in relation to any given measured value remains constant independently of the measured value, the computer taking into 30 account the conversion characteristic curve of the analogdigital converter in processing the values supplied by the - 30 latter.

32. 33. A meter according to any one of the preceding claims, wherein the or each analog-digital converter has a logarithmically gradated conversion characteristic curve the computer taking this curve into account in processing the values supplied by the analog-digital converter.

33. 34. A meter according to any one of claims 1 to 32,. wherein the or each analog-digital converter has a conversion characteristic curve gradated in binary form, the computer taking this characteristic curve into account in processing Of the values supplied by the analog-digital converter.

34. 35. A meter according to any one of the preceding claims, wherein the or each analog-digital converter is arranged to be controlled by the computer as a monitoring analog-digital conver ter in which, before sampling a particular quantity, the analogdigital converter is set by the computer to the previous sampled value of the same quantity, so that a comparator of the analogdigital converter only has to process the difference between the new and old sampled values.

35. 36. A meter according to claim 35, wherein, in the calculation of phase instantaneous power, the computer does not multiply the most recently sampled value but only the previous stored value of voltage and current by the current and voltage difference respectively.

36. 37. A meter according to any one of the preceding claims, wherein a parallel multiplication unit is employed for calculating phase instantaneous power.

37. 38. A meter according to any one of the preceding claims, Wherein the or each analog-digital converter, and the computer are integrated together on a single semiconductor chip. 4S2S1 - 31

38. 39. A meter according to claim 38, wherein the multiplexer is included on the chip.

39. 40. A meter according to any preceding claim, wherein electronically error-compensated current transformers are employed as input stages to the meter and the computer is so programmed that, in the calculation of instantaneous power, it eliminates from the values supplied by the analog-digital converter an erroneous offset unidirectional component which is contained unrein and which emanates from auxiliary amplifiers cf the current transformers.

40. 41. A meter according to any one of the preceding claims, wherein the computer is so programmed that, in the case of distorted input signals, it only takes into account the fundamental wave in each instance in the calculation of instantaneous power.

41. 42. A meter according to claim 41, wherein, for the purpose of computing the fundamental wave in the case of a distorted input signal, tne computer carries out a Fourier transformation.

42. 43. A mater according to any one of the preceding claims, wherein the computer is so programmed that it assesses as disturbances jumps in the sampled values which exceed a predetermined level, and that it either subtracts the excess part of or rejects such disturbed sampled values.

43. 44. A meter according to any one of the preceding claims, wherein the computer is so programmed that it effects a sliding mean-value operation in the calculation of instantaneous power or of the energy.

44. 45. ?. meter according to any one of the preceding claims, including an electrically programmable store the content of which can be maintained in the event of a power failure. fc* - 32

45. 46. A mater according to any one of the preceding claims, including an electrically programmable non-volatile store which is arranged to be refreshed at intervals sufficient to ensure that data stored therein is maintained.

46. 47. A meter according to claim 7 or to any dependent claim thereof, vzherein the existing energy consumption reading is arranged to be stored in the non-volatile store associated vzith said indicator means, in the event of a power failure.

47. 48. A meter according to claim 6 or to any dependent claim thereof, wherein said indicator means comprises an electronic indicating device.

48. 49. A meter according to any one of the preceding claims, provided vzith a protected connection for a tasting instrument, with the aid of vzhich the meter can be tested.

49. 50. A meter according to any one of the preceding claims, arranged to provide an output signal vzhich is representative of energy consumption and includes a meter identification code.

50. 51. A meter according to any one of the preceding claims, arranged to provide an output signal which is representative of energy consumption and includes a meter test.

51. 52. A meter according to any one of the preceding claims, arranged fc perform metering operations other than simple energy computation.

52. 53. A meter according to claim 52, arranged to operate as a maximum demand meter.

53. 54. A meter according to claim 52 or 53, arranged to operate as an excess demand meter.

54. 55. A meter according to claim 52, 53, or 54, wherein the computer is programmed to receive instructions from a centralised-control, and it is so programmed that it effects 33 45351 by digital filtering methods, not only decoding of the centralised-control instructions, but also the filtering-out of centralised-control pulses after appropriate level matching and signal processing.

55. 56. A meter according to claim 55, wherein the programming of the computer as a centralised-control receiver is contained in a non-volatile store,

56. 57. Α according to claim 55 or 56, wherein the computer, in vie performance of ths taste of a centralisedcontrol receiver, outputs control signals particular to a respective energy consumer.

57. 58. A meter according to any one of the preceding claims, comprising a store which is electrically programmable and reprogrammable as often as desired and which contains correction values or correction curves for all component elements of the meter which are not affected by the automatic zero-point and calibration correction, but of which the properties or values enter into the measurement, the computer being so programmed that it takes b··'..·.· account in the calculation of energy consumption the respective correction values or curves.

58. 59. A meter according to claim 58, wherein, in the determination of the deviations of those component elements which enter the measurement result from the desired values, the multiplexer and the analog-digital converter are simultaneously employed, suitable input signals being applied to the meter and the actual values being taken from the output of the analogdigital converter or from a suitable output of the computer in digital form, so that all the curves or correction values thus measured contain the individual deviations both of the multiplexer and of the analog-digital converter from the desired conversion characteristic curve between 0% and 100% of the measurement range, whereby, with an appropriate number of measuring points par correction curve, each deviation of the conversion characteristic curve from its desired form (linear, logarithmic, binary, etc) is compensated to witin any desired small residual error.

59. 60. A meter according to claim 58 or 59, comprising current transformers for each of which a correction curve for the magnitude and a correction curve for the error angle are ascertained, stored and taken into account in computation by the computer.

60. 61. A meter according to any one of the preceding claims wherein generic or individual temperature coefficients or temperature curves of component elements whieh enter the measurement result are stored in a read-only store (ROM) or in a programmable read-only store (PROM) or in an electrically alterable read-only store (EA.ROM) of the meter, the meter comprising a temperature sensor, the output signal cf which is applied to the computer, and the computer takes into account in the computation of energy consumption the measured operating temperature and the stored temperature dependencies.

61. 62. A meter according to any one of the preceding claims, wherein generic or individual ageing rates of component elements which enter the measurement result are stored in a read-only store (ROM) or in a programmable read-only store (PROM) or in an electrically alterable read-only store (EAROM) of the meter, the meter, has a freewheeling time base or a time base derived from a reference frequency, and the computer takes into account the computation of the energy consumption the ageing rates and the existing state of the time base.

62. 63. A meter according to claim 62, wherein the computer *82$ ι - 3S is so programmed that, on re-cheeking of the mater, new correction values or curves arising from the ageing which has meanwhila occurred in the component elements are written into a respective store. 5 64. A meter according to any one of the preceding claims, wherein an electrically programmable store of the meter contains a meter identification which consists of a serial number and optionally a data code and a test code. «35. ft r.-or according to any one of ths preceding claims, 10 being a kilowatt-hour meter.

63. 66. An electrical energy meter substantially as hereinbefore described with reference to Figure 2 or 3 of the accompanying drawings.