Sensor of a quantity in a medium
The sensor employs a feedback loop to adjust excitation frequency for constant sensitivity, addressing non-linear sensitivity issues in conventional sensors, enhancing accuracy and precision in permittivity measurements.
Patent Information
- Authority / Receiving Office
- FR · FR
- Patent Type
- Applications
- Current Assignee / Owner
- ELECTRICITE DE FRANCE
- Filing Date
- 2024-12-23
- Publication Date
- 2026-06-26
Smart Images
Figure 00000000_0000_ABST
Abstract
Description
Title of the invention: Sensor for a quantity of a medium technical field
[0001] The present disclosure falls within the domain of sensors, for characterizing a medium by one of its quantities such as, for example, by its permittivity. Previous technique
[0002] Such sensors can operate in particular by reflectometry, for example for the detection of defects in particular in conductive structures such as electrical cables.
[0003] These sensors can also operate by transmission provided that they include a measuring head (or "sensitive element" hereafter), equipped with at least two pins to be inserted into the medium to be measured, a waveguide being formed following the injection of an excitation signal (or "input signal" hereafter).
[0004] Taking the example of reflectometry, this is a measurement technique commonly used in fault detection, particularly in cables. It can be frequency-domain reflectometry (FDR) or time-domain reflectometry (TDR). In both cases, the principle of reflectometry is based on the analysis of a reflected wave (contributing to the formation of a signal called the "output signal" hereafter) as a function of the emitted wave (contributing to the formation of the aforementioned "input signal"). The variations of these waves in the measured medium are based on the variation of the characteristic impedance of the propagation medium formed by this medium, which causes the reflected wave ratio to vary. This variation is measurable after analysis of the input and output signals by electronic means.
[0005] Time-domain reflectometry (TDR) is primarily used to characterize and locate defects in metallic cables. The measuring tool sends a pulse to the input of a cable and measures the propagation time of the wave before it is reflected by a defect. Indeed, a defect in a cable causes a variation in the characteristic impedance, which results in the reflection of a portion of the emitted wave and the collection of a corresponding signal. Analysis of the reflected signal allows information to be deduced about the system or medium under consideration. Reflectometry is therefore a non-destructive testing method.
[0006] In what follows, the excitation and analysis are situated more in the frequency domain, and typically, frequency response reflectometry (FDR) incorporates the main characteristics of TDR. However, the analysis is carried out in the frequency domain and it is a standing wave ratio (“SWR”) which is generally analyzed.
[0007] During a measurement of water level or permittivity, particularly in what follows, the nature of the theoretical relationships reveals a strong non-linearity over the entire measurement range. The response curve of a conventional vector reflectometer with respect to a water level is given in [Fig. 1].
[0008] Fig. 1 illustrates by way of example the phase variation (in degrees) as a function of a water height varying from 0 to 1 m for a fixed frequency of 7 MHz.
[0009] In the graph of [Fig. 1], a strong variation in phase appears for a water height ranging from 0 to approximately 40 cm. This variation gradually decreases to become almost zero for greater water heights (typically above 40 cm).
[0010] To determine the accuracy of the measurement as a function of height, it is useful to know the intrinsic sensitivity of the sensor. One possible technique consists of deriving the variation of the phase of [Fig. 1] with respect to the water level, such as:
[0011] (4.1) i)h. dh. UILeau ulLeau
[0012] This intrinsic sensitivity is specific to the sensor developed and is expressed in ° / m (degrees per meter).
[0013] To obtain the sensor sensitivity in another way, for example in volts per meter (V / m), the sensing element is then excited at a fixed frequency. The final sensitivity of the sensor is then: e - , . (4.2) where sq> denotes the sensitivity, expressed in volts per degree, of a demodulation block used to obtain the measurand. In the case of this study, the sensitivity Sv is primarily due to the type of demodulation used.
[0014] Figure 2 shows the evolution of the intrinsic sensitivity of the sensor over the measurement range, for a water height ranging from 0 to 1 m. This evolution is not constant as the water height changes and is maximum at approximately 20 cm.
[0015] Such a variation, exhibiting at least one extremum, is unsatisfactory, and it is desirable to obtain, for example, a constant sensitivity value over the entire measurement range. Constant sensitivity over the entire measurement range would also allow for constant resolution and uncertainty over that entire range. Summary
[0016] This disclosure improves the situation.
[0017] A sensor is proposed for measuring a quantity of a medium, the sensor comprising: - at least one sensitive, suitable element: * to inject waves, the injection being controlled by the application of an input signal, and * to collect received waves, disturbed by physical properties of the medium related to said quantity, to form an output signal comprising, relative to the input signal, a modulation from which is determined a measurand indicative of said quantity of the medium, - a control unit, capable of generating said input signal with a wave frequency to be injected corresponding to a chosen excitation frequency, and - a processing unit applying a demodulation of the output signal to determine the corresponding measurand. Specifically, the processing unit is configured to cooperate with the control unit to generate the input signal, and the excitation frequency is adjusted based on the input and output signals using a feedback loop within the processing unit. This maintains a linear response of the output signal over a range of measurands by continuously adjusting the excitation frequency (the measurand range being thus wide, for example, exceeding a chosen threshold). The feedback loop is configured to ensure a constant (or nearly constant) sensor sensitivity over the aforementioned range.
[0018] Such an implementation thus makes it possible to have a "linear" (or at least strictly monotonic) response of the sensor with respect to the excitation, and this over a wide range of operands.
[0019] In an embodiment where the output signal is obtained in the form of electrical voltages, the processing unit includes a voltage-controlled oscillator (or "VCO") to adjust the excitation frequency according to the output signal at least.
[0020] For example, the processing unit can be configured to digitize electrical voltages and can thus include a microcontroller to process the digitized output voltages in order to adjust the excitation frequency of the input signal.
[0021] In this embodiment, the input signal can be generated in the form of a digital voltage of the SPI type (or "Serial Peripheral Interface").
[0022] The aforementioned feedback loop may include a controller of the type PID (or "Proportional Integral Differential"), PI ("Proportional Integral") or I ("Integral").
[0023] In an analog version of the processing unit, the latter may alternatively include a phase-locked loop (PLL) between the oscillator (VCO) and the control unit.
[0024] Such a technique with feedback loop to linearize the sensor response can be applied to any type of sensor, for example configured to operate in reflectometry, and / or in magnetometry, or others.
[0025] The sensor can operate at high frequencies, for example at an excitation frequency on the order of a few megahertz (more generally from a few tens of kHz to 2GHz, or even more by adapting the electronics).
[0026] For example, a satisfactory frequency may be on the order of (v / 4) / L (where v is the speed of the waves in the medium, L the length of the propagation medium (for a bifilar rod, for example)).
[0027] In one embodiment, the sensor has a sensitivity that is a function of the measurand, the excitation frequency, and an injection power of the input signal, and the injection power is fixed at a chosen value, for: - to make the sensor's sensitivity dependent only on the measurand and the excitation frequency, and - determine a pair of excitation frequency and produced measurand, for which the sensitivity of the sensor is maximal.
[0028] Such an implementation aimed at maximum sensitivity makes it possible to minimize measurement noise and to optimize it.
[0029] In an embodiment, the quantity of the medium that the sensor measures is the permittivity of the medium (typically the dielectric permittivity).
[0030] In such an embodiment, the permittivity of the medium depends on at least one parameter such as the temperature of the medium, and comprising at least one second sensitive element, the processing unit being configured to deduce, from the measurands obtained via the two sensitive elements, permittivity values characterizing the medium independently of said parameter such as the temperature of the medium.
[0031] In one embodiment, the aforementioned waves are radio frequency waves and the sensing element comprises at least two elements to be introduced into the medium for: * inject the waves via said elements, the injection being controlled by the application of an input signal, to form a waveguide between the elements and, from there, * collect the received waves.
[0032] In one embodiment, the two aforementioned elements can be identical metallic pins to be introduced into the medium in order to generate a waveguide between them, this waveguide then depending on the medium (and its dielectric permittivity in particular).
[0033] An application of the sensor presented above to the detection of defects in conductive structures such as, in particular, electrical cables is also envisaged.
[0034] A microcontroller comprising a processing circuit for implementing feedback based on digital signals is also targeted in the digital implementation version of the process.
[0035] Also referred to is a computer program comprising instructions for implementing digital signal-based feedback in such a process, when executed by a processor. According to another aspect, a non-transient, computer-readable recording medium is proposed on which such a program is recorded. Brief description of the drawings
[0036] Other features, details and advantages will become apparent from reading the detailed description below and from analyzing the accompanying drawings, in which: Fig. 1
[0037] [Fig.1] illustrates the phase variation (in degrees) as a function of water height varying from 0 to 1 m for a fixed frequency of 7 MHz with a prior art sensor. Fig. 2
[0038] [Fig.2] illustrates the evolution of the intrinsic sensitivity of a prior art sensor. Fig. 3
[0039] [Fig.3] is a schematic diagram of the loop of a reflectometer (version digital), according to a particular embodiment. Fig. 4
[0040] [Fig.4] is a schematic diagram implementing a PLL loop for looping of the output to the input (analog version), according to one embodiment. Fig. 5
[0041] [Fig.5] is a schematic diagram of the measuring head introduced into a material. Fig. 6a
[0042] [Fig.6a] is a schematic diagram of the measuring head introduced into a double material. Fig. 6b
[0043] [Fig. 6b] illustrates a simulation of the phase variation under the conditions of the [Fig.6a] Fig. 7
[0044] [Fig.7] illustrates the principle of variation of the relative permittivity as a function of the frequency and temperature. Fig. 8
[0045] [Fig.8] is a schematic diagram implementing a PLL loop for looping output to input and multiplexing on different measurement heads. Fig. 9
[0046] [Fig.9] illustrates the principle of measuring the depth of penetration into the soil envisaged on the one hand (part a), and with a humidity mapping on the other hand (part b). Fig. 10
[0047] [Fig. 10] is a schematic diagram of a magnetometer in particular with a feedback loop according to an embodiment of the present description. Fig. 11
[0048] [Fig. 11] illustrates the electronic architecture using the reflection coefficient Su, with a bidirectional coupler as in [Fig.3] in particular.
[0049] Fig. 12
[0050] [Fig. 12] illustrates the electronic architecture using the S21 transmission coefficient with here a directional coupler. Fig. 13
[0051] [Fig. 13] illustrates an electronic architecture using the reflection coefficient Su with a double excitation. Fig. 14
[0052] [Fig. 14] illustrates the electronic architecture using the transmission coefficient S2i with a double excitation. Description of the implementation methods
[0053] A sensor is proposed which may be of the reflectometer type, similar for example to a magnetic sensor, and thus capable of extracting the measurand from a modulation signal. To this end, the reflectometer, which may be vector-based here, relies on: - the use of a high-frequency excitation, - a sensitive element for capture (typically a measuring head), and - a demodulation unit (envelope, phase detection or so-called “quadrature” detection).
[0054] Like giant magnetoimpedance (GMI) magnetic sensors or so-called "magnonic" devices (magnons being excitation modes or "spin waves"), the response of measurement heads used in reflectometry is largely nonlinear. This nonlinearity is due to the propagation equations of the wave involved in this type of sensor. In order to obtain a linear response over a wide range of measurands and to guarantee identical sensitivity over this range of measurements, feedback of the output signal to the input must be implemented. However, in the context of a measurement by reflectometry, for example, with the different nature of the output and input signals, the output signal cannot interact directly with the measurand. To circumvent this problem related to In the sensor implementation, the output signal is processed before being injected as a variation in the sensor's excitation frequency or the phase between the two excitation sources. However, varying the phase between the two sources is difficult to implement in practice, and varying the excitation frequency is easier.
[0055] In the context of a magnetic sensor feedback loop or a phase-locked loop (PLL), the injected signal is of the same nature as the measurand. In the context of vector reflectometry, the principle of such a feedback loop can be used to linearize the sensor characteristic. The term "linearization" used here can be generalized to the principle of varying the sensor's response to excitation according to a strictly monotonic (e.g., increasing) function.
[0056] Thus, the range of measurands over which the sensor response is "linear" is thus broadened compared to the examples shown in Figures 1 and 2. For example, this range can be wider than a chosen threshold, for example over a distance in meters of water height, and linear for example over 20 meters, between 0 and 20 meters for example (which is a clear advantage over prior art sensors for a large number of applications).
[0057] To this end, it is preferable to describe the measurement principle of a vector network analyzer (VNA). This tool makes it possible to measure the linear characteristics of a circuit or a circuit element, as a function of: - the excitation frequency, - the injection power, and, - in this case, the measurand (quantity) to be measured.
[0058] This measurement results in complex parameters in the form of a matrix, denoted [Smn], such that: (5.1) where IS111 denotes the amplitude of the characterized parameter and q>n denotes the angle of this parameter. The size of the matrix depends on the type of system to be characterized. As an example here, the VNA may have two ports and the matrix may be 2x2 in size.
[0059] In the context of reflectometry, when a signal is excited on the head of As a measure, the parameter matrix simplifies by: [ç 1 - (5-2) PmnJ ' s**
[0060] The VNA vector network analyzer then makes it possible to characterize the measuring head in its environment as a function of: - the measurand, denoted XM, - the static part of the measurand, the excitation frequency / 0 and the injection power Po, as follows: (5.3)
[0061] These parameters thus measured are well associated with the characterization of the sensitive element, in particular in reflection.
[0062] An additional treatment, implemented here to ensure a feedback loop of the sensor, is described below.
[0063] By fixing the input power Po, it appears that the reflection parameter SI 1 depends only on the measurand XM and the excitation frequency. In this case, the sensor sensitivity, denoted Sv, also depends on these parameters. A detailed implementation of this type of sensor is presented later. In most cases, the objective is to find the frequency and measurand parameters for which the sensor sensitivity is maximized in order to minimize measurement noise and optimize it.
[0064] In the context of this description, the aim is to find, more precisely, a sensitivity that is not a maximum sensitivity as a function of the parameters ( / Ô,XM) around an operating point, but a maximum sensitivity that is also constant over a large operating range of the measurand (foe | / min ;f mJ,XMe[XMmin ; XMmax]). This implies adapting a variable excitation frequency from defmin to fmax.
[0065] To this end, a digital loop can be implemented, in an embodiment presented below.
[0066] To achieve the desired result, a feedback loop, particularly on the excitation frequency, is implemented. The sensor output voltages, denoted v, r, and vq, are digitized and processed by a microcontroller, for example, an Analog-to-Digital Converter (ADC) as illustrated in [Fig. 3]. After information processing, a control signal is generated by this microcontroller, which is proportional to the desired excitation frequency via a voltage-controlled oscillator (VCO). This control signal can be in the form of a voltage, a digital signal such as an SPI (Serial Peripheral Interface) signal, or any other digital form. The use of a digital VCO allows the excitation frequency to be managed directly by this microcontroller. Furthermore, the main advantage of this setup is that it provides considerable flexibility in managing the control loop, thus allowing the parameters to be chosen according to a specific use case.The numerical correction can therefore contain a “PID” type corrector, or “PI” or simply “I” (PID for “Proportional Integral Differential”, PI for . “Proportional Integral”, I for “Integral”). Fig. 3 represents the schematic diagram of such a numerical implementation.
[0067] An alternative embodiment may consist of using a Phase-Locked Loop (PLL) to control the excitation frequency or the output frequency of the sensor according to a given setpoint. Here, tracking the setpoint phase maximizes the sensor's sensitivity. A schematic diagram is shown in [Fig. 4]. The use of a PLL is based primarily on comparing the phases of the two input signals (CP). The output of this comparator is then filtered and constitutes the control signal for the oscillator (VCO). This oscillator then controls the excitation frequency of the measuring head.
[0068] In particular, [Fig.4] represents an analog implementation of a vector reflectometer using a phase-locked loop (PLL) for looping the output back to the input, while [Fig.3] represents a digital version with a digital controller preceded by an analog-to-digital converter and followed by a digital-to-analog converter.
[0069] Typically, the measuring head (h(f0)) is the sensing element that interacts with the material to be measured. The measuring head is excited by a high-frequency signal f0. Here, it may consist of pins propagating radio-frequency waves to form a waveguide between the pins, the properties of which depend on the medium, particularly its permittivity, thus allowing the medium to be characterized. Alternatively, however, it may be a coaxial or other structure introduced into the medium.
[0070] Next, couplers allow the input signal to be split into several channels for different processing. Amplifiers G2 and G1 increase the amplitude of the signals to make them usable by the other components. Mixers, particularly FQ, combine the signals to obtain in-phase (I) and quadrature (Q) components, which allows the signal to be demodulated. A phase comparator (CP) compares the phase of the reference signal rf in with that of the measured signal 10. The voltage-controlled oscillator (or "VCO") generates a signal whose frequency is controlled by an input voltage. This voltage is adjusted by the feedback loop to maintain the correct phase. A low-pass filter is also provided to eliminate the high-frequency components of the phase comparator's output signal, allowing only the low-frequency components to pass.In addition, voltage amplifiers G3 and G4 are provided to increase the amplitude of the reference signals v re / to make them usable by the other components. The frequency divider / - VCo / N divides the VCO output frequency by a factor N to obtain a stable reference frequency.
[0071] Thus, when the measuring head is excited by a high-frequency signal, the signal is split by the couplers and amplified by amplifiers G2 and G1. The mixers demodulate the signal into in-phase (I) and quadrature (Q) components. The phase comparator (CP) compares the phase of the reference signal rf in with that of the measured signal. In the feedback loop, the output of the phase comparator is filtered by a low-pass filter and used to control the frequency of the VCO, and the VCO adjusts its frequency to maintain the correct phase, thus creating this feedback loop. Finally, at the output, the output signals are proportional to the amplitude and phase of the reflected signal, respectively. These signals can then be processed to obtain the final measurements.
[0072] By such an embodiment, it is possible to linearize the response of the vector reflectometer over a wide range of measurements by using a phase-locked loop (PLL) in an analog version of the device. This makes it possible to obtain constant sensitivity and improved accuracy of permittivity measurements using phase-demodulation feedback looping techniques.
[0073] A situation in quasi-real conditions of the measurement of the penetration of the measuring head as a function of the permittivity of the ambient medium is presented below.
[0074] The measurement conventionally used to determine the penetration depth of the head in a material, denoted h, is based on a fixed permittivity value er. Figure 5 shows the measuring head inserted into a medium with a permittivity value er2 without compensation for this value.
[0075] According to a radio frequency model of an open-circuit line, the phase variation depends primarily on the permittivity value er. The higher this value, the greater the phase variation. Furthermore, the length of the rods can play a significant role in the chosen excitation frequency. Ideally, this length should remain less than one-quarter of the wavelength, which can be denoted Xe, i.e.: [0076 1 <5 - 4)
[0077] The variation in measurement with the penetration distance in a double-material is shown below by way of illustration.
[0078] By applying the measurement principle described above, it is possible to measure the penetration depth of a measuring head in a second material while knowing the height of the first material being probed. The specific procedure described below can be followed for this purpose.
[0079] The measurement is started while the head is not in contact with any material other than free air, a material in which the permittivity is such that er=l.
[0080] The measuring head is then introduced into the first material. The measured permittivity value increases, with er= er2>l, which generates a phase change.
[0081] The head continues its introduction and encounters the second material. The permittivity value varies with er3^ er2. In this case, the phase variation is modified and a "break" can be observed.
[0082] This procedure then makes it possible to probe different materials, constituting a typical heterogeneous medium.
[0083] Figure 6a illustrates the measuring head embedded in a dual material having different electrical properties. Figure 6b represents a simulation of the phase evolution for a dual material at a fixed frequency:
[0084] m fh * 3]
[0085] A clear break in slope appears in the phase evolution as a function of the penetration depth of the measuring head into the dual material. The simulation replicates the analysis of the height of a coffee with foam on top. The permittivity of the foam is approximately eri=5 and the permittivity of the liquid coffee is approximately er2=80. This value is close to that of water.
[0086] To facilitate the mathematical analysis of this variation, it is easier to use the derivative of the phase as a function of time or the height of the material. The term "sensitivity of the sensor" then appears:
[0087] _ (5.5) â / î
[0088] A jump in this sensitivity value then demonstrates a change in material. Typically, if this variation is beyond an error threshold e rr, then the sensor detects a new material, i.e.: [°089]
[0090] In the context of a large phase variation, it is normal for the sensitivity to change during the measurement. By introducing an error value, denoted e rr, it is possible to define a threshold above which the aforementioned break must be taken into account. Finally, by storing the height (the level) of each material change, it is then possible to determine the thicknesses of each material.
[0091] The simulations shown in [Fig. 6b] were performed, however, with theoretical permittivity values. However, studies show that this value can depend on both the temperature and the excitation frequency. The different curves in [Fig. 7] show a dependence of the permittivity depending on the frequency but also depending on the temperature of the environment (in this case, water).
[0092] In order to compensate for the variation of this second factor, for a precise height measurement, it is proposed to measure this permittivity value cyclically. The measurement is performed by at least one second sensitive element, distributed with the first sensitive element on the measuring head. Such an embodiment can be implemented in the event of temperature variations in the medium (or significant height variations).
[0093] A radio frequency (or "rf") multiplexer with n outputs then allows the measurement head to be directed to be interrogated. The permittivity measurement is thus performed at the excitation frequency of the height measurement. For example, for a parallel-plate capacitor, er is evaluated by:
[0094] _3 ab (5.7) £ra
[0095] Where a and b represent respectively the height and width of the two plates of the capacitor, and d the distance separating them. Finally, C represents the capacitance of the capacitor and eO represents the permittivity of free space.
[0096] This results in a slight change to the measurement electronics as shown in [Fig. 8]. A multiplexer is added to the measurement head to select the sensitive element to be addressed. Addressing can be performed by a microcontroller.
[0097] An application to the measurement of the humidity of a non-homogeneous material is described below.
[0098] By cyclically measuring permittivity and depth, the sensor can measure height as closely as possible to reality. The feedback condition can be referenced with respect to time or with respect to a measured penetration distance. When the feedback is described as "temporal," a new permittivity measurement is timed according to a clock or "timer" on the microcontroller. In the case of a "spatial" feedback, the device takes a new permittivity value only when a minimum height is reached.
[0099] For example, such a measurement technique makes it possible to establish a map of soil moisture along several horizons, as illustrated in [Fig.9] (part b for the map).
[0100] A linearization of a magnetic sensor around an operating point, based on the principles described above, is described below.
[0101] The operating principle of a magnetometer is recalled beforehand, for this purpose.
[0102] The field of magnetometry uses the principle of linearization in order to extend the measurement range in which the response of the measuring instrument is linear.
[0103] A magnetic sensor based on a high-frequency system is primarily composed of an excitation, a sensing element, and a demodulation stage. The objective of the reflectometer is identical to that of a magnetic sensor, namely to extract the measurand from a modulation signal. The vector reflectometer then relies on the use of a high-frequency excitation, a sensing element (or "measuring head"), and demodulation electronics (envelope, phase detection, or quadrature detection).
[0104] Like magnetic sensors based on GMI or magnonic devices, the response of measuring heads used in reflectometry is largely non-linear. In order to obtain a linear response over a wide range of measurands and to guarantee identical sensitivity over this range of measurements, it is proposed to feed the output signal back to the input. However, in the context of a reflectometry measurement, the output signal cannot interact with the measurand. To circumvent this sensor implementation problem, the output signal is processed before being injected as a variation in the sensor's excitation frequency or in the phase between the two excitation sources. Indeed, in the context of feedback on magnetic sensors or a PLL (phase-locked loop), the injected signal is a signal of the same nature as the measurand.
[0105] Figure 10 shows the schematic diagram of a field-controlled magnetometer around an operating point, in which the sensing element transforms the measurand (Bext, the magnetic field to be detected) into a voltage vs. This voltage is then transformed into a magnetic field Bcr, using the feedback loop [3]. Thus, the magnetometer is subjected to the magnetic field difference ei = Bext - Bcr. As with a closed-loop system, this value, ei, must be zero in order to avoid non-linearity in the characteristic of a magnetic sensor as a function of the field.
[0106] A possible embodiment using high-sensitivity magnetometry is described below.
[0107] Magnetometers are increasingly present in connected objects and applications, particularly in consumer applications such as smartphones with integrated magnetic compasses, or in more specific applications ranging from precision agriculture with product flow measurement on a machine to magnetic guidance and non-destructive testing. They are very often the first link in a measurement and transmission chain that can be extremely complex.
[0108]
[0109]
[0110] [YES]
[0112]
[0113]
[0114]
[0115]
[0116]
[0117]
[0118]
[0119] A magnetometer is a measuring instrument that outputs a quantity proportional to the magnetic field being measured. It consists primarily of a magnetic sensor that converts variations in the magnetic field into an electrical quantity, and conditioning electronics that maintain the sensor in an optimal operating mode and shape the output signal. In a complete measurement chain, this signal is then digitized and transmitted to a processor for data processing. Regardless of the technology used, the general principles of sensor implementation remain the same. Sensors provide an output signal yc (most often a voltage or a current) which is a representation of the measurand xM. The desired outcome here is that the output and input be proportional to each other. The proportionality coefficient is typically related to the sensor's sensitivity, denoted S. Therefore, the sensor is said to be "linear". However, the dependence of the output signal on the measurand is generally non-linear. Furthermore, disturbances such as random fluctuations or influencing quantities (temperature, pressure, etc.) can also be taken into account. As mentioned above, since the response is generally non-linear, a Taylor approximation allows us to linearize the response around an operating point by considering small-amplitude variations in the measurand. The linearized form allows us to define a sensitivity at the operating point by: A? ~ The limitation of linear operation to small variations can be a major drawback. However, by looping the measurement system back to the measurand using an integrator, the system can be linearized over a wide range of the input signal. In the context of a magnetic sensor of the "magnetometer" type, this feedback can be carried out as follows, where T denotes a transfer function. The output voltage of a magnetometer is expressed by: T (6.2) “— “ Ber) I cul With an open-loop transfer function TBo, we have:
[0120] T ( 6.3 ) T ____ £> * Sf? ~ / Ùi
[0121] And / or a closed-loop transfer function TBF, we have:
[0122] ~ (6.4) i —y— » T "ïr ' i
[0123] It can be noted that if y »,
[0124] then y ( 6.5 )
[0125] which has the advantage of simplifying the treatments to be implemented, in practice.
[0126] More generally, in the field of magnetometry, linearization around an operating point can provide a linear representation of the magnetic field around that operating point. In this specific case, the feedback loop generates a field that is inversely proportional to and equal to the field to be measured, thus recentering the operating point on the chosen one. The feedback loop is therefore based on the measured quantity.
[0127] In the case of reflectometry described above, for example for measuring water level, applying a quantity inverse to the measured value would amount to controlling the height of the measuring head based on that height. Thus, the height of the measuring head would be a reflection of the water level. However, this implementation is not practical because it greatly complicates the system. Furthermore, the environment does not allow for the installation of such equipment.
[0128] Since feedback on the input quantity is not feasible, and the phase of the signal to be measured varies not only according to the water height, but also according to the excitation frequency, the option of an adjustable excitation frequency is retained here. Industrial application
[0129] The present technical solutions may be applied in particular to the detection of dielectric constants, in particular permittivity, for any conductive materials, in particular electrical cables or heights of water or mixtures of water and mud to be sounded, etc.
[0130] Furthermore, this disclosure is not limited to the embodiment examples presented above in the context of reflectometry only by way of example, but it encompasses other variants.
[0131] In the examples above, the water level measurement is based on measuring the variation in dielectric permittivity, denoted er, in a radio frequency waveguide. This waveguide can take various forms, such as:
[0132] two parallel rods, or a bifilar line,
[0133] or two parallel plates,
[0134] or even a “microstrips” line,
[0135] or line couplers, etc.
[0136] These line, waveguide or coupler technologies can be used in both reflection and transmission. As previously discussed and illustrated in [Fig. 1 1], the feedback principle can be implemented using the parameter Sn(θ, er) as a sensitivity parameter.
[0137] However, a transmission line with, for example, a two-microstrip coupler can be used to transmit data. The loop can be achieved using the parameter S2i( / ô, er) as the sensitivity parameter and as illustrated in [Fig. 12].
[0138] A variant of the architectures illustrated in Figures 11 and 12 can take the form of a double excitation as illustrated respectively in Figures 13 and 14.
Claims
Demands
1. A sensor for a quantity of a medium, the sensor comprising: - at least one sensitive element, capable of: * injecting waves, the injection being controlled by the application of an input signal, and * collecting received waves, perturbed by physical properties of the medium related to said quantity, to form an output signal having, with respect to the input signal, a modulation from which is determined a measurand indicative of said quantity of the medium, - a control unit, capable of generating said input signal with a wave frequency to be injected corresponding to a chosen excitation frequency, and - a processing unit applying a demodulation of the output signal to determine the corresponding measurand, characterized in that the processing unit is configured to cooperate with the control unit to generate the input signal,and in that the excitation frequency is adjusted according to the input and output signals by means of a feedback loop included in the processing unit, to maintain a linear response of the output signal over a range of measurands by continuously adjusting the excitation frequency, the range of measurands being wider than a chosen threshold, the feedback loop being thus configured to ensure constant sensitivity of the sensor over said range.
2. Sensor according to claim 1, wherein, the output signal being obtained in the form of electrical voltages, the processing unit includes a voltage-controlled oscillator (VCO) to adjust said excitation frequency as a function of the output signal at least.
3. Sensor according to claim 2, wherein the processing unit is configured to digitize electrical voltages and includes a microcontroller to process the digitized output voltages in order to adjust the excitation frequency of the input signal.
4. Sensor according to claim 3, wherein the input signal is generated in the form of a digital voltage of type SPI (Serial Peripheral Interface).
5. Sensor according to claim 4, wherein the feedback loop includes a controller of the PID (Proportional Integral Differential), PI (Proportional Integral) or I (Integral) type.
6. Sensor according to claim 2, wherein the processing unit includes a phase-locked loop (PLL) between the oscillator (VCO) and the control unit.
7. Sensor according to any one of the preceding claims, configured to operate in reflectometry.
8. Sensor according to any one of the preceding claims, configured to operate in magnetometry.
9. Sensor according to any one of the preceding claims, wherein the excitation frequency is on the order of a few megahertz.
10. Sensor according to any one of the preceding claims, comprising a sensitivity that is a function of the measurand, the excitation frequency and an injection power of the input signal, wherein the injection power is fixed at a chosen value, to: - make the sensitivity of the sensor depend only on the measurand and the excitation frequency, and - determine a pair of excitation frequency and produced measurand, for which the sensitivity of the sensor is maximum.
11. Sensor according to any one of the preceding claims, wherein the quantity of the medium that the sensor measures is the permittivity of the medium.
12. Sensor according to claim 11, wherein the permittivity of the medium depends on at least one parameter such as the temperature of the medium, and comprising at least one second sensitive element, the processing unit being configured to deduce, from the measurands obtained via the two sensitive elements, permittivity values characterizing the medium independently of said parameter such as the temperature of the medium.
13. Sensor according to any one of the preceding claims, wherein the waves are radio frequencies and the sensing element comprises at least two elements to be introduced into the medium to: * inject the waves via said elements, the injection being controlled by the application of an input signal, to form a waveguide between the elements and, therefrom, * collect the received waves.
14. Microcontroller comprising a processing circuit for implementing digital signal-based feedback in the process according to any one of claims 3 to 5.
15. Computer program comprising instructions for implementing digital signal-based feedback in the process according to any one of claims 3 to 5, when executed by a processor.