Estimation of power consumed on phase despite fraud

Abstract

A power estimation method implemented in a three-phase electricity meter (1), and the method includes the following steps: detecting fraud that tampers with the measurement of a first phase voltage (V1) present on a first phase (Ph1) without tampering with the measurement of a second phase voltage present on a second phase (Ph2); reconstructing first estimated voltage samples, images of the first phase voltage (V1) over time from second voltage samples, images of the second phase voltage (V2) over time; and estimating at least one first electric power consumed on the first phase by using the first estimated voltage samples.

Description

Technical Field

[0001] The present invention relates to the field of three-phase electricity meters. Background Art

[0002] Electricity meters used in industrial applications often measure power and energy distributed via very high current levels, often up to 2000 A. These meters are typically three-phase meters.

[0003] This instrument includes a current input terminal and a voltage input terminal.

[0004] Current sensors, such as Rogowski sensors or external current transformers, are usually installed on the phases. Each current sensor is connected to the current input via two wires. Thus, for each phase, the instrument acquires an image of the current circulating on said phase.

[0005] Each voltage input is typically connected to a phase by a wire. Thus, the meter has direct access to the voltage on each phase via the associated wire.

[0006] Known fraud involves cutting one or more of the wires connecting the phases to the voltage input terminals. The purpose of this type of fraud is to underestimate the total energy consumption measured by the meter in order to reduce the bill paid to the electrical energy distributor.

[0007] In the event of fraud, the prior art meter continues to make metering measurements on the three phases without taking into account the fraud and therefore actually underestimates the energy consumed. Even if fraud is detected, the distributor does not know the actual energy consumed by the fraudster's facility. However, it can prove useful to know the actual energy consumed, for example to recover payment for unbilled energy from the fraudster (in addition to possible penalties, intended to sanction the fraud attempt). Summary of the invention

[0008] The present invention aims to estimate the total energy consumed by a facility despite the presence of fraud including false voltage measurements on the phases.

[0009] In view of achieving this goal, a power estimation method is proposed, which is implemented in a three-phase electricity meter and includes the following steps:

[0010] detecting fraud that tampered with a measurement of a first phase voltage present on a first phase of a power distribution network without tampering with a measurement of a second phase voltage present on a second phase of the power distribution network;

[0011] reconstructing first estimated voltage samples, images of the first phase voltage over time from second voltage samples, images of the second phase voltage over time, the second voltage samples being offset over time by an expected theoretical phase shift between the first phase voltage and the second phase voltage in the absence of fraud;

[0012] At least one first electrical power consumed on a first phase is estimated by using the first estimated voltage samples.

[0013] When fraud is detected on the first phase, the measurement of the first phase voltage is falsified and therefore cannot be used.

[0014] Therefore, the power estimation method according to the present invention replaces the measurement of the first phase voltage with the first estimation sample, which is reconstructed from the measurement of the second phase voltage, while the measurement of the second phase voltage itself has not been tampered with by fraud.

[0015] Then, the method estimates a first electric power consumed on the first phase by using the first estimated voltage sample.

[0016] Thus, despite the presence of fraud, energy distributors are able to obtain relatively accurate estimates of the total electrical power and total energy actually consumed by the facility.

[0017] Furthermore, a power estimation method such as described above is proposed, wherein the reconstruction comprises the step of acquiring a first index offset corresponding to the expected theoretical phase shift, and for each first estimated voltage sample associated with the first index:

[0018] If the first index offset is an integer, assigning to the first estimated voltage sample the value of a second voltage sample having the first index offset from the first index offset as a second index;

[0019] Otherwise, the first estimated voltage sample is assigned an estimated value obtained from interpolation using the at least one second voltage sample.

[0020] In addition, a power estimation method such as described above is proposed, wherein the first index is offset by Δ1 such that:

[0021] Δ1=2N / 3 or Δ1=N / 3,

[0022] Where N is the number of first voltage samples per cycle of the first phase voltage,

[0023] And where if Δ1 is an integer, then the following occurs:

[0024]

[0025] In addition, a power estimation method such as that described above is proposed, including if Δ1 is not an integer, by and the second voltage sample A linear interpolation is performed between the steps to calculate the estimated value, and INT is the integer part function.

[0026] Additionally, a power estimation method such as described above is proposed, wherein the reconstruction comprises the following steps:

[0027] Acquiring a pre-recorded reference table containing reference samples of a sinusoidal signal;

[0028] measuring the maximum value of the second phase voltage and multiplying the reference sample by a crest factor proportional to the maximum value to obtain a resulting reference sample, and

[0029] For each first estimated voltage sample associated with the first index, the first estimated voltage sample is assigned the value of a resulting reference sample having as index the first index offset from an index offset corresponding to an expected theoretical phase shift between the first phase voltage and the second phase voltage in the absence of fraud.

[0030] In addition, a power estimation method such as that described above is proposed, wherein at least one first electrical power includes a first active power estimated from a first estimated voltage sample and a first current sample, an image of a first phase current circulating on a first phase over time, and / or a first reactive power estimated from level samples at the fundamental frequency of the first phase voltage and level samples at the fundamental frequency of the first phase current, and / or a first apparent power estimated from the effective value of the first phase voltage and the effective value of the first phase current.

[0031] In addition, a power estimation method such as described above is proposed, wherein the level samples at the fundamental frequency of the first phase voltage and the level samples at the fundamental frequency of the first phase current are obtained by applying low-pass filtering to the first estimated voltage samples and the first current samples, respectively.

[0032] In addition, a power estimation method such as that described above is proposed, wherein the low-pass filtering is a first-order Butterworth filter with a Z-transform H(Z), such that:

[0033] in:

[0034] a=0.7569801925.

[0035] In addition, a power estimation method such as that described above is proposed, which also includes the following steps: if the fraud also tampers with the measurement of the third phase voltage present on the third phase, the second voltage sample is used again to reconstruct the third voltage sample, the image of the third phase voltage over time, and at least one third electric power consumed on the third phase is estimated by using the third estimated voltage sample.

[0036] In addition, a power estimation method such as described above is proposed, which also includes the following steps: detecting whether the power distribution network includes a neutral line in addition to the three phases, and if this is not the case, calculating a second corrected voltage sample and calculating at least one second corrected electric power consumed on the second phase from the second corrected voltage sample.

[0037] In addition, a power estimation method such as described above is proposed, wherein the calculation of the second corrected voltage sample comprises the following steps: acquiring a second index offset corresponding to an expected theoretical phase shift of the measurement of the second phase voltage due to fraud; multiplying the second voltage sample by a predetermined factor to obtain a second result sample; and for each second corrected voltage sample associated with the second corrected index:

[0038] if the second index offset is an integer, assigning to the second corrected voltage sample the value of the second resultant voltage sample having the second corrected index offset from the second index offset;

[0039] Otherwise, the second corrected voltage sample is assigned an estimated value obtained from interpolation using the at least one second result voltage sample.

[0040] In addition, a power estimation method such as the one described above is proposed, wherein when the angle between the second phase voltage and the third phase voltage is The absence of a neutral line is detected when:

[0041]

[0042] X is a positive predefined value and is less than 30.

[0043] In addition, a power estimation method as described above is proposed, wherein fraud is detected when the first phase voltage is less than a predefined voltage threshold and the first phase current is greater than a predefined current threshold.

[0044] Furthermore, a three-phase electricity meter is proposed, comprising a processing component arranged to implement a power estimation method such as described above.

[0045] In addition, a computer program is proposed, comprising instructions for driving a processing component of a three-phase electricity meter such as described above to perform the steps of a power estimation method such as described above.

[0046] In addition, there is proposed a recording medium readable by a computer on which a computer program such as described above is recorded.

[0047] The present invention will best be understood from the following description of specific non-limiting embodiments of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Reference will be made to the accompanying drawings, in which:

[0049] [ Figure 1 ] Figure 1 A three-phase electricity meter in a four-wire configuration according to the present invention is shown;

[0050] [ Figure 2 ] Figure 2 A diagram is shown which includes a sinusoidal curve and enables the explanation of a linear interpolation method;

[0051] [ Figure 3 ] Figure 3 A three-phase electricity meter in a three-wire configuration according to the present invention is shown;

[0052] [ Figure 4 ] Figure 4 represents the Fresnel diagram of the voltage and current measured on each phase in a four-wire configuration in a normal situation (on the left-hand side) and in a fraudulent situation (on the right-hand side);

[0053] [ Figure 5 ] Figure 5 is with Figure 4 A similar diagram but in a three-wire configuration;

[0054] [ Figure 6 ] Figure 6 is a flow chart showing the steps of a power estimation method according to the present invention. DETAILED DESCRIPTION

[0055] See also Figure 1 , the power estimation method according to the first embodiment of the invention is implemented in a three-phase electricity meter 1, which is intended to measure the electric energy supplied to a user's facility 2 through a distribution network 3. The facility 2 is an industrial facility.

[0056] In this case, this is a four-wire configuration.

[0057] Thus, the distribution network 3 comprises a first phase Ph1, a second phase Ph2, a third phase Ph3 and a neutral line N. A circuit breaker 4 is placed at the "border" between the distribution network 3 and the installation 2. The three phases Ph1, Ph2, Ph3 and the neutral line N enter the circuit breaker 4 and emerge from it towards the installation 2. The circuit breaker 4 enables the electrical energy distribution to be cut off.

[0058] The meter 1 comprises four voltage input terminals Ue1, Ue2, Ue3, UeN and three current input terminals Ie1, 1e2, Ie3 (each current input terminal comprises two ports).

[0059] The voltage input terminal Ue1 is connected to the first phase Ph1 through the wire f1, the voltage input terminal Ue2 is connected to the second phase Ph2 through the wire f2, the voltage input terminal Ue3 is connected to the third phase Ph3 through the wire f3, and the voltage input terminal UeN is connected to the neutral line N through the wire fN.

[0060] The current input terminal Ie1 is connected to the external current transformer CT1 installed on the first phase Ph1, the current input terminal 1e2 is connected to the external current transformer CT2 installed on the second phase Ph2, and the current input terminal Ie3 is connected to the external current transformer CT3 installed on the third phase Ph3. The external current transformer is located outside the meter 1.

[0061] In this case, each external current transformer has a transformation ratio equal to 2000.

[0062] In order to realize the metering measurement, the meter 1 collects the images of the phase currents I1, I2, I3 circulating on the phases Ph1, Ph2 and Ph3 at the current inputs Ie1, Ie2, Ie3 and due to the external transformers CT1, CT2 and CT3. The meter 1 also collects the phase voltages V1, V2, V3 present on the phases Ph1, Ph2, Ph3 and the neutral voltage VN present on the neutral line N at the voltage inputs Ue1, Ue2, Ue3 and UeN.

[0063] The meter 1 also comprises a measuring assembly 5 comprising a front assembly 5a and an analog-to-digital converter 5b (or a single analog-to-digital converter with a sufficient number of input terminals).

[0064] The front assembly 5a is connected to the current input terminal and the voltage input terminal. The front assembly 5a enables the current on the current input terminal to be converted into a voltage suitable for the analog-to-digital converter 5b. The front assembly 5a also enables the voltage on the voltage input terminal of the analog-to-digital converter 5b to be transformed.

[0065] The device 1 further comprises a processing component 6 .

[0066] The processing component 6 is adapted to execute the instructions of the program to implement the power estimation method according to the present invention. The program is stored in a memory 7, which is integrated in or connected to the processing component 6. The processing component 6 is, for example, a conventional processor, a microcontroller, a DSP (digital signal processor) or a programmable logic circuit, such as an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit).

[0067] In this case, the processing component 6 is a microcontroller 6 .

[0068] The microcontroller 6 is connected to the analog-to-digital converter 5 b and acquires the measurement samples generated by the analog-to-digital converter 5 b.

[0069] The microcontroller 6 thus acquires a first voltage sample and the first current sample The microcontroller 6 also collects the second voltage sample and the second current sample The microcontroller 6 also collects the third voltage sample. and the third current sample The images of the third phase voltage V3 and the second phase current I3 respectively over time.

[0070] At regular intervals, the microcontroller 6 attempts to detect fraudulent cutting of one or more wires f1, f2, f3 that falsifies the measurement of one or more phase voltages.The microcontroller 6 detects the cutting of a wire of a phase in the absence of voltage and the presence of current on said phase.

[0071] For example, for the first phase Ph1 , when the first phase voltage is less than a predefined voltage threshold and the first phase current is greater than a predefined current threshold, the microcontroller 6 detects fraud.

[0072] In this case, the microcontroller 6 calculates V1 as the effective value of the first phase voltage V1. RMS and I1 as the effective value of the first phase current I1 RMS , and fraud on the first phase Ph1 is detected when:

[0073] V1 RMS <Vf

[0074] And I1 RMS >If,

[0075] Vf is a predefined voltage threshold and If is a predefined current threshold.

[0076] Advantageously, If is chosen such that:

[0077] Istart<If<Imin,

[0078] Istart is the current value at which the meter 1 starts measuring the consumed energy, and Imin is the current value at which the meter 1 must measure the current with nominal accuracy.

[0079] In this case, the following situation occurs: Istart=1 mA and Imin=10 mA.

[0080] Therefore, for example, If=5 mA is selected.

[0081] For example, select Vf=40V.

[0082] This fraud detection step is performed in the same way for the other two phases.

[0083] It is now assumed that the microcontroller 6 has detected fraud (only) on the first phase Ph1 .

[0084] The measurement of the first phase voltage V1 is therefore fraudulently tampered with, and the first voltage sample is unusable.

[0085] Therefore, the microcontroller 6 will estimate at least one first electric power consumed on the first phase Ph1 by using the voltage measurement on a reference phase not subject to fraud. In this case, for example, the reference phase is the second phase Ph2 (but could also be the third phase Ph3).

[0086] To this end, the microcontroller 6 obtains the second voltage sample The image of the second phase voltage V2 over time reconstructs the first estimated voltage sample A graph of the first phase voltage V1 over time.

[0087] The reconstruction is performed by successive measurement cycles, each measuring cycle having a duration of, for example, 1 s.

[0088] Each phase voltage has a network frequency equal to 50 Hz, and the network period is therefore equal to 20 ms.

[0089] For each network cycle the number of samples is recorded N (thus, N is the number of first voltage samples per cycle of the first phase voltage), and σ is the number of network cycles per measurement cycle. In this case, the following situation occurs: σ=50. Therefore, there are σN samples per 1s measurement cycle.

[0090] Advantageously, N is a multiple of 4. In this case, for example, N=52.

[0091] Therefore, the sampling frequency f E So that:

[0092] f E =2600Hz.

[0093] The microcontroller 6 first acquires a first index offset, which corresponds to the expected theoretical phase shift between the first phase voltage V1 and the second phase voltage V2 in the absence of fraud. The first index offset has been pre-recorded in the memory 7.

[0094] The expected theoretical phase shift is equal to +120°.

[0095] By considering the expected theoretical phase shift, and by considering that an index shift equal to the total number of samples per network cycle (ie, N) will correspond to a phase shift of 360°, the first index shift Δ1 is the number of samples where the second voltage sample must be offset to produce a voltage having the first phase voltage V1 in this phase.

[0096] In this case, the following occurs:

[0097] Δ1=2N / 3.

[0098] For each first estimated voltage sample If the first index offset Δ1 is an integer, the microcontroller 6 assigns the value of the second voltage sample to the first estimated voltage sample, the second voltage sample having the first index n offset from the first index offset Δ1 as the second index. Otherwise, the microcontroller 6 assigns the value of the second voltage sample to the first estimated voltage sample. A value estimated from interpolation using at least one second voltage sample is assigned.

[0099] Therefore, if 2N / 3 is an integer, there is a second voltage sample And the microcontroller 6 thinks:

[0100]

[0101] Therefore, using a known sample But it shifts by 2 / 3 of the network cycle over time.

[0102] However, if Δ1 is not an integer, then does not exist, and the microcontroller 6 uses at least one second existing voltage sample to generate a first estimated voltage sample The estimated value of .

[0103] In this case, the microcontroller 6 generates a voltage sample at the second voltage level. With the second voltage sample between and Linear interpolation is performed between the values ​​to calculate the estimated value.

[0104] INT is the "round" function.

[0105] The microcontroller 6 thus determines the sample at the precise instant n-2N / 3 The value of .

[0106] See also Figure 2 , by definition, the time offset between each sample is equal to T E (sample period), that is, for f E =2600Hz is 384.62μs.

[0107] The principle of linear interpolation consists in passing through two points and A2 n-INT(2N3).TE; V2n-INT(2N3) tangent) to estimate the time T E The corresponding sine point at

[0108] Therefore, the coordinates The point A determines the vertical coordinate.

[0109] The equation of the tangent line through these two points is of the type y = a.x + b and is very easy to determine. The ordinate on the tangent line corresponding to point A, the x-axis of the tangent line is Therefore, it is shown that:

[0110]

[0111] Here, in this case, 2N / 3 is not an integer, and estimation by linear interpolation is systematically performed to reconstruct each first estimated voltage sample.

[0112] Every second, the microcontroller 6 therefore calculates the value of at least one first electrical power consumed on the first phase V1 .

[0113] In this case, at least one first electric power calculated by the microcontroller 6 comprises a first estimated voltage sample and from the first current sample An estimated first active power, a first reactive power estimated from level samples at the fundamental frequency of the first phase voltage V1 and the first phase current I1, and a first apparent power estimated from the effective value of the first phase voltage V1 and the effective value of the first phase current I1. A measure of the energy consumed over time is derived from these power estimates.

[0114] In this case, the microcontroller 6 therefore calculates:

[0115] (first active power);

[0116] (first reactive power);

[0117] S1=V1 RMS .I1 RMS (First apparent power).

[0118] is the nth sample of the level at the fundamental frequency of the first phase voltage V1.

[0119] is the nth sample of the level at the fundamental frequency of the first phase current I1.

[0120] These values ​​V1 RMS and I1 RMS From the sample and Obtained by calculation.

[0121] It will therefore be appreciated that the interest in choosing the number of samples N representing four times the network period enables the calculation of reactive power to be facilitated by simply shifting the level samples by a quarter of the period at the fundamental frequency of the first phase voltage V1 .

[0122] Level samples at the fundamental frequency of the first phase voltage V1 and the level samples at the fundamental frequency of the first phase current I1 Respectively through the (reconstructed) first estimated voltage sample and the first current sample obtained by low-pass filtering.

[0123] In this case, the low-pass filtering uses a first-order Butterworth filter.

[0124] The Z transform of a first-order Butterworth filter is represented by the following relationship:

[0125] in

[0126] Advantageously, in this case, the following is selected:

[0127] a=0.7569801925,

[0128] It is possible to obtain a 3dB cutoff frequency of 70 Hz when the network frequency f1 is such that: f1 = 50 Hz.

[0129] Interestingly, noting that the network frequency f1, knowing the sample frequency f E So that fE=N.f1 (where N=52 in this case), Therefore, the gain and phase shift of the filter at frequency f1 are are the same.

[0130] In the case where the filter is linear, by noting the input to the filter, E(Z), and the output of the filter, S(Z), the following relationship emerges:

[0131] S(Z)=H(Z).E(Z)

[0132] At the input of the filter, the samples In the case of digital filtering, samples are given at the output The following relationships emerge:

[0133] Where a=0.7569801925.

[0134] and Initialized to 0.

[0135] Likewise, for the other filters described below, the following relationships occur:

[0136] Where a=0.7569801925

[0137] and Initialized to 0.

[0138] Where a=0.7569801925.

[0139] and Initialized to 0.

[0140] Where a=0.7569801925.

[0141] and Initialized to 0.

[0142] Where a=0.7569801925.

[0143] and Initialized to 0.

[0144] Where a=0.7569801925.

[0145] and Initialized to 0.

[0146] If the fraud also tampered with the measurement of the third phase voltage V3 present on the third phase Ph3, the only available reference phase is the second phase Ph2.

[0147] The microcontroller 6 therefore first reconstructs a first estimation sample for the first phase voltage V1 and calculates as above:

[0148] (first active power);

[0149] (first reactive power);

[0150] S1=V1RMs .I1 RMS (apparent power).

[0151] The microcontroller 6 then uses the second voltage sample again To reconstruct the third estimated voltage sample And by using the third estimated voltage sample To estimate at least one third electric power consumed on the third phase Ph3.

[0152] To this end, the microcontroller 6 first acquires a third index offset Δ3 , which corresponds to the expected theoretical phase shift between the third phase voltage V3 and the second phase voltage V2 in the absence of fraud.

[0153] The expected theoretical phase shift is equal to -120°.

[0154] By considering the expected theoretical phase shift, and by considering that an index shift equal to the total number of samples per network cycle (ie, N) will correspond to a phase shift of 360°, the third index shift Δ3 is the number of samples where the second voltage sample must be offset to generate a phase voltage having a third phase voltage V3.

[0155] So, in this case, the following occurs:

[0156] Δ3=N / 3.

[0157] If the index offset is an integer, that is, if N / 3 is an integer, then there is a second voltage sample And the microcontroller 6 thinks:

[0158]

[0159] However, if Δ3 is not an integer, there is no And the microcontroller 6 uses at least one second existing voltage sample to generate a third estimated voltage sample The estimated value of .

[0160] In this case, the microcontroller 6 generates a voltage sample at the second voltage level. With the second voltage sample between and Linear interpolation is performed between the values ​​to calculate the estimated value.

[0161] The microcontroller 6 thus determines the sample at the precise instant nN / 3 The value of .

[0162] Here, in this case, N / 3 is not an integer, and estimation by linear interpolation is systematically performed.

[0163] The microcontroller 6 therefore calculates:

[0164] (Third active power);

[0165] (third reactive power);

[0166] S3=V3 RMS .I3 RMS (Third apparent power).

[0167] By applying a Butterworth filter such as that described above the samples are calculated and These values ​​V3 RMS and I3 RMS From the sample and Obtained by calculation.

[0168] refer to Figure 3 , this time a three-wire configuration is used.

[0169] The power distribution network 3 includes a first phase Ph1 , a second phase Ph2 , and a third phase Ph3 , but has no neutral line.

[0170] The meter 1 comprises three voltage input terminals Ue1, Ue2, Ue3 and three current input terminals Ie1, 1e2, Ie3 (each current input terminal comprises two ports).

[0171] The meter 1 again comprises a measuring element 5a and an analog-to-digital converter 5b, as well as a microcontroller 6 connected to the analog-to-digital converter 5b.

[0172] Again, use is made of the case in which the wire f1 has been cut by a malicious user.

[0173] Therefore, the following occurs:

[0174] V1 RMS <40V

[0175] And I1 RMS >5mA.

[0176] The microcontroller 6 has therefore detected a fraud which has tampered with the measurement of the first phase voltage V1 .

[0177] It is assumed that the fraudster has not tampered with the measurement of the second phase voltage V2, nor has he tampered with the measurement of the third phase voltage V3.

[0178] The implementation of the present invention is slightly different in a three-wire configuration.

[0179] For different reasons, Figure 4 and5 is illustrated in the diagram.

[0180] Figure 4 Indicated in Figure 1 Fresnel plots of voltage and current in the normal case (on the left-hand side) and in the fraudulent case involving cutting wire f1 (on the right-hand side) in the four-wire configuration in FIG.

[0181] Observe that the absence of a voltage measurement (in this case the V1 measurement) has no effect on the amplitude and angle of the other voltages. In the absence of two voltages, the same applies to the remaining voltages.

[0182] Figure 5 Indicated in Figure 3 Fresnel diagrams of voltage and current in the normal case (on the left-hand side) and in the fraudulent case (on the right-hand side) involving cutting wire f1 in the three-wire configuration in FIG.

[0183] Observe that no voltage measurement (in this case the V1 measurement) has any effect on the measured amplitude (the amplitude is proportional to decrease), but will also have an effect on the measured angles of the other voltages (V2 is measured at +30°, while V3 is measured at -30°).

[0184] Therefore, the microcontroller 6 first preliminarily detects whether the power distribution network 3 includes a neutral line other than the three phases.

[0185] If this is not the case, i.e. if the configuration is a three-wire configuration, the microcontroller 6 takes into account the factor and phase shifts of +30° and -30° to estimate electrical power.

[0186] The microcontroller 6 will first calculate the second corrected voltage sample and the third corrected voltage sample

[0187] Therefore, the microcontroller 6 will calculate at least one second corrected electric power consumed on the second phase Ph2 (in this case, the second active power, the second reactive power, and the second apparent power) from the second corrected voltage sample, and calculate at least one third corrected electric power consumed on the third phase Ph3 (in this case, the third active power, the third reactive power, and the third apparent power) from the third corrected voltage sample.

[0188] To calculate the second corrected voltage sample The microcontroller 6 acquires a second index offset Δ2 which corresponds to the expected theoretical phase shift of the measurement of the second phase voltage V2 due to fraud.

[0189] For the second phase, the theoretical phase shift is equal to 30°, and therefore the second index is shifted by Δ2 such that:

[0190] Δ2=11N / 12.

[0191] The microcontroller 6 first converts the second voltage sample Multiply by a predetermined factor, which in this case is equal to This factor enables compensation of the amplitude which has been reduced by the inverse factor. A second result sample is obtained.

[0192] For each second corrected voltage sample associated with the second corrected index If the second index offset is an integer, the microcontroller 6 sends a second corrected voltage sample The value of the second result voltage sample is assigned to the second result voltage sample, which has the second corrected index offset from the second index offset as the second index. Otherwise, the microcontroller 6 will assign the value of the second corrected voltage sample to the second result voltage sample. A value estimated using interpolation from at least one second result voltage sample is assigned.

[0193] Thus, it is possible to obtain, without calculating linear interpolation, that the “ideal” time shift of the second corrected sample corresponds to the exact instant n−11N / 12 at which the second voltage sample was present.

[0194] Therefore, N must be divisible by 12, which is not the case here when N=52, so that linear interpolation is systematically calculated.

[0195] Similarly, to calculate the third corrected voltage sample The microcontroller 6 acquires a third index offset, which corresponds to an expected theoretical phase shift of the measurement of the third phase voltage V3 due to fraud.

[0196] For the third phase, the theoretical phase shift is equal to -30°, so the third index is shifted by Δ3 such that:

[0197] Δ3=N / 12.

[0198] Therefore, the microcontroller 6 converts the third voltage sample Multiply by a predetermined factor, which in this case is equal to This factor enables compensation of the amplitude which has been reduced by the inverse factor. A third result sample is obtained.

[0199] For each third corrected voltage sample associated with the third corrected index If the third index offset is an integer, the microcontroller 6 assigns the value of the third result voltage sample to the third corrected voltage sample, which has the third corrected index offset from the third index offset as the third index. Otherwise, the microcontroller 6 assigns the value estimated using interpolation from at least one third result voltage sample to the third corrected voltage sample.

[0200] Therefore, it is possible to obtain, without calculating linear interpolation, the ideal acquisition time offset of the third corrected sample corresponding to the exact instant nN / 12 at which the third voltage sample was obtained.

[0201] Therefore, N must be divisible by 12, which is not the case here when N=52, so that linear interpolation is systematically calculated.

[0202] The microcontroller 6 then calculates the second corrected electric power consumed on the second phase Ph2 and the third corrected electric power consumed on the third phase Ph3:

[0203] (second corrected active power);

[0204] (second corrected reactive power);

[0205] S2=U2 RMS .I2 RMS (second corrected apparent power);

[0206] (third corrected active power);

[0207] (third corrected reactive power);

[0208] S3=U3 RMS .I3 RMS (Third corrected apparent power).

[0209] sample and is obtained by respectively sampling the corrected voltage at the second and the second current sample is calculated by applying the Butterworth filter as described above.

[0210] Likewise, the sample and is obtained by respectively applying the third corrected voltage sample and the third current sample is calculated by applying the Butterworth filter as described above.

[0211] Therefore, the first estimated voltage sample is reconstructed by applying exactly the same method as in the four-wire configuration over all 1s periods. However, this time the second corrected voltage sample used

[0212] Therefore, the microcontroller 6 calculates the first electric power consumed on the first phase Ph1 as follows:

[0213] (first active power);

[0214] (first reactive power);

[0215] S1=V1 RMS .I1 RMS (First apparent power).

[0216] sample and is obtained by respectively first corrected voltage samples and the first current sample is calculated by applying the Butterworth filter as described above.

[0217] As can be seen, the microcontroller must detect whether the configuration is a four-wire or three-wire configuration, i.e. whether the power distribution network 3 includes or does not include a neutral line in addition to the three phases. The estimation of the first electric power is actually different depending on the actual configuration. Likewise, the correction applied on the second and third electric powers depends on the configuration.

[0218] The power estimation method therefore comprises a preliminary step consisting in detecting whether the network 3 and the meter 1 are in a four-wire or three-wire configuration.

[0219] To this end, after having detected fraud on the first phase Ph1 , the microcontroller 6 measures the angle between the two remaining voltages, namely between the second phase voltage V2 and the third phase voltage V3 (In the case where the measurement of the second phase voltage V2 and the third phase voltage V3 has not been tampered with).

[0220] Therefore, the microcontroller 6 determines the angle Is it at an absolute value close to 180°, if this is the case, a three-wire configuration is detected.

[0221] Microcontroller 6 verification angle Is it a predefined interval?

[0222]

[0223] X is a positive predefined value and is less than 30. For example, the following situation occurs: X=20.

[0224] If the angle belongs to the predefined interval, the microcontroller 6 considers the angle In an absolute value close to 180°, and thus a three-wire configuration is detected. Otherwise, the microcontroller 6 detects a four-wire configuration.

[0225] Note that in a three-wire configuration it is possible to estimate the power consumed without fraud only if only one voltage input is lost (fraud on a single phase, as in three-wire, if two phases are lost the meter is no longer powered), while in a four-wire configuration it is possible to estimate them in the case of loss of two voltage inputs.

[0226] It should also be noted that in the event that the microcontroller 6 detects the loss of two voltage inputs (fraud on two phases), the microcontroller 6 may systematically detect a four-wire configuration (because the meter 1 is powered); therefore, the meter 1 performs the estimation method under the corresponding conditions. In fact, when the configuration is a three-wire configuration and two voltages are lost, the meter automatically shuts down.

[0227] The different steps of the estimation method can be found in Figure 6 As seen in.

[0228] The method starts at step E0.

[0229] The microcontroller 6 attempts to detect at least one fraud on one of the three phases (step E1 ).

[0230] If no fraud is detected, the microcontroller 6 calculates estimates of the active power, reactive power and apparent power for each of the three phases, typically by using the voltage and current samples sent by the analog-to-digital converter 5 b (step E2 ). It is not necessary to reconstruct the voltage samples.

[0231] The calculation of the power estimate is performed at fixed intervals (eg, a period of 1 s).

[0232] The method returns to step E1.

[0233] If, in step E1 , the microcontroller 6 detects fraud on both phases, the microcontroller 6 infers directly from this that the configuration is a four-wire configuration (see what has just been explained).

[0234] For example, assume that the measurement of the first phase voltage V1 and the measurement of the third phase voltage V3 are tampered with.

[0235] The microcontroller 6 first reconstructs an image of the first voltage sample, the second phase voltage (reference voltage), estimated from the second voltage sample. Thus, the microcontroller 6 estimates the first electric power consumed on the first phase (step E3).

[0236] Then, the microcontroller 6 reconstructs a third voltage sample estimated from the image of the second voltage sample, the second phase voltage (reference voltage). Thus, the microcontroller 6 estimates the third electric power consumed on the third phase (step E4).

[0237] The calculation of the power estimate is performed at fixed intervals (eg, a period of 1 s).

[0238] The method returns to step E1.

[0239] It is now assumed that in step E1 the microcontroller 6 detects fraud on a single phase; it is assumed that the measurement of the first phase voltage has been falsified.

[0240] The microcontroller 6 thus detects whether the network 3 and the meter 1 are in a four-wire or three-wire configuration (step E5 ).

[0241] If a three-wire configuration is detected, the microcontroller 6 generates a second corrected voltage sample (amplitude and phase corrected) of the second phase Ph2 and uses the second corrected voltage sample to calculate the second corrected electric power consumed on the second phase Ph2. Similarly, the microcontroller 6 generates a third corrected voltage sample of the third phase Ph3 and uses the third corrected voltage sample to calculate the third corrected electric power consumed on the third phase Ph3: step E6.

[0242] Then, the microcontroller 6 reconstructs the first voltage sample estimated from the second corrected voltage sample (or from the third corrected voltage sample) and thereby estimates the first electric power consumed on the first phase (step E7 ).

[0243] The calculation of the power estimate is performed at fixed intervals (eg, a period of 1 s).

[0244] The method returns to step E1.

[0245] In step E5 , if a four-wire configuration is detected, the microcontroller 6 reconstructs the first voltage sample estimated from the second voltage sample (or from the third voltage sample) and thus estimates the first electrical power consumed on the first phase (step E8 ).

[0246] The calculation of the power estimate is performed at fixed intervals (eg, a period of 1 s).

[0247] The method returns to step E1.

[0248] Now, a power estimation method according to a second embodiment of the present invention is described.

[0249] Again it is considered that the fraud has tampered with the measurement of the first phase voltage V1 present on the first phase Ph1 , without tampering with the measurement of the second phase voltage V2 present on the second phase Ph2 . The second phase voltage V2 is again a reference voltage.

[0250] The main difference between the second embodiment and the first embodiment is related to the reconstruction step, which is implemented to reconstruct the first estimated voltage samples.

[0251] A reference table has been recorded in the memory 7. The reference table contains reference samples of a sine function, in this case a cosine function.

[0252] Each reference sample is written as cos(θi), and therefore has a value that is a function of the value of the cosine (y-axis) of the angle θi (x-axis).

[0253] The following situation occurs: cos(θ0)=1.

[0254] The value of i corresponds to the position of the value cos(θi) in the table. The value of i is also the index of the reference sample cos(θi).

[0255] For example, the pitch angle θi is equal to 0.1° (ie, for all i, θi+1, θi=0.1°).

[0256] Note that it is sufficient for the reference table to contain reference samples for i varying from 0° to 90°. 1 A value of / 4 periods is indeed sufficient to reconstruct the entire period by horizontal symmetry along the x-axis and vertical symmetry along the y-axis.

[0257] To reconstruct the first estimated voltage sample, the microcontroller 6 first acquires a reference table.

[0258] Therefore, the microcontroller 6 measures the maximum value of the second phase voltage V2 and multiplies the reference sample by an amplitude factor proportional to the maximum value to obtain a result reference sample.

[0259] In this case, the crest factor is equal to the maximum value V2 max , which makes:

[0260] In a four-wire configuration, and

[0261] In a three-wire configuration.

[0262] Then, for each first estimated voltage sample associated with the first index, the microcontroller 6 assigns to said first estimated voltage sample the value of a resulting reference sample having as index the first index offset from an index offset corresponding to the expected theoretical phase shift between the first phase voltage and the second phase voltage in the absence of fraud.

[0263] Therefore, the microcontroller 6 shifts the reference samples contained in the reference table by ±120° (±0.1°).

[0264] This approach makes it possible to avoid any calculations and save a lot of CPU time.

[0265] Therefore, the microcontroller 6 estimates the first electric power consumed on the first phase by using the first estimated voltage sample.

[0266] In a variant, regardless of the configuration (four-wire or three-wire), in the case where the fraud affects only the first phase Ph1, it is possible to estimate the first electric power, reconstructing the first estimated voltage sample from both the second voltage sample (second phase voltage V2) and the third voltage sample (third phase voltage V3).

[0267] To this end, the microcontroller 6 calculates an arithmetic mean of the first estimated voltage sample reconstructed from the second voltage sample and the first estimated voltage sample reconstructed from the third voltage sample to generate a first unified voltage sample for estimating the first electric power.

[0268] Naturally, the invention is not limited to the embodiments described but encompasses any variant coming within the scope of the invention, such as defined by the claims.

[0269] This shows that the expected theoretical phase shift between the first phase voltage and the second phase voltage in the absence of fraud is equal to +120°. Naturally, this value depends on the phases called "first phase" and "second phase"; the expected theoretical phase shift may be equal to -120°. Therefore, the first index offset Δ1 can be such that:

[0270] Δ1=N / 3

[0271] Likewise, the expected theoretical phase shift due to fraudulent measurement of the second phase voltage may be equal to -30°, and the expected theoretical phase shift due to fraudulent measurement of the third phase voltage may be equal to +30°.

[0272] The architecture of the instrument may differ from that described in this case. The analog-to-digital converter(s) may thus be integrated into a microcontroller, for example.

Claims

1. A power estimation method implemented in a three-phase electricity meter (1), and the method comprises the following steps: Detecting fraud that tampers with the measurement of a first phase voltage (V1) present on a first phase (Ph1) of a power distribution network (3) without tampering with the measurement of a second phase voltage (V2) present on a second phase (Ph2) of the power distribution network (3); reconstructing a first estimated voltage sample, an image of the first phase voltage (V1) over time, from a second voltage sample, an image of the second phase voltage (V2) over time, the second voltage sample being offset over time by an expected theoretical phase shift between the first phase voltage and the second phase voltage in the absence of fraud; At least one first electrical power consumed on the first phase is estimated by using the first estimated voltage sample.

2. The power estimation method according to claim 1, characterized in that: The reconstruction comprises the steps of acquiring a first index offset corresponding to the expected theoretical phase shift, and for each first estimated voltage sample associated with the first index: If the first index offset is an integer, assigning to the first estimated voltage sample the value of a second voltage sample having as a second index the first index offset from the first index offset; Otherwise, the first estimated voltage sample is assigned an estimated value obtained from interpolation using at least one second voltage sample.

3. The power estimation method according to claim 2, characterized in that: The first index is offset by Δ1 such that: Δ1=2N / 3 or Δ1=N / 3, Where N is the number of first voltage samples per cycle of the first phase voltage, And where if Δ1 is an integer, then the following occurs:

4. The power estimation method according to claim 3, characterized in that: If Δ1 is not an integer, then and the second voltage sample Linear interpolation is performed between the steps to calculate the estimated value, and INT is a rounding function.

5. The power estimation method according to claim 1, wherein: The reconstruction comprises the following steps: Acquiring a pre-recorded reference table containing reference samples of a sinusoidal signal; measuring a maximum value of the second phase voltage and multiplying the reference sample by an amplitude factor proportional to the maximum value to obtain a resulting reference sample, and For each first estimated voltage sample associated with a first index, the first estimated voltage sample is assigned the value of a resulting reference sample having as index the first index offset from an index offset corresponding to an expected theoretical phase shift between the first phase voltage and the second phase voltage in the absence of fraud.

6. The power estimation method according to any one of claims 1 to 5, characterized in that: The at least one first electrical power includes a first active power estimated from a first estimated voltage sample and from a first current sample, an image of the first phase current (I1) circulating on the first phase over time, and / or a first reactive power estimated from a level sample at the fundamental frequency of the first phase voltage (V1) and a level sample at the fundamental frequency of the first phase current (I1), and / or a first apparent power estimated from the effective value of the first phase voltage (V1) and the effective value of the first phase current (I1).

7. The power estimation method according to claim 6, characterized in that: The level samples at the fundamental frequency of the first phase voltage (V1) and the level samples at the fundamental frequency of the first phase current (I1) are obtained by applying low-pass filtering to the first estimated voltage samples and the first current samples, respectively.

8. The power estimation method according to claim 7, characterized in that: The low-pass filtering is a first-order Butterworth filter with a Z-transform H(Z) such that: in: a＝0.7569801925。 9. The power estimation method according to any one of claims 1 to 5, characterized in that: The following steps are also included: If the fraud also tampered with the measurement of the third phase voltage (V3) present on the third phase (Ph3), the second voltage sample is again used to reconstruct a third estimated voltage sample, an image of the third phase voltage over time, and at least one third electric power consumed on the third phase is estimated by using the third estimated voltage sample.

10. The power estimation method according to any one of claims 1 to 5, characterized in that: The following steps are also included: It is detected whether the power distribution network (3) comprises a neutral line (N) in addition to the three phases, and if this is not the case, a second corrected voltage sample is calculated and at least one second corrected electric power consumed on the second phase is calculated from the second corrected voltage sample.

11. The power estimation method according to claim 10, characterized in that: The calculation of the second corrected voltage sample comprises the following steps: acquiring a second index offset corresponding to an expected theoretical phase shift of the measurement of the second phase voltage (V2) due to fraud; multiplying the second voltage sample by a predetermined factor to obtain a second result sample; and for each second corrected voltage sample associated with a second corrected index: if the second index offset is an integer, assigning to the second corrected voltage sample the value of a second result voltage sample having as a second index a second corrected index offset from the second index offset; Otherwise, the second corrected voltage sample is assigned an estimated value obtained from interpolation using at least one second result voltage sample.

12. The power estimation method according to claim 9, characterized in that: When the angle between the second phase voltage (V2) and the third phase voltage (V3) The absence of neutral (N) is detected when: X is a positive predefined value and is less than 30.

13. The power estimation method according to any one of claims 1 to 5, characterized in that: Fraud is detected when the first phase voltage is less than a predefined voltage threshold and the first phase current is greater than a predefined current threshold.

14. A three-phase electricity meter comprising a processing component (6) arranged to implement a power estimation method as claimed in one of claims 1 to 13.

15. A computer-readable recording medium having a computer program including instructions recorded thereon, wherein the instructions cause the processing component (6) of the three-phase electricity meter (1) according to claim 14 to execute the steps of the power estimation method according to any one of claims 1 to 13.

16. A computer program product comprising instructions for causing a processing component (6) of a three-phase electricity meter (1) according to claim 14 to execute the steps of a power estimation method according to one of claims 1 to 13.

Images