Method for estimating a signal property

Abstract

A method for estimating a property of a signal (1) sensed in an electrical system, comprises the steps of sensing the signal (1) and estimating a fundamental period of a fundamental of the signal (1) by comparing the sensed signal (1) with at least one threshold (2) to detect threshold crossings and estimating the fundamental period from the threshold crossings. The signal property is then estimated from the fundamental period and/or from the sensed signal (1) during an interval of a length of the fundamental period. An electronic device according to the invention comprises a micro controller and/or an analogue circuitry which performs the method for estimating a property of a signal. Preferably, the micro controller and/or analogue circuitry controls other parts of the electronic device depending on the results obtained by the method for estimating a property of a signal.

Claims

1. A method for protection of a switched mode power supply, a) whereby a property of a signal x(t) sensed in an electrical system is estimated by i) sensing the signal; ii) estimating a fundamental period of a fundamental of the signal by (1) comparing the sensed signal with at least one threshold to detect threshold crossings, and (2) estimating the fundamental period from the threshold crossings; b) initializing a tracker variable with a starting value when the signal crosses a first threshold in a first direction; c) modifying the value of the tracker variable by a first mathematical operation linking a first weighting value and the value of the tracker variable when the signal crosses a second threshold in the first direction; d) modifying the value of the tracker variable by a second mathematical operation linking a second weighting value and the value of the tracker variable when the signal crosses the second threshold in a second direction; e) defining at least two points of time E.sub.i-1, E.sub.i, at which the value of the tracker variable fulfils a trigger-condition; f) estimating the signal property based on a) a distance in time between two subsequent points of time, Di=Ei−Ei−1 and/or b) the signal x(t) sensed between two subsequent points of time; and g) whereby the estimated property of the signal is used to protect the switched mode power supply by adapting internal switching frequencies or switching parts of the switched mode power supply off.

2. The method according to claim 1, whereby the sensed signal is compared with two or more thresholds to detect the threshold crossings.

3. The method according to claim 1, comprising the step of integrating a function f(x(t)) corresponding to the signal x(t), over time between the two points of time E.sub.n, E.sub.m.

4. The method according to claim 1, whereby the trigger-condition is an equality of the value of the tracker variable and a comparison value.

5. The method according to claim 1 whereby, a) the first and the second threshold are both at a positive signal value and the first threshold is greater than the second threshold, b) the first direction is negative, c) the second direction is positive.

6. The method of claim 1 further including a method for creating an event record, whereby the points of time are stored in an event record, wherein the event record comprises a list of all detected point of time E.sub.i.

7. The method of claim 1 further including a method for creating a distances record, comprising the steps of a) sorting the points of time by ascending time of occurrence and numbering the points of time in order with the ascending time of occurrence, thereby assigning each point of time a number, if the numbers of the points of times or numbers of distances are stored in the distance record, b) determining distances D between every two subsequent points of time E.sub.i, E.sub.i-1 by subtracting an earlier point of time from a later point of time, c) assigning each distance D.sub.i a number equal to the number of the later point of time, if the numbers of distances are stored in the distance record, d) storing the distances D.sub.i in the distances record, a. which is ordered and the distances are stored in the order of the occurrence of the earlier or the later of the two points of time between which the respective distance was evaluated or b. which comprises data-tuples and each data-tuple comprises one of the distances and the number of the earlier or the later of the two points of time between which the respective distance was evaluated  or c. which is ordered and the distances are stored in the order of the numbers of the distances  or d. which comprises data-tuples and each data-tuple comprises one of the distances and the number of the one of the distances which is comprised in the respective data-tuple.

8. The method of claim 7 further including a method for determining a sliding and weighted half period of a distance D.sub.X, comprising the steps of: a) determining distances D.sub.0, . . . D.sub.X, or reading the distances record, b) setting a value of the sliding and weighted half period at distance D.sub.X (SWHP(D.sub.X)) to a maximum value of a list containing a. a fixed quantify A of distances D.sub.i divided by a divider Div, i. whereby 1. the distances D.sub.i have a number smaller or equal to the number of D.sub.X 2. or the distances D.sub.i are in a sorted distances record in time before or at the same record as the distance D.sub.X, 3. or the distances D.sub.i occur before or at the distance D.sub.X b. and a minimum distance, D.sub.min.

9. The method of claim 8 further including a method for determining if a distance is reliable, comprising the steps of: a) determining distances D.sub.0, . . . D.sub.X or by reading the distances record, b) determining a distance Y between a first distance D.sub.X+1 with the number Nr(D.sub.X+1) and a second distance D.sub.Fi with the number Nr(D.sub.Fi), a. by addition of all distances with numbers between the number of the second distance D.sub.FiNr(D.sub.Fi) plus+1 and the number of the first distance D.sub.X+1 Nr(D.sub.X+1), b. whereby the second distance D.sub.Fi is a reliable distance and whereby the second distance D.sub.Fi has the largest number of all reliable distances which is smaller than the number of the first distance D.sub.X+1 Nr(D.sub.X+1), c) defining the first distance D.sub.X+1 to be reliable if a. the distance Y is larger or equal the sliding and weighted half period at the first distance D.sub.X+1, (SWHP(D.sub.X+1)), or b. if the distance Y is larger than a maximum distance D.sub.max.

10. The method according to claim 9, whereby the signal property to be estimated is an effective signal X.sub.eff, which is estimated by a) integrating the square of the sensed signal x(t) between the later one of the two points of time used to evaluate a first distance D.sub.m and the later one of the two points of time used to evaluate a second distance D.sub.n, and by b) dividing this integrated signal through the distance between the first and the second distance.

11. The method of claim 10 further including determining an improved effective signal X.sub.eff,imp, whereby a floating average of an even number of subsequent effective signal values X.sub.eff (E.sub.m0,E.sub.n0), X.sub.eff (E.sub.m1,E.sub.n1), . . . , X.sub.eff (E.sub.mx,E.sub.nx) as estimated according to claim 10 is calculated by a) addition of the subsequent effective signal values and b) division by the number of added effective signal values, c) whereby a first relationship between the two points in time (E.sub.mi,E.sub.ni), between which each one of the involved subsequent effective signals is determined, is the same for all involved subsequent effective signal values, and d) whereby all of the points of time (5) (E.sub.m0,E.sub.m1, . . . , E.sub.mx) which are the earlier ones of the two points in time, between which one of the involved subsequent effective signals is determined, are in a second relationship towards each other.

12. The method according to claim 9, whereby the signal property to be estimated is a characteristic number for the amount of harmonic content H, which is evaluated by a) using a. the number and/or the distribution of points of time and/or b. the number and/or the differences of the reliable distances.

13. The method according to claim 1 whereby the signal x(t) is a voltage signal u(t) or a current signal i(t) or a power signal w(t).

14. The method of claim 6 further including a method for creating a distances record, comprising the steps of a) collecting points of time by reading the event record, b) sorting the points of time by ascending time of occurrence and numbering the points of time in order with the ascending time of occurrence, thereby assigning each point of time a number, if the numbers of the points of times or numbers of distances are stored in the distance record, c) determining distances D.sub.i between every two subsequent points of time E.sub.i, E.sub.i-1 by subtracting an earlier point of time from a later point of time, d) assigning each distance D.sub.i a number equal to the number of the later point of time, if the numbers of distances are stored in the distance record, e) storing the distances D.sub.i in the distances record, a. which is ordered and the distances are stored in the order of the occurrence of the earlier or the later of the two points of time between which the respective distance was evaluated or b. which comprises data-tuples and each data-tuple comprises one of the distances and a number of the earlier or the later of the two points of time between which the respective distance was evaluated or c. which is ordered and the distances are stored in the order of numbers of the distances, or d. which comprises data-tuples and each data-tuple comprises one of the distances and the number of the one of the distances which is comprised in the respective data-tuple.

15. The method of claim 14 further including a method for determining a sliding and weighted half period of a distance D.sub.X, comprising the steps of: a) determining distances D.sub.0, . . . D.sub.X, or reading the distances record b) setting a value of the sliding and weighted half period at distance D.sub.X (SWHP(D.sub.X)) to a maximum value of a list containing c. a fixed quantify A of distances D.sub.i divided by a divider Div, i. whereby 1. the distances D.sub.i have a number smaller or equal to the number of D.sub.X 2. or the distances D.sub.i are in a sorted distances record in time before or at the same record as the distance D.sub.X, 3. or the distances D.sub.i occur before or at the distance D.sub.X d. and the minimum distance, D.sub.min.

16. The method of claim 15 further including a method for determining if a distance is reliable, comprising the steps of: a) Determining distances D.sub.0, . . . D.sub.X or reading the distances record b) determining a distance Y between a first distance D.sub.X+1 with the number Nr(D.sub.X+1) and a second distance D.sub.Fi with the number Nr(D.sub.Fi), a. by addition of all distances with numbers between the number of the second distance D.sub.FiNr(D.sub.Fi) plus 1 and the number of the first distance D.sub.X+1 Nr(D.sub.X+1), b. whereby the second distance D.sub.Fi is a reliable distance and whereby the second distance D.sub.Fi has the largest number of all reliable distances which is smaller than the number of the first distance D.sub.X+1Nr(D.sub.X+1)  c) defining the first distance D.sub.X+1 to be reliable if a. the distance Y is larger or equal the sliding and weighted half period at the first distance D.sub.X+1, (SWHP(D.sub.X+1)), or b. if the distance Y is larger than a maximum distance D.sub.max.

17. The method of claim 16, whereby the signal property to be estimated is the effective signal X.sub.eff, which is estimated by a) integrating the square of the sensed signal x(t) between the later one of the two points of time used to evaluate a first distance D.sub.m and the later one of the two points of time used to evaluate a second distance D.sub.n, and by b) dividing the integrated signal through the distance between the first and the second distance.

18. The method of claim 17 further including a method for determining an improved effective signal X.sub.eff,imp, whereby a floating average of an even number of subsequent effective signal values X.sub.eff (E.sub.m0,E.sub.n0), X.sub.eff (E.sub.m1,E.sub.n1), . . . , X.sub.eff (E.sub.mx,E.sub.nx) as estimated according to claim 17 is calculated by a) addition of the subsequent effective signal values and b) division by the number of added effective signal values, c) whereby a first relationship between the two points in time (E.sub.mi,E.sub.ni), between which each one of the involved subsequent effective signals is determined, is the same for all involved subsequent effective signal values, and whereby all of the points of time (E.sub.m0,E.sub.m1, . . . , E.sub.mx) which are the earlier ones of the two points in time, between which one of the involved subsequent effective signals is determined, are in a second relationship towards each other.

19. The method of claim 16, whereby the signal property to be estimated is a characteristic number for the amount of harmonic content H, which is evaluated by a) using a. the number and/or the distribution of points of time and/or b. the number and/or the differences of the reliable distances.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The drawings used to explain the embodiments show:

(2) FIG. 1 A first embodiment with two thresholds and an alternating voltage as signal

(3) FIG. 2 A second embodiment with two thresholds and a rectified voltage as signal

(4) FIG. 3a A third embodiment with three thresholds and a rectified voltage as signal showing how reliable points of time are detected

(5) FIG. 3b The same situation as shown in FIG. 3a but instead of the tracker variable, the distances to the points of time and the reliable points of time are shown

(6) FIG. 4 The third embodiment with three thresholds and a rectified voltage as signal showing the integral of the voltage between two reliable points of time

(7) FIG. 5a A fourth embodiment with 4 thresholds and an alternating voltage as signal, showing the effect of a short black out on the detection of points of time and reliable points of time when the tracker variable is never blocked

(8) FIG. 5b The same situation as shown in FIG. 5 but with a tracker variable which can be blocked

(9) FIG. 6 A power supply which uses the method for estimating a signal property to protect itself from critical input voltages whereby the signal is an input voltage

(10) In the figures, the same components are given the same reference symbols.

PREFERRED EMBODIMENTS

(11) All figures illustrate the method for estimating a property of a signal on the example of a voltage signal. Instead of the voltage signal, also another signal as e.g. a current signal, a power signal, an energy signal, a charge signal, a magnetic field signal or any other signal which can be sensed in an electrical system can use the method and the device illustrated in the following. Signals which cannot have negative values, like e.g. a power or an energy signal, can replace a rectified voltage signal in the following examples. Signals which can have positive and negative values, like e.g. a current signal, can replace a voltage signal in the following examples. In both cases, the threshold values need to be replaced by values of the same unit at the signal: For example thresholds could be at 0.05, 0.1 and 0.2 Amperes instead of 50, 100 and 200 V in the case of a current signal or at 0.5, 1 and 2 Watt in the case of a power signal.

(12) FIG. 1 illustrates a first embodiment of a method for estimating a property of a voltage u(t). The waveform of the voltage 1 is in this case a sinusoidal curve with a fundamental frequency and added onto this curve is the 6th harmonic with an amplitude of 40% of the amplitude of the fundamental. The waveform 1 is plotted in a diagram where the left y-axis 9 shows the voltage, the right y-axis 8 the tracker variable 3 and the x-axis 7 the time. In addition to the waveform of the voltage 1 there are two thresholds 2.1 and 2.2 depicted. The first threshold 2.1 is at a positive voltage value which is larger than the one of the second threshold 2.2. The voltage value of the second threshold 2.2 is positive, too. The voltage waveform 1 crosses both thresholds 2.1 and 2.2 multiple times and in both, positive and negative directions. A crossing in a positive direction is a situation where the waveform 1 increases in the nearest neighbourhood around the crossed threshold, seen in the direction of the positive x-axis 7, i.e. time. A crossing in a negative direction is a situation where the waveform 1 decreases in the nearest neighbourhood around the crossed threshold, seen in the direction of the positive x-axis 7, i.e. time. There is a positive crossing of the first threshold 1.12, a positive crossing of the second threshold 1.22, a negative crossing of the first threshold 1.11 and a negative crossing of the second threshold 1.21.

(13) The positive crossing of the first threshold 1.12 causes a tacker variable 3 to be locked. This means that the tracker variable 3 cannot change its value until it is unlocked. A negative crossing of the first threshold 1.11 unlocks the tracker variable 3 and sets it to the value zero. A positive crossing of the second threshold 1.22 subtracts the value “two” from the value of the tracker variable 3 and a negative crossing of the second threshold 1.21 adds the value “one” to the tracker variable 3.

(14) The first threshold 2.1 is a first threshold in the sense of the invention. The second threshold 2.2 is a second threshold in the sense of the invention. The first direction is negative in this embodiment. The first mathematical operation is an addition of the first weighting value 6.21 which is “one” in this embodiment. The second mathematical operation can be either a subtraction of a second weighting value 6.22 being “two” or an addition of a second weighting value 6.22 of “minus two”.

(15) In FIG. 1, the value of the tracker variable 3 is plotted, too. The left y-axis 8 is the relevant axis for the tracker variable 3. Going from left to right, the voltage waveform 1 starts at a voltage value of 0. It increases and crosses the second threshold in positive direction 1.22. Thereby the tracker variable gets the value of −2. Shortly afterwards, the voltage waveform crosses the second threshold in a positive direction 1.12. This causes the tracker variable 3 to be locked, which is not visible in FIG. 1. A short time later, the voltage waveform 1 crosses the first threshold in a negative direction 1.11, which sets the tracker variable 3 to zero. A positive crossing of the first threshold shortly afterwards locks the tracker variable 3 again until the next negative crossing of the first threshold unlocks it and sets it again to zero. This is not visible as the tracker variable 3 was already zero before. Shortly afterwards the voltage waveform 1 crosses the second threshold in negative direction 1.21 which causes the tracker variable 3 to increase by “one” and to get the value 1. The trigger-condition 4 of this embodiment is the equality of the tracker variable 3 to the comparison value of 1. The comparison value is shown as dashed line in FIG. 1. The trigger-condition 4 is fulfilled at this negative crossing of the second threshold 1.21 and therefore a point of time 5 is detected. The point of time 5 is marked with a cross in FIG. 1. With the next positive crossing of the second threshold 2.2, the tracker variable becomes (−1). The following positive crossing of the first threshold 2.1 locks it again and the next negative crossing of the first threshold 2 sets it to zero.

(16) Although not shown, one can calculate the sequence of tracker variable values occurring when the waveform oscillates around the second threshold without crossing the first threshold: 1, −1, 0, −2, −1, −3 . . . In other words: with every new oscillation, the distance between the tracker variable and the comparison value becomes greater. Only the first negative crossing of the second threshold after a negative crossing of the first threshold fulfils the trigger-condition and leads to the detection of a point of time 5.

(17) An oscillation of the voltage waveform 1 around the first threshold locks and unlocks the tracker variable but keeps it at the constant value of 0.

(18) Assuming, that in this embodiment, jumps are not treated as a sequence of threshold crossings and that a jump is defined as a voltage change from one sample to the next which crosses two or more thresholds at once:

(19) A jump in the voltage from a value above the first threshold to a value below the second threshold keeps the tracker variable looked: Although the voltage may cross the second threshold, the tracker variable cannot change its values and therefore no points of times are detected until the first threshold is crossed in a negative direction again.

(20) A jump in the voltage from below the second threshold to a voltage above the first threshold keeps the tracker variable unlocked. This is no problem if there are no further jumps in the time until the first threshold is crossed again in a negative direction, setting the tracker variable to zero again.

(21) Assuming that, in this embodiment, jumps are defined as a voltage change between two subsequent voltage samples which crosses two or more thresholds at once and that jumps are treated as a series of threshold crossings:

(22) A jump in the voltage from a value above the first threshold to a value below the second threshold is therefore treated as a negative crossing of the first threshold immediately followed by a negative crossing of the second threshold: This is exactly the situation in which a point of time is detected.

(23) A jump in the voltage from a value below the second threshold to a value above the first threshold is therefore treated as a positive crossing of the second threshold immediately followed by a positive crossing of the first threshold: The tracker variable would be reduced by 2 and then locked. The trigger-condition cannot be fulfilled in this way and no point of time is detected.

(24) The difference between the two ways of treating jumps is therefore if a sudden voltage drop should cause a detection of a point of time or not. Treating a jump as a sequence of threshold crossings may be reasonable if there is a chance that the sampling rate is slow compared to the fundamental frequency. Having two thresholds with the same sign and a risk of black-outs or brown-outs, ignoring jumps may improve the estimate for properties of the undisturbed voltage signal.

(25) If the amplitude of possible harmonics is a priori difficult to estimate, increasing the distance between the two thresholds decreases the probability of false detections due to crossings of all thresholds by harmonics and improves therefore the robustness of the method. On the other hand, the thresholds should be chosen so that they are crossed during nominal operations. Choosing two thresholds with the same sign helps to detect measurements including black-outs.

(26) FIG. 2 shows a illustrates a second embodiment of a method for estimating a property of a voltage u(t).

(27) The setup of the diagram is the same as in FIG. 1. However the voltage 1 is rectified. Further, the thresholds, the weighting values and the trigger-condition are different.

(28) The first threshold 2.1 is at the positive voltage but lower than the second threshold 2.2. A positive crossing of the first threshold 6.11 sets the tracker variable 3 to 0. A positive crossing of the second threshold 6.21 adds 4 to the tracker variable 3 and a negative crossing of the second threshold 6.22 adds 0 to the tracker variable 3. The trigger-condition 4 is that the tracker variable 3 should be greater than 1 for the first time after being less or equal to 1.

(29) Again, an oscillation of the voltage 1 around the first threshold 2.1 keeps the tracker variable at zero.

(30) An oscillation of the voltage 1 around the second threshold 2.2 increases the tracker variable further and further. Therefore the trigger-condition keeps being fulfilled and no new points of time are detected.

(31) In this embodiment it is assumed that, a jump is a voltage change between two samples which crosses two or more thresholds at once.

(32) Ignoring voltage jumps prevents again the detection of points of time:

(33) A voltage jump from a voltage above the second threshold to a voltage below the first threshold does not change the tracker variable differently from what regular negative crossings of the second and the first threshold would do.

(34) A voltage jump from a voltage below the first threshold to a voltage above the second threshold causes that the tracker variable is not set back to zero and that nothing is added. However, as the trigger-condition was already fulfilled because it was not set back to zero, the missing addition of the first weighting value to the tracker variable does not change the trigger-condition. Therefore no new points of time are detected.

(35) If jumps are treated as a sequence of threshold crossings on the other hand, a point of time would be detected in a jump from below the first threshold 2.1 to above the second threshold 2.2.

(36) While the distance between two points of time is approximately the period of the fundamental in the embodiment of FIG. 1, it is approximately half of the period of the unrectified fundamental in the embodiment of FIG. 2.

(37) FIGS. 3a and b illustrate the idea and the method to determine reliable points of time 5.1. FIGS. 3a and b show the same voltage waveform and the same thresholds. FIG. 3a shows in addition the tracker variable 3. FIG. 3b shows instead the distances to the last point of time 5 and to the last reliable point of time 6 as well as the sliding and weighted half-period 15.

(38) The voltage waveform 1 in FIGS. 3a and b is a rectified fundamental with a 16.sup.th harmonic whereby the amplitude of the 16.sup.th harmonic is 40% of the fundamental's amplitude.

(39) There are three thresholds: the first threshold 2.1 is at the highest voltage of these thresholds and a positive crossing of it unlocks the tracker variable 3 and sets it to zero while a negative crossing does not change the tracker variable 3.

(40) The second threshold 2.2 has a voltage between the first 2.1 and the third threshold 2.3. Crossing it in positive direction adds 4 to the tracker variable 3 while crossing it in negative direction adds 0 to the tracker variable 3.

(41) The third threshold 2.3 is at the lowest voltage of all thresholds but still greater than zero. Crossing it in positive direction adds 0 to the tracker variable 3 while crossing it in negative direction adds (−1) to the tracker variable 3.

(42) The trigger-condition is that a point of time 5 is detected if the tracker variable 3 rises above or to 1 after being smaller before.

(43) As a result, a point of time 5 is detected if the second threshold 2.2 is crossed in positive direction after a positive crossing of the first threshold 2.1 and with less than four positive crossings of the third threshold 2.3 in between. This can be seen in FIG. 3a.

(44) Due to the shape of the voltage waveform 1 and the choice of the thresholds, multiple points of time 5 are detected around every crossing of 0-Volt of the fundamental of the voltage waveform 1.

(45) Looking at the distribution of points of time 5 along the time axis 7, there are therefore clusters around the crossings of 0-Volt separated by intervals of approximately one half-period of the fundamental.

(46) In order to estimate voltage properties, one assumes that the distance 13 between two subsequent points of time 5 is approximately one half-period of the fundamental in the case of a rectified voltage waveform 1. This assumption is in the case shown in FIG. 3 obviously wrong in most situations. In order to improve this, the points of time 5 are filtered to determine which ones of them are reliable 6.

(47) This filter consists of a register containing a fixed number of entries. If a point of time 5 is detected, all entries of the register are shifted by one and the distance 13 between the newly detected point of time and the previous one is written in the first entry of the register. This distance 13 is indicated by the peaks of the dashed line in FIG. 3b. Thereby, the value of the oldest entry before the detection of the new point of time 5 is lost. However, preferably this change in the register is only performed if the distance 13 between the newly detected point of time and the previous one is in an interval between a minimal and a maximal distance. This prevents, in the case shown in FIG. 3b, that the sliding and weighted half-period 15 increases after the peak in the distance 13 to the previous point of time at the end of the black-out. The distance value of 1000 at which first the reliable point of time 6 is detected and shortly afterwards distance 13 from the last point of time is reduced, is the maximum distance in this embodiment.

(48) The largest entry of this register is divided by a divider Div. The divider Div is here larger or equal to one. The result of this calculation is preferably compared to the minimal distance. The larger one of these two values is called the sliding and weighted half-period 15. This value is shown in FIG. 3b: The sliding and weighted half-period 15 is essentially constant. In the case shown here, Div is set to 2 which is why its value is so far off the maximum value of the distances 13 between to subsequent points in time 5.

(49) In parallel, the distance 14 of the newly detected point of time 5 to the last reliable point of time 6 is determined. In FIG. 3b this distance 14 is shown as the peaks of the continuous line except for the one of 1000. If the distance 14 to the last reliable point of time 6 is larger than the sliding and weighted half-period 15 and a point of time 5 is detected, this point of time 5 is assumed to be reliable 6.

(50) However there is in parallel also another way to determine a reliable point of time 6: The distance 14 of the momentary time to the last reliable point of time 6 is monitored, too. In FIG. 3b this distance 14 of the momentary time to the last reliable point of time 6 is shown as the continuous line. If this distance 14 is equal to a maximum distance, in the shown embodiment the maximum distance is 1000, a reliable point of time 6 is detected, even if there is no point of time 5. This reliable point of time 6 which is no point of time 5 is in FIGS. 3a and 3b marked with a circle without a cross in it.

(51) In FIGS. 3a and b, the register is always filled with short and long distances 13. The long distances 13 are slightly shorter than one half period of the fundamental. Therefore the sliding and weighted half period 15 is in this region also only slightly shorter than one half period of the fundamental divided by the divider Div, being 2 in this case. The points of time 5 around a single crossing of 0-Volt of the fundamental are in much smaller distances 13 to each other and are therefore no “reliable points of time”. However the first point of time 5 of the next “cluster” is about one half-period away from the last reliable point of time 6 and is therefore detected as being a reliable point of time 6.

(52) Following these first three crossings of 0-Volt of the fundamental, the voltage 1 drops to zero and stays there for a time greater than one half period. No points of time 5 are detected. However, the distance 14 of the momentary time to the last reliable point of time increases and reaches finally the maximum distance of 1000: A reliable point of time 6 is “detected” at a position where there is no point of time 5.

(53) After this zero-voltage region is one reliable point of time 5 found which does not fit in the regular pattern seen before the zero-voltage region. This is because the jumps at the beginning and the end of the zero-voltage region are treated as a sequence of threshold crossings. In this embodiment, a jump is a voltage change between two samples which crosses two or more thresholds at once. Further, the sliding and weighted half-period 15 is quite small compared to the half-period of the fundamental and by chance, the distance 14 between the point of time 5 and the last reliable point of time 6 is just long enough to fulfil the criteria for finding a reliable point of time 6.

(54) Due to this, in principle wrongly found, reliable point of time 6, the distance 14 to the next point of time 5 is also quite small. Nevertheless, it is correctly recognised in this case due to the small value of the sliding and weighted half-period 15.

(55) By reducing the divider Div the sliding and weighted half-period 15 increases and reduces the chances of wrongly detecting reliable points of time 6. However, if there is such a wrong detection, it may take more regular half-periods until reliable points of time 6 are correctly detected again.

(56) Also the maximum distance is a free parameter which can be chosen: If the value is set too small, a low fundamental frequency cannot be detected. On the other hand, if the value is to large, a longer black-out will not be detected as such but enter the register as long distance 13 between two points of time. This may increase the sliding and weighted half-period 15 to an unrealistic high value so that reliable points of time 6 are missed. However, this problem disappears as soon as the large value leaves the register. A suitable size of the register may be chosen to limit this problem.

(57) FIG. 4 shows another embodiment of a method for estimating a property of a voltage u(t).

(58) The setup of the diagram is the same as in FIG. 2. Again, the voltage 1 is rectified. The thresholds, the weighting values and the trigger-condition differ from FIGS. 1 and 2:

(59) The first threshold 2.1 is the one with the highest voltage value. The second threshold 2.2 has a lower voltage than the first but is still greater than the third threshold 2.3.

(60) Crossing the first threshold 2.1 in positive direction causes the tracker variable 3 to be initialized and set to zero. Crossing the first threshold 2.1 in a negative direction has no effect of the tracker variable 3. The first direction is therefore the positive direction in this embodiment.

(61) Crossing the second threshold 2.2 in a positive direction causes the first weighting value of 4 to be added to the tracker variable 3. Crossing the second threshold 2.2 in a negative direction causes the second weighting value of 0 to be added to the tracker variable 3.

(62) Crossing the third threshold 2.3 in a positive direction causes the third weighting value of 0 to be added to the tracker variable 3. Crossing the third threshold 2.3 in a negative direction causes the fourth weighting value of −1 to be added to the tracker variable 3.

(63) The trigger-condition 4 is fulfilled if the tracker variable 3 is greater or equal to one for the first time after being smaller than one.

(64) This choice of weighting factors and thresholds has the effect that many oscillations around the third and lowest threshold 2.3 can prevent tracker variable 3 from reaching the trigger-condition 4. There is no need to cross the third and lowest threshold 2.3 to reach the trigger-condition 4. Such a choice of weighting factors and thresholds may be useful in situations where many oscillations at low voltages are an indication of a problem or an otherwise special situation.

(65) The waveform shown in FIG. 4 does not have enough oscillations around the third and lowest threshold 2.3 in order to prevent the tracker variable 3 from reaching the trigger-condition. Therefore two point of time 5 are detected during the observed time. Both points of time 5 are identified to be reliable points of time 6. In order to estimate a voltage property like the average voltage, the voltage values 1 are integrated over the time between the two subsequent points of time 5. This is the integral 11. The lines 10.1 and 10.2 indicate the points of time 5.

(66) FIGS. 5a and 5b show two evaluations of the same voltage waveform 1 with the same thresholds, weighting values and trigger-condition. The two figures differ in the way jumps are treated: In FIG. 5a a jump is treated as a sequence of threshold crossings, in FIG. 5b a jump does not influence the tracker variable. In both cases, a jump is a voltage change between two samples which crosses two or more thresholds at once, in this embodiment.

(67) The voltage waveform 1 is a fundamental together with its sixth and sixteenth harmonic. The amplitudes of both harmonics are 40% of the fundamental's amplitude. Further there is a time during which the voltage drops to zero before continuing as before, simulating a black-out.

(68) In FIGS. 5a and 5b the waveform 1 is depicted by many points indicating the individual sampled voltage values. Thereby, the jump at the end of the black-out becomes clearly visible.

(69) The first threshold 2.1 is in this embodiment at −50 V and a positive crossing locks the tracker variable 3 while a negative crossing unlocks the tracker variable and sets it to zero.

(70) The second threshold 2.2 is in this embodiment at −250 V and both, a positive and a negative crossing add one to the tracker variable 3.

(71) The third threshold 2.3 is at 250 V and a positive crossing (2.32) multiplies the tracker variable 3 with (−1) (12.32) while a negative crossing (2.31) multiplies it with zero (12.31).

(72) A fourth threshold 2.4 located at 50 V and a positive crossing (2.42) unlocks the tracker variable (12.42) while a negative crossing (2.41) locks it (12.41).

(73) The trigger-condition is that the tracker variable is smaller than zero for the first time after being greater or equal to zero. Therefore a point of time 5 is detected if the second threshold 2.2 was crossed followed by a positive crossing of the third threshold 2.3 and without any negative crossing if the first threshold 2.1 in between.

(74) At the end of the black-out, the voltage jumps from zero, located between the first 2.1 and the fourth threshold 2.4 to below the second threshold 2.2. This is a jump in this embodiment, because two thresholds are crossed at once. The jumped-over thresholds are the first 2.1 and the second threshold 2.2 and the jump is in negative direction.

(75) In FIG. 5a, the jump is treated as a negative crossing of the first threshold 2.1, unlocking the tracker variable 3 and setting it to zero, followed by a negative crossing of the second threshold, adding one to the tracker variable 3. Therefore the value of the tracker variable 3 equals one after the jump.

(76) In FIG. 5b, the jump does not change the tracker variable 3. As the fourth threshold 2.4 was crossed in negative direction just before the blackout, the tracker variable 3 was locked and is still locked when the black-out ends. Due to the jump, there is no threshold crossing which unlocks the tracker variable 3 again and therefore its value equal 0 and it stays locked until the fourth threshold 2.4 is crossed in positive direction of until the first threshold 2.1 is crossed in a negative direction.

(77) This difference between FIGS. 5a and 5b results in a point of time 5 being detected in FIG. 5a, but being not detected in FIG. 5b. As a consequence, the distance 13 between two detected points of time 5 is in FIG. 5b so large, that a reliable point of time 6 is inserted by the method in between the two detected points of time 5.

(78) FIG. 6 shows an electronic device which uses the method for estimating a voltage property. A voltage u(t) 1 is delivered to the electronic device. In this case the electronic device is a power supply. The power supply consists of a conventional power supply 17, a measurement and evaluation device 18 and a correction device 19. Before the conventional power supply 17 the voltage u(t) is measured by a voltmeter 18.1 and the method for estimating a voltage property is executed in real time by subdevice 18.2. Both, the voltmeter 18.1 and the subdevice 18.2 are part of the measurement and evaluation device 18. The measurement and evaluation device 18, in particular its subdevice 18.2, contains either a suitable analogue circuitry or suitable digital devices like microprocessors or a mixture of both. The measurement and evaluation device 18 measures the voltage u(t) and executes the method for estimation of a voltage property. The resulting estimated voltage properties are delivered via a signal channel 20 to the correction device 19 which protects the conventional power supply 17 from input voltages with critical properties. The signal channel 20 can transmit the estimated voltage property with a wire or wireless.

(79) In other embodiments, only an indicator signal is transmitted to the correction device 19 or the conventional power supply 17 itself. The indicator signal may indicate that one or more voltage properties are critical. It is also possible that one or more of the voltage properties are transmitted to the conventional power supply device 17 itself. The correction device 19 may be included in the measurement and evaluation device 18 and one or both may be included in the conventional power supply 17. It is also possible that the voltmeter 18.1 and the subdevice 18.2 are integrated in a single unit or that they are separated and integrated in different devices or being single units. Further, the voltage property estimate or an indicator signal indicating a critical value of one of more voltage properties may be transmitted to a storage device or to a supervisor.

(80) An electronic device which uses the method for estimating a current property is similar to the one shown in FIG. 6, but the voltmeter 18.1 is replaced by an ampere meter which is placed in the current path of interest. An electronic device which uses the method for estimating a power property may comprise both, a voltmeter 18.1 and an ampere meter and multiply the two values to evaluate a power signal. In another embodiment, it comprises a power meter instead of the voltmeter 18.1. One way of producing an energy signal is to integrating a power signal over a fixed duration. Electronic devices with use the method for estimating other signal properties replace the voltmeter 18.1 with a measurement device for the desired signal: e.g. a hall-sensor for a magnetic field signal, a thermocouple for a temperature signal. Of course, it is also possible to use sensors based on other measurement principles, too.

(81) In summary, it is to be noted that the method can be used for alternating voltages as well as for rectified voltages. The choice of the amount of thresholds, their values, the weighting factors and the mathematical operations shown in the different embodiments are only examples: They can be adapted to the needs of the application in which the method should be used.

LIST OF REFERENCE SYMBOLS

(82) TABLE-US-00001 1 Sensed Voltage u(t) 1.11 Crossing of first threshold in first direction 1.21 Crossing of second threshold in first direction 1.12 Crossing of first threshold in second direction 1.22 Crossing of second threshold in second direction 1.31 Crossing of third threshold in first direction 1.32 Crossing of third threshold in second direction 1.41 Crossing of fourth threshold in first direction 1.42 Crossing of fourth threshold in second direction 2.1 First threshold 2.2 Second threshold 2.3 Third threshold 2.4 Fourth threshold 3 Tracker variable 4 Trigger condition 5 Point of time 6 Reliable Point of time 12.11 Set tracker variable to starting value (bei 1.sup.st threshold in 1.sup.st direction) [FIG. 1] 12.12 Locking of tracker variable (bei 1.sup.st threshold in 2.sup.st direction) 12.21 First weighting value (bei 2.sup.nd threshold in 1.sup.st direction) [FIG. 1, 2] 12.22 Second weighting value (bei 2.sup.nd threshold in 2.sup.st direction)[FIG. 1] 12.31 Modify fourth weighting value (bei 3.sup.rd threshold in 1.sup.st direction) [FIG. 5a] 12.32 Modify by fifth weighting value (bei 3.sup.rd threshold in 2.sup.st direction) [FIG. 5a] 12.41 fourthprotecting tracker variable (bei 4.sup.rd threshold in 1.sup.st direction) [FIG. 5a] 12.42 fourthAllowing changes of tracker variable (bei 4.sup.rd threshold in 2.sup.st direction) [FIG. 5a] 7 Time axis 8 Axis for tracker variable and trigger condition 9 Axis for sensed signal u(t) curve 10.1, Orientation lines 10.2 11 Integral over signal between two subsequent reliable points of time 13 Distance from the last point of time 14 Distance from the last reliable point of time 15 Sliding and weighted half-period (SWHP) 16 Axis for distances and SWHP 17 Conventional power supply 18 Measurement and evaluation device; 18.1 Voltmeter 18.2 subdevice 19 Correction device 20 Signal channel