CN115629230A

CN115629230A - Power grid current use measurement method

Info

Publication number: CN115629230A
Application number: CN202211183214.8A
Authority: CN
Inventors: 刘晓羽
Original assignee: Individual
Current assignee: Individual
Priority date: 2022-09-27
Filing date: 2022-09-27
Publication date: 2023-01-20

Abstract

The invention provides a power grid current use measurement method, which verifies the accuracy of an integral algorithm under different conditions by establishing a simulation model of a complex interference magnetic field environment; the system structure and the hardware circuit of the current sensor are designed; the measuring device prototype is manufactured, an experiment platform is built, performance test is carried out, and the measuring device prototype has high measuring precision and external magnetic field interference resistance, and comprises the following steps: s1: current measurement based on sensor array and numerical integration; the method comprises the measurement principle and a Newton-Cotes product solving formula; s2: a first simulation experiment and result analysis, including error analysis of a simulation model, an experiment result and a product calculation algorithm; s3: performing a second simulation experiment and result analysis, including hardware circuit design, zero point elimination of the sensor and software program design; s4: testing a system experiment; TMR current sensor experiment platform, experimental result and analysis.

Description

Power grid current use measurement method

Technical Field

The invention relates to the technical field of current measurement research, in particular to a power grid current use measurement method.

Background

In China, a neutral point is commonly grounded through an arc suppression coil. The existing national standard (GBJ 63-83) of China stipulates a power distribution network with 60kV or below, and the grounding residual current of a fault point cannot exceed 10A. The accurate and rapid measurement of the capacitance current of the power distribution network is the basis for determining the capacity of the arc suppression coil and the tuning control of the arc suppression coil. There are two main methods of signal injection: one is that three current signals with different frequencies are injected from a voltage transformer, the voltage amplitude of the triangular side of an opening is measured, and a simultaneous equation is used for solving the capacitance current of the power distribution network; and the other method is to inject two constant current signals with different frequencies into the opening triangle side of the voltage transformer of the power distribution network, measure the return voltage signal and calculate the capacitance current of the power distribution network. The two injection current methods are difficult to select the injection signal frequency, the influence of the damping resistance formed by connecting arc suppression coils in series is ignored, the measurement error is large, and a power grid current measurement method is urgently needed at home and abroad;

use of a magnetic sensor in a non-contact current measurement technique. However, the magnetic saturation problem caused by the use of the magnetic gathering structure limits the application of the magnetic gathering structure in a large-current measurement scene, and also causes the increase of the volume and the weight of the device. The current solving algorithm adopted by the sensor array is generally only used for analyzing the specific situation of a single parallel interference wire and has no universality.

Therefore, how to avoid using a magnetic gathering or shielding structure and weaken the influence of the background magnetic field on the measurement precision under different interference conditions is the key of the magnetic sensing type current measurement research.

The magnetic saturation problem brought by the magnetic gathering structure in the prior art limits the application of the magnetic gathering structure in a large-current measurement scene, and increases the volume and weight of the device. The current solving algorithm adopted by the sensor array is generally only used for analyzing the specific condition of a single parallel interference wire and is not universal.

Disclosure of Invention

Aiming at the defects in the prior art, the accuracy of the product-solving algorithm under different conditions is verified by establishing a simulation model of a complex interference magnetic field environment; the system structure and the hardware circuit of the current sensor are designed; the measuring device prototype is manufactured, an experimental platform is built, performance test is carried out, the measuring device prototype has high measuring precision and high capacity of resisting external magnetic field interference, the maximum relative errors in a 0-16A current measuring range are respectively 0.31% of direct current and 1.00% of power frequency alternating current under the condition of no magnetic gathering ring, and the high precision can be still maintained when the position of a current lead and the posture of a sensor are changed; a grid current usage measurement method is provided.

In order to achieve the above purpose, the invention provides:

a grid current usage measurement method comprising the steps of:

s1: current measurement based on sensor array and numerical integration; the method comprises the measurement principle and a Newton-Cotes product solving formula;

s2: a first simulation experiment and result analysis, including error analysis of a simulation model, an experiment result and a product calculation algorithm;

s3: performing a second simulation experiment and result analysis, including hardware circuit design, sensor zero elimination and software program design;

s4: testing a system experiment; TMR current sensor experiment platform, experiment result and analysis;

s5: and (6) summarizing the experiment.

Preferably, in S1, the measurement principle includes the following steps:

firstly, a TMR element utilizes the tunnel magnetoresistance effect of a magnetic multilayer film material, is essentially a magnetic resistor, and encapsulates four magnetic resistors in a chip in a bridge structure to form a TMR magnetic sensor with output voltage changing along with an external magnetic field;

secondly, on the premise of determining the transfer relation of current-magnetic field-output voltage, indirectly calculating the current value by measuring the output of the sensor;

thirdly, a method for measuring current by using a TMR sensor array is adopted, a plurality of magnetic sensors are uniformly distributed around a lead through which current I flows, and the method is characterized in that the method comprises the following steps:

∮ _l B·dl＝μ ₀ I (1)

b is magnetic induction intensity, which is formed by superposing a magnetic field generated by the measured current and other interference magnetic fields; and l is a circular closed integration path containing a magnetoresistive sensor. Since the current generating the disturbing magnetic field is outside the integration path, the current does not influence the integration result, i.e. the integral contribution of the disturbing magnetic field on l is 0, so that the influence of the disturbing magnetic field can be eliminated by accurately calculating the integral. Taking the circumference as an integral path, decomposing the magnetic induction B of any point P on the circle according to the tangent B1 of the circumference, the normal B2 and the direction B3 vertical to the plane formed by B1 and B2, and only considering the integral of the tangential direction component B1 on l because B2 and B3 are both vertical to the vector infinitesimal dl. A plurality of single-axis TMR sensors are arranged in a mode that the sensitive axis is located on the circumferential tangent line to measure B1, and the output signals of the sensors are integrated to obtain the current I:

μ ₀ I＝∮ _l (B ₁ +B ₂ +B ₃ )·dl＝∮ _l B ₁ ·dl (2)

it can be seen from the derivation process that the key to accurately measuring the current is to accurately measure the magnetic field and ensure the accuracy of the integral calculation.

Preferably, in S1, for an integrand f (x) in the interval [ a, b ], the Newton-Cotes product formula performs Lagrange interpolation function Ln (x) approximation f (x) on the known function values f (xk) at several different nodes xk in the interval to obtain an interpolation type product formula:

where ω k is a weight coefficient determined by Ln (x). If equation (3) holds for any polynomial of degree not exceeding p, but there is a partial polynomial of degree p +1 that does not hold, it is said to have p-degree algebraic precision. When the interpolation nodes are equidistant, the value of ω k depends only on m and k, called Cotes coefficients C (m, k), and equation (3) can be expressed as Newton-Cotes

And (4) a formula. If m =1, a trapezoidal product-solving formula which needs 2 integral nodes and has 1-order algebraic precision is obtained:

if m =2, a Simpson product formula is obtained which requires 3 integration nodes and has 3-order algebraic precision:

equally dividing [ a, b ] into n small intervals [ xi, xi +1], wherein i =0,1, ·, n-1, the length h = (b-a)/n. And (3) respectively using the formula (4) and the formula (5) to obtain a complex trapezoidal formula and a complex Simpson formula:

wherein oddi and eveni denote that i is odd and non-zero even, respectively. When calculating the current, the plurality of TMR sensors form a mutually different node xk, the magnetic induction measurement value is f (xk), the magnetic induction B (x) is a function of position, and f (x) is B (x) in equations (6) and (7).

Preferably, in S2, a simulation model and an experimental result are established, in order to verify the accuracy of the numerical product algorithm and study the application range of the numerical product algorithm, the simulation model is established by using ANSYSMaxwell and Matlab, and a scene with a complex interference magnetic field is simulated.

As shown in fig. 2, a plurality of uniaxial TMR sensors are uniformly arranged around a current I to be measured with r as a radius to form a circular array l, the sensitive axes are all located in the circumferential tangential direction, a plurality of currents exist outside the array, and magnetic fields generated by the currents jointly act on the sensors to cause serious interference on measurement;

under the condition that the current to be measured is not changed, the integral true value is constant, but is limited by the practical condition, and only a limited number of nodes can be used for approximate solution;

the number of sensors and the selection of the integration formula will affect the integration accuracy, and the variation of the relative positions of the sensors and the conductors will also cause the output of the sensors to vary, thereby changing the calculated value of the integration. Therefore, it is necessary to study the influence of various factors on the measurement accuracy. The number n of the sensors, the radius r of the annular array and the eccentric distance d of the lead are respectively changed, the influence of the sensors on the relative error of the quadrature algorithm is tested, and the calculation steps are as follows:

(1) Setting the current to be measured as 600A, and generating 5 interference currents with random sizes (0-1000A) and random positions within a space range of +/-1.5 m by taking the current to be measured as a center; (2) Calculating a space magnetic field excited by all currents together according to the Bio-Safahr law to obtain a magnetic induction intensity tangent component B1 of the position where the sensor is located, substituting the magnetic induction intensity tangent component B1 into a formula (6) or a formula (7) to calculate an integral, calculating the current to be measured according to a formula (2), and obtaining a relative error between a current calculation value and a true value 600A;

(3) Changing n (2-12), r (3 cm-30 cm) and d (0-60 mm) respectively, and researching the change condition of relative errors;

taking a direct current experiment as an example, a complex trapezoidal formula and a complex Simpson formula are respectively applied to perform calculation, and a simulation result is shown in FIG. 3.

Preferably, in S2, the error analysis of the product algorithm: the basic idea of numerical value integration is that an constructor Ln (x) replaces an integrand B (x), and the definite integral of Ln (x) is used to approximate the integral true value, and FIG. 4 shows the geometrical meaning of two integration formulas. The formula (6) approximates an integral value by a broken line and an area (namely, a trapezoidal area) enclosed by an x axis in each small interval; the area of the parabola is enclosed with the x axis in every two cells to approximate an integral value in the formula (7);

the remainder is the error source of the product formula, and the geometric meaning refers to the algebraic sum of the areas enclosed by B (x) and Ln (x). As can be seen from fig. 4, the size of the remainder is closely related to the variation of B (x). When B (x) changes violently, ln (x) in a peak interval deviates from B (x) seriously, ln (x) of two formulas cannot be well fitted with a real curve, the B (x) curve between adjacent TMR sensors is generally closer to an S shape and is between a straight line and a parabola, and the errors of two kinds of integration algorithms are increased reversely;

increasing n will increase the approximation degree of Ln (x) and B (x) no matter what kind of product formula is adopted, when n tends to infinity, the both approach infinitely, the integral value also tends to true value; the radius r of the annular array has important influence on the capacity of resisting the interference of an external magnetic field, the smaller r is, the smaller the influence of the external magnetic field is, and the more accurate the integral calculation result is because B (x) changes when the radius is smaller;

the integral calculation algorithm tends to be gentle, the integral calculation algorithm has better performance, and on the contrary, the larger r is, the larger the influence of the interference magnetic field is, the larger B (x) oscillation is, and the integral error is increased; when the wire deviates from the center, the magnetic field at the position of the sensor changes along with the wire, and the output of the sensor close to the current wire is increased sharply due to excessive deviation, so that the integration precision is reduced due to the existence of a peak in the B (x) curve.

Preferably, in S3, a hardware circuit design is performed, and fig. 5 is a hardware circuit block diagram of the measurement system, which includes a power supply module, a TMR sensor array, a signal conditioning circuit, a microcontroller, and a wireless communication module.

A TMR2001 single-sensitive-axis magnetic sensor is used, which comprises 4 non-shielding high-sensitivity TMR sensing elements and adopts a push-pull Wheatstone full-bridge structure, and when an external magnetic field is parallel to the sensitive direction of the sensor, the bridge provides differential voltage output. The linear magnetic field interval is +/-5 Oe, the magnetic sensitivity is 8mV/V/Oe, the saturation magnetic field is-25 Oe and +40Oe, and a typical output curve of the sensor when 1V is powered is shown in figure 6.

In order to improve the measurement accuracy of the system, TMR2001 and ADC use a 3V reference voltage provided by a high-accuracy voltage source chip ADR4530, and the maximum output of TMR2001 is +/-120 mV in a linear region. Since the sensor is a differential output and the ADC conversion range is not fully utilized, signal conditioning is subsequently required. The voltage signal is amplified by the instrument amplifier AD8421, and a 1.5V forward bias is provided for a reference voltage pin of the AD8421 to raise an output voltage in order to realize negative voltage conversion. By combining the effective range of ADC, the optimal gain (1 +9.9k omega/RG) of AD8421 is 12.5 times, the gain resistance RG is 860.8 omega, 1k omega precision resistance with 0.1% precision is selected, and the actual gain is 10.9 times. The sensor array is designed into a circular ring PCB with the inner diameter of 60mm and the outer diameter of 100mm, 8 TMR2001 chips are uniformly arranged on the circumference with the radius of 35mm in a mode that a sensitive shaft is tangent to the circumference, and the position accuracy of an element is guaranteed by utilizing screen printing positioning and the surface tension of molten soldering tin during welding.

Preferably, in S3, the zero point of the sensor is eliminated, and another premise of accurately measuring the current is that the sensor can accurately reflect the magnetic field strength. However, due to the difference in manufacturing processes, the TMR element also outputs a voltage called an offset voltage V0 when the magnetic field strength H =0, and table 1 shows some performance parameters of the TMR2001 when power is supplied at 1V.

It can be seen that the offset voltage V0 is large (-30 mV-30 mV), and the effect of this voltage on the system zero is not negligible.

The TMR sensor output voltage Vout under the effect of magnetic induction B can be expressed as:

V _out ＝SEN·V _cc ·B+V ₀ (8)

wherein Vcc is the sensor supply voltage; SEN and V0 are the sensitivity and offset voltage of the sensor, respectively, then:

the actual value of the magnetic induction B can be calculated according to equation (9), and table 1 already gives typical values of SEN, but V0 has only a rough range and cannot be treated as a fixed value, so Vout and V0 need to be measured. Vout can be measured by an ADC, and V0 needs to be measured independently for each TMR chip under the condition of magnetic shielding, so that the implementation difficulty is certain when the number of sensors is large. Another measurement method was derived based on the principle of the array method. As can be seen from equations (6) and (7), the current I is a weighted sum of the magnetic induction intensities Bi:

substituting formula (9) for formula (10) yields:

in the formula: subscript i represents the ith TMR sensor, when no current exists in the loop, the above formula should be equal to zero, then the ADC is used for collecting the sensor output voltage Vout in any environment, the offset voltage integral of TMR can be reversely deduced, and when the current is measured, the correction is carried out according to the formula (11), namely, the zero point of the current sensor can be eliminated. According to the scheme, the influence of the offset voltage of the TMR element is eliminated by a simple method, repeated and complicated measurement work is avoided, and zero compensation of the current sensor can be realized only by simple calculation.

Preferably, in S3, after the software programming and the system assembly are completed, the voltage acquisition, the data processing, the zeroing and other operations are implemented by an internal program of the microcontroller, and the specific process is as follows:

(1) Zero setting operation is required before current measurement, a zero setting instruction is sent by an upper computer when no current exists in a loop, the microcontroller starts the ADC to measure the output voltage of 8 paths of AD8421 after receiving the zero setting instruction, and offset voltage integral is calculated according to a formula (11);

(2) A 16-bit hardware Timer _ A in the microcontroller sets that timing interruption is generated every 1ms, an ADC is started to execute sequence channel conversion, 8 paths of voltages are collected and weighted and summed, correction is carried out according to a formula (11), and a current instantaneous value is calculated;

(3) Every 1s, averaging all the instantaneous values of the current to calculate the direct current component, calculating the root mean square to calculate the alternating current component, and sending the current data to the upper computer through the wireless module, wherein the program flow of the microprocessor is shown in fig. 7.

Preferably, in S4, a TMR current sensor test platform is built in a laboratory environment, as shown in fig. 8, including an ac/dc current source, a TMR current sensor, an oscilloscope (tack TDS 2012B), a clamp ammeter (Flukei 30S), a long straight conductor (a copper bar 90cm in length and 30mm in diameter), and a plurality of leads (2.5 mm2 single-core copper wires).

When the system is tested, a standard source generates standard direct current or power frequency current to be tested, a lead is connected with a power supply and a copper bar conductor, the conductor penetrates through the TMR sensor array, a clamp meter Flukei30s observes the current to be tested, an oscilloscope observes the output waveform of the clamp meter and the output signal waveform of a sensor conditioned by an instrumentation amplifier, a notebook computer is used as an upper computer to send a control instruction, and current measurement data is received and displayed, and fig. 9 is a test platform real object diagram.

Preferably, in S4, experimental results and analysis

When no current exists in the loop, under the combined action of the offset voltage of the sensor and the background magnetic field of a laboratory, the output of 8 instrumentation amplifiers is superposed with an interference amount on the basis of 1.5V offset, and the output voltage value is distributed in the range of 1.164V-1.701V. The influence of the background magnetic field is eliminated after the numerical integration, but the existence of offset voltage enables the current measurement value at the moment to still show-3.11A, and the necessity of the zero setting operation of the system is proved again. The measured value of the TMR current sensor after zero setting is 0.03A, the position and the posture of the current sensor are randomly adjusted, and the measured value of the current fluctuates within a small range of-0.02A-0.06A, which shows that the system has good shielding effect on a background magnetic field in a laboratory environment, and the zero setting is still effective even if the position of the system is changed. When power frequency currents of 6A and 12A are given, the output voltage waveform of one AD8421 is shown in fig. 10, so that the conditioned TMR output signal is raised by 1.62V, is in a sine waveform without obvious distortion, has the same frequency and phase with the measured current, and has the amplitude which is increased in equal proportion to the measured current.

During testing, the conductor is perpendicular to the plane of the TMR sensor array and is positioned at the center of the circumference, so that the best measuring effect can be achieved. And taking the complex trapezoidal formula as a product algorithm, and respectively measuring the 0-16A direct current and the power frequency current after the zero setting operation is finished.

The maximum relative error of the prototype in the measurement range of 0-16A is 0.31% of direct current and 1.00% of alternating current, and least square linear fitting is respectively carried out on the direct current measurement data and the alternating current measurement data, and a fitting curve is shown in figure 11. The fitting result shows that the correlation coefficients of direct current linear fitting and alternating current linear fitting are 99.9994% and 99.9976%, the nonlinear errors are 0.215% and 0.462%, the measurement system presents very high linearity, and the slope and the Root Mean Square Error (RMSE) of the linear fitting curve also show that the approximation degree of the current measurement value and the true value is high.

In order to test the measurement error of the system under the non-ideal test condition, the conductor is respectively processed by eccentricity of 15mm and inclination of 45 degrees, and the influence of current position and attitude adjustment on the measurement precision is researched; applying a 6A close-distance parallel conductor current outside the circular ring to test the anti-magnetic field interference capability of the system; in addition, a hard straight copper bar conductor is replaced by a soft bent wire, and the influence of the shape of the wire on the measurement result of the device is researched. In the experimental process, the product calculation algorithm is changed, the test result shown in the effective graph 12 of two product calculation formulas of a complex trapezoid and a complex Simpson is compared with the error change trend of a simulation experiment, and the error change trend is reduced along with the reduction of the interference magnetic field and is increased along with the increase of the eccentric distance. The effect of the complex trapezoidal formula is slightly better than that of the complex Simpson formula. When small current is measured, a magnetic field generated by the current to be measured is weaker than a background interference magnetic field, the TMR sensor is mainly influenced by the interference magnetic field, and magnetic field images along an integral path are often changed violently, so that the error is slightly larger. When the current is increased, the magnetic field of the current to be measured gradually occupies a dominant position, and the measurement error is reduced.

Fig. 12 (a), (b) show that the current eccentricity or inclination to be measured causes slight interference on measurement, but the device is not sensitive to the small-amplitude position and posture adjustment of the lead, and the measurement error is still small even in the case of eccentricity of 15mm (42.9% of the circumferential radius of the sensitive shaft) or inclination of 45 degrees. Fig. 12 (c) shows that the close-range strong interference current brings a certain error to the measurement result, especially when the current to be measured is small, the error is large, but as the distance between the two increases, or when the current to be measured increases to be equivalent to the interference current, the error is recovered to be small; FIG. 12 (d) shows that the change in the shape and thickness of the wire has little influence on the measurement results. The device has no special requirement on the shape of the lead.

From the linear range (± 5 Oe) of TMR2001 and the sensor array size (radius r =35 mm), the linear measurement range of the current sensor can be calculated to be 87.5A. The current above 16A is not tested herein, limited by the output current capability of standard current sources. But it can be determined that the larger the measured current is, the less the external interference is, the signal-to-noise ratio is improved, and the measurement accuracy should be higher

In order to achieve a good measurement effect, the peak of the B (x) curve should be avoided, and the sensor should be prevented from falling into a nonlinear magnetic field range, so that the device is required to restrict the position and the posture of the conducting wire, and the current is not allowed to be tightly attached to the TMR chip. In practical application, the sensor array can be controlled within a reasonable range by adopting measures of reducing the inner diameter, increasing the outer diameter, increasing the thickness of a device shell and the like.

Compared with the prior art, the invention has the advantages that:

the invention discloses a non-contact current sensor design using a circular TMR sensor array, and points out that the key for realizing high-precision measurement is to accurately measure a magnetic field and improve numerical value integration precision.

The method establishes a simulation model of a complex interference magnetic field environment, researches the influence of various factors on the quadrature precision, analyzes error sources and provides specific measures for reducing errors; based on the working principle of an array method, a new zero compensation method is provided, and repeated and tedious measurement work is avoided;

the invention manufactures a current sensor prototype and sets up an experiment platform for testing; the result shows that the device has good shielding effect on the background interference magnetic field, has higher linearity and measurement accuracy, and has the nonlinear errors of 0.215 percent of direct current and 0.462 percent of power frequency alternating current in the measurement range of 0-16A, and the maximum relative errors of 0.31 percent and 1.00 percent respectively;

the device developed by the invention is insensitive to the position of a current conductor and the posture of a sensor, the small-amplitude eccentric inclination of a lead or the change of the shape and thickness of the lead can not cause obvious influence on the measurement result, and certain measurement precision can be achieved under the environment of a strong interference magnetic field of an external current.

Drawings

FIG. 1 is a schematic diagram of TMR sensor array current measuring principle of the present design;

FIG. 2 is a simulation model of the present design;

FIG. 3 is a schematic diagram showing the influence of the main factors on the measurement error in the present design;

FIG. 4 is a geometric meaning diagram of the product formula;

FIG. 5 is a block diagram of a measurement system circuit;

FIG. 6 is a schematic diagram of an exemplary output curve of a TMR2001 sensor of the present design;

FIG. 7 is a flowchart of the system program of the present design;

FIG. 8 is a schematic diagram of an experimental test platform in the present design;

FIG. 9 is a diagram of a sample experimental test platform in the present design;

FIG. 10 is a waveform of the output of the instrumentation amplifier in the present design;

FIG. 11 is a plot of a linear fit of measured data in the present design;

FIG. 12 is a graph of measurement error for non-optimal test cases in the present design.

Detailed Description

Embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

Referring to fig. 1-12, a grid current usage measurement method includes the steps of:

s5: and (6) summarizing the experiment.

Preferably, in S1, the measurement principle includes the following steps:

firstly, the TMR element utilizes the tunnel magnetoresistance effect of a magnetic multilayer film material, is essentially a magneto-resistor, and encapsulates four magneto-resistors in a chip by a bridge structure to form a TMR magnetic sensor with the output voltage changing along with an external magnetic field;

secondly, on the premise of determining the transfer relation of current-magnetic field-output voltage, the current value is indirectly calculated by measuring the output of the sensor;

∮ _l B·dl＝μ ₀ I (1)

b is magnetic induction intensity, which is formed by superposing a magnetic field generated by the measured current and other interference magnetic fields; and l is a circular closed integration path including a magnetoresistive sensor. Since the current generating the disturbing magnetic field is outside the integration path, the current does not influence the integration result, i.e. the integral contribution of the disturbing magnetic field on l is 0, so that the influence of the disturbing magnetic field can be eliminated by accurately calculating the integral. Taking the circumference as an integral path, decomposing the magnetic induction B of any point P on the circle according to the tangent B1 of the circumference, the normal B2 and the direction B3 vertical to the plane formed by B1 and B2, and only considering the integral of the tangential direction component B1 on l because B2 and B3 are both vertical to the vector infinitesimal dl. A plurality of single-axis TMR sensors are arranged in a mode that the sensitive axis is located on the circumferential tangent line to measure B1, and the output signals of the sensors are integrated to obtain current I:

μ ₀ I＝∮ _l (B ₁ +B ₂ +B ₃ )·dl＝∮ _l B ₁ ·dl (2)

it can be seen from the derivation process that the key to accurately measuring the current is to accurately measure the magnetic field and ensure the accuracy of the integral calculation. In scientific calculation, when an integrand is given by a series of measuring point data, definite integral cannot be accurately obtained by utilizing a Newton-Leibniz formula, and only approximate calculation can be carried out by a numerical value quadrature formula, wherein the Newton-Cotes type and the Gauss type are most widely applied. In the former, an interpolation polynomial is constructed by equidistant nodes to replace an integrand, the higher the polynomial degree is, the higher the algebraic precision of the integrand is, but the Runge phenomenon can be generated at the same time; the latter has the advantages of stability and high precision, but the product node contains irrational numbers, which brings certain difficulty to the positioning of the TMR chip on the PCB. And comprehensively considering, selecting a complex low-order interpolation Newton-Cotes formula.

In this embodiment, in S1, for an integrand f (x) in an interval [ a, b ], a Newton-Cotes product formula obtains an interpolation-type product formula by performing Lagrange interpolation function Ln (x) approximation f (x) on a function f (xk) at a plurality of different nodes xk in the known interval:

In this embodiment, in S2, a simulation model and an experimental result are established, in order to verify the accuracy of the numerical value product algorithm, the application range of the simulation model is studied, and a scene with a complex interference magnetic field is simulated by building the simulation model using ANSYSMaxwell and Matlab.

(1) Setting the current to be measured as 600A, and generating 5 interference currents with random sizes (0-1000A) and random positions within a space range of +/-1.5 m by taking the current to be measured as a center; (2) Calculating a space magnetic field excited by all currents together according to the Biao-Safahr law to obtain a magnetic induction intensity tangent component B1 of the position of the sensor, substituting the magnetic induction intensity tangent component B1 into a formula (6) or a formula (7) to calculate an integral, and calculating the current to be measured according to a formula (2) to obtain a relative error between a current calculation value and a true value 600A;

taking a dc experiment as an example, a complex trapezoidal formula and a complex Simpson formula are respectively applied to perform calculation, and a simulation result is shown in fig. 3.

In this embodiment, the error analysis of the product-finding algorithm: the basic idea of numerical value integration is that an constructor Ln (x) replaces an integrand B (x), and the definite integral of Ln (x) is used to approximate the integral true value, and FIG. 4 shows the geometrical meaning of two integration formulas. The formula (6) approximates an integral value by a broken line and an area (namely, a trapezoidal area) enclosed by an x axis in each small interval; in the formula (7), the area of each two small intervals is approximate to an integral value by a parabola and an x axis;

no matter what kind of product formula is adopted, increasing n will improve the approximation degree of Ln (x) and B (x), when n approaches infinity, the two approach infinitely, the integral value also approaches to the true value; the radius r of the annular array has important influence on the capacity of resisting the interference of an external magnetic field, the smaller r is, the smaller the influence of the external magnetic field is, and the more accurate the integral calculation result is because B (x) changes when the radius is smaller;

In this embodiment, in S3, a hardware circuit design is performed, and fig. 5 is a hardware circuit block diagram of the measurement system, which includes a power supply module, a TMR sensor array, a signal conditioning circuit, a microcontroller, and a wireless communication module. In consideration of power supply convenience, a microUSB interface is reserved, a 5V mobile phone charger is matched to supply power to the whole system, and multilevel voltage generated by conversion of a power supply module meets the requirement of electric energy supply of each element. After being conditioned by an instrument amplifier, the output signal of the sensor is sent to an ADC (analog to digital converter) for sampling, is calculated and processed by a microcontroller msp430F149, and finally is communicated with an upper computer through a wireless transceiver, wherein the upper computer is responsible for sending a control instruction, receiving and displaying current data

A TMR2001 single-sensitive-axis magnetic sensor is used, which comprises 4 non-shielding high-sensitivity TMR sensing elements and adopts a push-pull Wheatstone full-bridge structure, and when an external magnetic field is parallel to the sensitive direction of the sensor, the bridge provides differential voltage output. The linear magnetic field interval is +/-5 Oe, the magnetic sensitivity is 8mV/V/Oe, the saturation magnetic field is-25 Oe and +40Oe, and a typical output curve of the sensor when the power is supplied at 1V is shown in figure 6.

In order to improve the measurement accuracy of the system, TMR2001 and ADC use a 3V reference voltage provided by a high-accuracy voltage source chip ADR4530, and the maximum output of TMR2001 is +/-120 mV in a linear region. Since the sensor is a differential output and the ADC conversion range is not fully utilized, signal conditioning is subsequently required. The voltage signal is amplified by the instrument amplifier AD8421, and a 1.5V forward bias is provided for a reference voltage pin of the AD8421 to raise an output voltage in order to realize negative voltage conversion. By combining the effective range of ADC, the optimal gain (1 +9.9k omega/RG) of AD8421 is 12.5 times, the gain resistance RG is 860.8 omega, 1k omega precision resistance with 0.1% precision is selected, and the actual gain is 10.9 times. The sensor array is designed into a circular ring PCB with the inner diameter of 60mm and the outer diameter of 100mm, 8 TMR2001 chips are uniformly arranged on the circumference with the radius of 35mm in a mode that a sensitive shaft is tangent to the circumference, and the position accuracy of an element is guaranteed by utilizing silk-screen positioning and surface tension of molten soldering tin during welding.

In this embodiment, the zero point of the sensor is eliminated, and another premise for accurately measuring the current is that the sensor can accurately reflect the magnetic field strength. However, due to the difference in manufacturing processes, the TMR element also outputs a voltage called an offset voltage V0 when the magnetic field strength H =0, and table 1 shows some performance parameters of the TMR2001 when the power is supplied at 1V.

It can be seen that the offset voltage V0 is large (-30 mV), which has a non-negligible effect on the system zero.

TABLE 1TMR2001 magnetic sensor Performance parameters (Vcc = 1.0V)

The TMR sensor output voltage Vout under the influence of magnetic induction B can be expressed as:

V _out ＝SEN·V _cc ·B+V ₀ (8)

the actual value of the magnetic induction B can be calculated according to equation (9), and table 1 already gives typical values of SEN, but V0 has only a rough range and cannot be treated as a fixed value, so Vout and V0 need to be measured. Wherein Vout can be measured by ADC, and V0 is measured separately for each TMR chip under magnetic shielding condition, when the number of sensors is large

Certain implementation difficulty. Another measurement method is derived based on the principle of the array method. As can be seen from equations (6) and (7), the current I is a weighted sum of the magnetic induction intensities Bi:

substituting formula (9) for formula (10) yields:

In this embodiment, after software programming and system assembly are completed, voltage acquisition, data processing, zeroing and other operations are implemented by an internal program of the microcontroller, and the specific flow is as follows:

(3) Every interval of 1s, all the instantaneous values of the current are averaged to calculate a direct current component, a root mean square is calculated to calculate an alternating current component, current data are sent to an upper computer through a wireless module, and the program flow of a microprocessor is shown in fig. 7.

In this embodiment, a TMR current sensor test platform is built in a laboratory environment, as shown in fig. 8, the TMR current sensor test platform includes an ac/dc current source, a TMR current sensor, an oscilloscope (taeks TDS 2012B), a clamp ammeter (Flukei 30 s), a long straight conductor (length 90cm, diameter 30mm copper rod) and a plurality of wires (2.5mm 2 single core copper wire).

During system test, a standard source generates standard direct current or power frequency current to be tested, a lead is connected with a power supply and a copper bar conductor, the conductor penetrates through the middle of a TMR sensor array, a clamp meter Flubei 30s observes the current to be tested (Flubei 30s is only used for observing current waveforms, is not used for measurement, and the current is sent out by the standard source and is a known quantity), an oscilloscope observes output waveforms of the clamp meter and sensor output signal waveforms conditioned by an instrument amplifier, a notebook computer is used as an upper computer to send control instructions and receive and display current measurement data, and fig. 9 is a test platform object diagram.

In this embodiment, in S4, in the experiment result and analysis, when no current exists in the loop, under the combined action of the offset voltage of the sensor and the background magnetic field of the laboratory, the outputs of 8 instrumentation amplifiers are superimposed with an interference amount on the basis of 1.5V offset, and the output voltage value is distributed in the range of 1.164V to 1.701V. The influence of the background magnetic field is eliminated after the numerical integration, but the existence of offset voltage enables the current measurement value at the moment to still show-3.11A, and the necessity of the zero setting operation of the system is proved again. The measured value of the TMR current sensor after zero setting is 0.03A, the position and the posture of the current sensor are randomly adjusted, and the measured value of the current fluctuates within a small range of-0.02A-0.06A, which shows that the system has good shielding effect on the background magnetic field in the laboratory environment, and the zero setting is still effective even if the position of the system is changed. When power frequency currents of 6A and 12A are given, the output voltage waveform of one AD8421 is shown in fig. 10, so that the conditioned TMR output signal is raised by 1.62V, is in a sine waveform without obvious distortion, has the same frequency and phase with the measured current, and has the amplitude which is increased in equal proportion to the measured current.

In this embodiment, the best measurement effect can be achieved by placing the conductor perpendicular to the TMR sensor array plane and at the center of the circumference during the test. Taking a complex trapezoidal formula as a product algorithm, respectively measuring 0-16A direct current and power frequency current after completing zero setting operation, and the measurement data are shown in table 2.

TABLE 2 TMR Current sensor measurement data

In the embodiment, in order to test the measurement error of the system under the non-ideal test condition, the conductor is respectively processed by eccentricity of 15mm and inclination of 45 degrees, and the influence of current position and posture adjustment on the measurement precision is researched; applying a 6A close-distance parallel conductor current outside the circular ring to test the anti-magnetic field interference capability of the system; in addition, a hard straight copper bar conductor is replaced by a soft bent wire, and the influence of the shape of the wire on the measurement result of the device is researched. In the experimental process, the product calculation algorithm is changed, the test result shown in the effective graph 12 of two product calculation formulas of a complex trapezoid and a complex Simpson is compared with the error change trend of a simulation experiment, and the error change trend is reduced along with the reduction of the interference magnetic field and is increased along with the increase of the eccentric distance. The effect of the complex trapezoidal formula is slightly better than that of the complex Simpson formula. When small current is measured, a magnetic field generated by the current to be measured is weaker than a background interference magnetic field, the TMR sensor is mainly influenced by the interference magnetic field, and magnetic field images along an integral path are often changed violently, so that the error is slightly larger. When the current is increased, the magnetic field of the current to be measured gradually occupies a dominant position, and the measurement error is reduced.

In this embodiment, fig. 12 (a), (b) show that the current eccentricity or inclination to be measured causes slight interference to the measurement, but the device is not sensitive to the small-amplitude position and posture adjustment of the wire, and the measurement error is still small even under the condition of eccentricity of 15mm (42.9% of the circumferential radius of the sensitive shaft) or inclination of 45 °; fig. 12 (c) shows that the close-range strong interference current brings a certain error to the measurement result, especially when the current to be measured is small, the error is large, but as the distance between the two increases, or when the current to be measured increases to be equivalent to the interference current, the error is recovered to be small; FIG. 12 (d) shows that the change in the shape and thickness of the wire has little influence on the measurement results. The device has no special requirement on the shape of the lead.

In this embodiment, from the linear range (± 5 Oe) of the TMR2001 and the sensor array size (radius r =35 mm), the linear measurement range of the current sensor can be calculated to be 87.5A. The current above 16A is not tested herein, limited by the output current capability of standard current sources. But it can be determined that the larger the measured current is, the less the external interference is, the signal-to-noise ratio is improved, and the measurement accuracy should be higher

In order to achieve a good measurement effect, the peak of the B (x) curve should be avoided, and the sensor should be prevented from falling into a nonlinear magnetic field range, so that the device is required to restrict the position and the posture of the conducting wire, and the current is not allowed to be tightly attached to the TMR chip. In practical application, the sensor array can be subjected to measures of reducing the inner diameter, increasing the outer diameter, increasing the thickness of a device shell and the like, so that the distance between the current and the sensor is controlled within a reasonable range.

Compared with the existing research, the TMR current sensor uses the magnetic sensor array to be matched with numerical values to calculate the current value by means of integration, a magnetism gathering or magnetic shielding structure is not needed, the problem of magnetic ring saturation does not exist, the measuring range only depends on the TMR current sensor, and the TMR current sensor is very suitable for measuring large current. The linear motor has the advantages of high precision, high linearity, small volume, low cost, simple structure and the like. By changing the array size and the model of the TMR sensor, the TMR sensor can be easily applied to current measurement in different ranges.

The working mode is as follows:

(1) The design of the non-contact current sensor using the circular TMR sensor array indicates that the accurate measurement of the magnetic field and the improvement of the numerical value integration precision are the keys for realizing high-precision measurement.

(2) Establishing a simulation model of a complex interference magnetic field environment, researching the influence of various factors on the quadrature precision, analyzing error sources and providing specific measures for reducing errors; based on the working principle of an array method, a new zero compensation method is provided, and repeated and tedious measurement work is avoided;

(3) Manufacturing a current sensor prototype and building an experiment platform for testing; the result shows that the device has good shielding effect on the background interference magnetic field, has higher linearity and measurement accuracy, and has the nonlinear errors of 0.215 percent of direct current and 0.462 percent of power frequency alternating current in the measurement range of 0-16A, and the maximum relative errors of 0.31 percent and 1.00 percent respectively;

(4) The developed device is insensitive to the position of a current conductor and the posture of a sensor, small-amplitude eccentric inclination of a lead or change of the shape and thickness of the lead can not cause obvious influence on a measurement result, and certain measurement precision can be achieved under the environment of a strong interference magnetic field of an impressed current.

The present invention is not limited to the above preferred embodiments, and any other various products can be obtained by anyone in light of the present invention, but any changes in shape or structure thereof, which are similar or equivalent to the technical solution of the present invention, are within the protection scope.

Claims

1. A grid current usage measurement method comprising the steps of:

s1: a current measurement based on a sensor array and a numerical integration; the method comprises the measurement principle and a Newton-Cotes product solving formula;

s5: and (6) summarizing the experiment.

2. A method of measuring grid current usage according to claim 1, wherein: in S1, the measurement principle comprises the following steps:

first, the TMR element utilizes the tunnel magnetoresistance effect of a magnetic multilayer film material;

b is magnetic induction intensity, which is formed by superposing a magnetic field generated by the measured current and other interference magnetic fields; l is a ring-shaped closed integration path containing a magnetoresistive sensor;

because the current generating the interference magnetic field is outside the integration path, the influence on the integration result is not generated, namely the integral contribution of the interference magnetic field on the l is 0, and the influence of the interference magnetic field can be eliminated by accurately calculating the integral;

taking a circle as an integral path, decomposing the magnetic induction B of any point P on the circle according to the tangent line B1 and the normal line B2 of the circle and the direction B3 vertical to the plane formed by the B1 and the B2, and only considering the integral of the tangential direction component B1 on l as the B2 and the B3 are vertical to the vector infinitesimal dl;

a plurality of single-axis TMR sensors are arranged in a mode that the sensitive axis is located on the circumferential tangent line to measure B1, and the output signals of the sensors are integrated to obtain the current I:

the key to accurately measuring the current is to accurately measure the magnetic field and ensure the accuracy of the integral calculation.

3. A grid current usage measurement method as claimed in claim 1, wherein: in S1, for an integrand f (x) in an interval [ a, b ], given function values f (xk) at a plurality of different nodes xk in the interval, a Lagrange interpolation function Ln (x) is approximated to f (x) to obtain an interpolation type integrand formula:

where ω k is a weight coefficient determined by Ln (x); if the formula (3) is true for the polynomial with any degree not exceeding p, but the polynomial with partial degree of p +1 is not true, the precision is called as p-degree algebraic precision; when the interpolation nodes are equidistant, the value of ω k depends only on m and k, called Cotes coefficients C (m, k), and equation (3) can be expressed as a Newton-Cotes equation; if m =1, a trapezoidal product-solving formula which needs 2 integral nodes and has 1-order algebraic precision is obtained:

equally dividing [ a, b ] into n small intervals [ xi, xi +1], wherein i =0,1, ·, n-1, the length h = (b-a)/n of each small interval; and (3) respectively using the formula (4) and the formula (5) to obtain a complex trapezoidal formula and a complex Simpson formula:

wherein oddi and eveni respectively represent that i is an odd number and a nonzero even number; when calculating the current, the plurality of TMR sensors form a mutually different node xk, the magnetic induction measurement value is f (xk), the magnetic induction B (x) is a function of position, and f (x) is B (x) in equations (6) and (7).

4. A grid current usage measurement method as claimed in claim 1, wherein: in the step S2, a simulation model and an experimental result are established, in order to verify the accuracy of the numerical value quadrature algorithm and research the application range of the numerical value quadrature algorithm, the simulation model is established by using ANSYSMAXwell and Matlab, and a scene with a complex interference magnetic field is simulated.

5. A grid current usage measurement method as claimed in claim 1, wherein: in S2, error analysis of the product-finding algorithm: the basic idea of numerical value integration is that a constructor Ln (x) replaces an integrand B (x), a fixed integral of the Ln (x) is used for approximating an integral true value, FIG. 4 is the geometrical meaning of two integration formulas, and a formula (6) forms an area approximation integral value by a broken line and an x axis in each cell; in the formula (7), the area of each two small intervals is approximate to an integral value by a parabola and an x axis; the remainder is an error source of the product formula, and the geometric meaning refers to the algebraic sum of areas surrounded by B (x) and Ln (x); the size of the remainder is closely related to the change of B (x);

when B (x) changes violently, ln (x) in a peak interval deviates from B (x) seriously, ln (x) of two formulas cannot be well fitted with a real curve, the B (x) curve between adjacent TMR sensors is generally closer to an S shape and is between a straight line and a parabola, and the errors of two kinds of integration algorithms are increased reversely;

no matter what kind of product formula is adopted, increasing n will improve the approximation degree of Ln (x) and B (x), when n approaches infinity, the two approach infinitely, the integral value also approaches to the true value; the radius r of the annular array has important influence on the interference resistance of the external magnetic field, the smaller r is, the smaller the influence of the external magnetic field is, and the more accurate the integral calculation result is, because B (x) changes when the radius is smaller;

the integral calculation algorithm tends to be gentle, the integral calculation algorithm has better performance, and on the contrary, the larger r is, the larger the influence of the interference magnetic field is, the larger B (x) oscillation is, and the integral error is increased; when the wire deviates from the center, the magnetic field at the position of the sensor changes along with the wire, and the output of the sensor close to the current wire is sharply increased due to excessive deviation, so that the B (x) curve has a peak to cause the reduction of the integration precision.

6. A method of measuring grid current usage according to claim 1, wherein: in the S3, a hardware circuit design is carried out, and a hardware circuit block diagram of the measurement system comprises a power supply module, a TMR sensor array, a signal conditioning circuit, a microcontroller and a wireless communication module; in consideration of power supply convenience, a microUSB interface is reserved, a 5V mobile phone charger is matched to supply power to the whole system, and multilevel voltage generated by conversion of a power supply module meets the requirement of electric energy supply of each element; after being conditioned by an instrument amplifier, the output signal of the sensor is sent to an ADC (analog to digital converter) for sampling, is calculated and processed by a microcontroller msp430F149, and finally is communicated with an upper computer through a wireless transceiver, and the upper computer is responsible for sending a control instruction, receiving and displaying current data.

A TMR2001 single-sensitive-axis magnetic sensor is used, which comprises 4 non-shielding high-sensitivity TMR sensing elements, and adopts a push-pull Wheatstone full-bridge structure, when an external magnetic field is parallel to the sensitive direction of the sensor, the bridge provides differential voltage output, the linear magnetic field interval is +/-5 Oe, the magnetic sensitivity is 8mV/V/Oe, the saturation magnetic fields are-25 Oe and +40Oe, and a typical output curve of the sensor when power is supplied at 1V is shown in figure 6.

7. A method of measuring grid current usage according to claim 1, wherein: in S3, the zero point of the sensor is eliminated, and another premise of accurately measuring the current is that the sensor can accurately reflect the magnetic field strength, and due to the difference in the manufacturing process, the TMR element also outputs a voltage, called an offset voltage V0, when the magnetic field strength H =0, where table 1 shows a part of performance parameters of the TMR2001 when power is supplied at 1V;

the offset voltage V0 is large (-30 mV-30 mV), and the influence of the voltage on the system zero point is not negligible;

V _out ＝SEN·V _cc ·B+V ₀ (8)

the actual value of the magnetic induction B can be calculated according to equation (9), the SEN typical value is given in table 1, but V0 has only an approximate range and cannot be treated as a fixed value, so Vout and V0 need to be measured;

vout can be measured by an ADC (analog to digital converter), V0 needs to be measured independently for each TMR chip under the condition of magnetic shielding, and certain implementation difficulty exists when the number of sensors is large; deriving another measuring method based on the principle of an array method; as can be seen from equations (6) and (7), the current I is a weighted sum of the magnetic induction Bi:

by substituting formula (9) for formula (10), we obtain:

in the formula: subscript i represents the ith TMR sensor, when no current exists in the loop, the above formula should be equal to zero, then the ADC is used for collecting the sensor output voltage Vout in any environment, the offset voltage integral of TMR can be reversely deduced, and when the current is measured, the correction is carried out according to the formula (11), namely, the zero point of the current sensor can be eliminated.

8. A grid current usage measurement method as claimed in claim 1, wherein: in S3, after software program design and system assembly are completed, voltage acquisition, data processing, zero setting and other operations are all realized by internal programs of the microcontroller.

9. A grid current usage measurement method as claimed in claim 1, wherein: in S4, a TMR current sensor test platform is set up in a laboratory environment and comprises an alternating current and direct current source, a TMR current sensor, an oscilloscope, a clamp ammeter, a long straight conductor and a plurality of leads;

experimental results and analysis, during testing, the conductor is perpendicular to the plane of the TMR sensor array and is positioned at the center of the circumference, so that the best measuring effect can be achieved, the complex trapezoidal formula is used as a product algorithm, and 0-16A direct current and power frequency current are respectively measured after zero setting operation is completed.