RU2399888C1 - Method of measuring level of material in reservoir - Google Patents

Abstract

FIELD: physics.

SUBSTANCE: inside zone of high measurement error caused by presence of interfering reflector, the level is determined by predicting the value of the level, determined at a previous measurement cycle, rate of change of the level measured outside the zone of high measurement error and a measured time interval between neighbouring measurement cycles. Such determination of the level excludes the effect of an interfering reflector on measurement results and measurement error falls sharply. The rate of change of the level is determined on areas where the level changes, where the effect of the interfering reflector is insignificant. The direction of change of the level or absence of any change in the level is determined from the value of an auxiliary control signal. In order to determine the border zones with high measurement error, pre-exposure is carried out, where the position of all interfering reflectors which exceed a certain preset level is determined in an empty reservoir, and in the vicinity of each, two borders of zones with high measurement accuracy are indicated in the size of the tripled value of the resolution element of the frequency-modulated signal with indicated frequency deviation.

EFFECT: high accuracy of measuring level in the presence of interfering signals caused by reflections from components of the reservoir.

10 cl, 5 dwg

Description

The invention relates to the field of measurement technology and can be used for precision level measurement of liquid materials.

Known radar method of measuring the level [1], including measuring the propagation time of radio waves radiated in the direction to the surface of the medium and reflected from it, and calculating the measured propagation time of the radio waves of the distance to the surface of the medium. The specified method does not allow to measure the level with sufficient accuracy in the presence of interfering reflections caused by the design features of the tank with liquid material, since interfering reflections distort the waveform and thereby lead to a large error in measuring the delay time.

A known method of measuring distance, implemented in the device [2], which consists in the fact that they emit a frequency-modulated signal in the direction of the contents of the tank, receive, after the propagation time, the reflected signal and mix it with part of the emitted signal to obtain a differential frequency signal (RF) . The phase of this signal is used to measure the distance to the surface of the controlled medium, provided that the difference frequency itself is kept constant by controlling the modulation period. In this case, the phase of the difference frequency signal during distance measurement will continuously vary within 2πN + φ in proportion to the distance change. Here N is an integer number of periods of the RHF contained in the modulation period, φ is the number corresponding to the remaining part of the period, that is, the initial phase of the RHF. Thus, the determination of the distance is reduced to counting the number N, measuring the phase φ and calculating the distance.

The disadvantage of this method is the inability to measure the level with a given accuracy in the presence of interfering reflections caused by structural elements of the tank, since the presence of interference greatly changes the phase of the signal and leads to a large error.

Also known is a method of measuring distance [3], which includes generating and emitting a probing signal with periodic frequency modulation in the initial frequency modulation range; receiving an echo signal, extracting part of the probe signal and mixing it with the received echo signal; calculation of the spectrum according to the RMS obtained for half the modulation period and determination of its center frequency; calculating the distance from the measured center frequency of the RFS and the number of minimum distance intervals that fit into the measured distance corresponding to the number of periods of the RFI at the half-period of modulation; reducing the initial range of frequency modulation to obtain an integer number of distance intervals corresponding to an integer number of periods of the UHF at the half-period of modulation and the maximum spectral component corresponding to the measured distance and calculating the frequency distance of this maximum spectral component of the UHF.

The disadvantage of this method is the inability to measure the level with high accuracy near the interfering reflectors, since the main and side lobes of the interference spectrum distort the shape of the main lobe of the signal spectrum and thereby cause a large error in determining the true position of the extremum of the spectrum.

The closest technical solution for the totality of essential features (prototype) is a level measurement method [4] with a radar oriented to the maximum of the echo signal from the measured surface and including radiation from a sequence of microwave signals in which discrete angular frequencies are uniformly distributed over the scanned frequency range. The signal reflected from the measured surface is mixed with a part of the emitted signal to obtain two differential frequency signals phase-shifted relative to each other by an angle π / 2 (quadrature signals), which, after analog-to-digital conversion, are fed to the microprocessor. Processing the sequence of quadrature samples includes the inverse discrete Fourier transform, the use of the MUZIC high-resolution method for localizing the sources of the most powerful radiation, the selection of echo signals in the time domain, and ultimately obtaining the distance to the measured level.

The disadvantage of this method is the insufficiently high accuracy of level measurement in the presence of interfering reflections caused by reservoir structural elements and inhomogeneities of the antenna-waveguide path due to their mutual influence on the useful signal. An analysis of the description and patent claims allows us to conclude that the high-resolution method of the MUZIC type, or other methods that improve the frequency resolution, can increase the measurement accuracy at a distance near the interfering reflector in comparison with methods based on determining the position of the maximum spectral component, for example [3]. However, the main reason for the lack of accuracy in the presence of interfering reflections remains - this is the mutual influence of the spectra of the useful and interfering signals, leading to a distortion of the ratios of the intensities of the spectral components of the useful and interfering reflectors and, as a result, to confusing the useful and false reflectors when they are localized, which causes a large level measurement error , manifested in an abrupt change in the measurement results.

The purpose of the invention is to reduce the error of level measurement in the presence of interfering reflections from the structural elements of the tank.

This goal is achieved by the fact that in the level measurement method, including generating a probing radio frequency signal with a linear stepwise periodic frequency modulation with known values of the carrier frequency and frequency deviation, the radiation of the generated signal in the direction of the probed material, reception, after the propagation time of the reflected signal, mixing it with a part of the emitted signal, extracting the differential frequency signal to obtain samples from the output of the analog-to-digital converter, calculating e of the spectrum of this signal using a fast Fourier transform, a rough estimate of the distance to the material from a rough estimate of the frequency of the maximum spectral component, an accurate estimate of the distance to the material from an accurate estimate of the difference frequency corresponding to the maximum spectral component found by calculating the continuously discrete Fourier transform and varying the difference frequencies within one frequency discrete spectrum around a rough estimate of the frequency of the maximum spectral component in both directions s, the determination of the level in the zone of increased measurement error near the interfering reflector is performed by predicting the level from the level value determined in the previous measurement cycle, the rate of change of the level measured outside the zone with increased measurement error and the measured time interval between two adjacent measurement cycles, moreover, the prediction of the position of the level is made taking into account the direction of its change, determined by the value of the auxiliary control signal.

With this determination of the level, the influence of the interfering reflector on the measurement result is eliminated and the measurement error decreases sharply. The rate of change of level is determined in areas of level change, where the influence of the interfering reflector is not significant, according to the results of measuring the level change and the corresponding time interval between two adjacent measurements and averaging the results over several measurement cycles. When predicting the level inside the zone with increased measurement error and when assessing the rate of level change outside the zone with increased measurement error, the direction of the level change or the absence of any level change is determined by the value of the auxiliary control signal.

To determine the boundaries of zones with a high measurement error, preliminary training of the measuring device is performed, in which the position of all interfering reflectors exceeding a predetermined level in terms of the reflected signal is determined on an empty tank, and in the vicinity of each of them two boundaries of zones with a high error are indicated, differing on both sides from the distance to the corresponding interfering reflector by a triple value of the resolution element of the frequency-modulated signal with indicated deviation of frequency.

The inventive method of measuring the level of material in the tank has a combination of features not known from the prior art for methods of this purpose, which allows us to conclude that the criterion of "novelty".

To prove the inventive step, it is necessary to take into account that the known method of level measurement [4], based on the high resolution method of the MUZIC type, allows to increase the resolution and, accordingly, the measurement accuracy in a certain area near the interfering reflection in comparison with methods based on determining the position of the maximum spectral component , for example [3]. However, the main reason for the lack of accuracy in the presence of interfering reflections remains - this is the mutual influence of the spectra of the useful and interfering signals, causing an incorrect transmission of the ratios of the readings of the spectrum of the useful and interfering reflectors, and the presence of spurious amplitude modulation and non-linear distortions of the RMS. Due to the incorrect transmission of amplitude ratios in the known measurement method, the useful and interfering reflectors are mixed up when the useful reflector is localized by the maximum spectral component, resulting in a large distance measurement error in the form of individual error spikes. Moreover, the greater the noise level, the more this phenomenon manifests itself.

The claimed method does not have this drawback, since when it changes the level inside the zone with an increased measurement error, starting from any of its boundaries, the current level is predicted by the value of the level determined on the previous measurement cycle, the estimated rate of change of level, taking into account the direction of change and a measured time interval elapsed between two measurement cycles. Therefore, the measurement error is significantly reduced at all points of the zone with a high measurement error near the interfering reflector, including directly near the interfering reflector, where the known method leads to increased error. Moreover, the prediction results are not affected by the noise level and the ratio of the intensities of the useful and interfering reflections.

These differences do not follow explicitly from available scientific and technical sources, which allows us to conclude that the claimed technical solution meets the criteria of the invention - "Inventive step".

These differences lead to the appearance of a qualitatively new property of the claimed method - the ability to measure the level of the material in the presence of interfering reflections with an arbitrary ratio of signals reflected from the material and the interfering reflector. This new property improves the accuracy of measurements.

The implementation of the claimed method is illustrated using the drawings shown in figures 1-5.

Figure 1 shows a device for measuring the level in the presence of interfering reflectors in the tank.

Figure 2 shows a block diagram of a level prediction program taking into account the control signal in the level measurement mode.

Figure 3 shows a block diagram of a signal processing program when performing training.

Figure 4 shows the dependence of the error of distance measurement based on the Fourier transform and the proposed method, obtained by modeling the measurement process on a computer.

Figure 5 shows the dependence of the error of distance measurement based on the Fourier transform and a known method based on the high resolution method MUSIK obtained by simulating the measurement process on a computer.

The level measuring device comprises a signal shaper (Ф) 1, the output of which is connected to the input of the microwave amplifier (UHF) 2, a directional coupler (HO) 3, the output of the microwave amplifier 2 is connected to the input of HO 3, the circulator (C) 4, the input of which connected to the first output of the directional coupler 3, the antenna (A) 5 connected to the first output of the circulator 4, a mixer (Cm) 6, the inputs of which are connected to the second outputs of the directional coupler 3 and the circulator 4, and the output is connected through series-connected amplifier (Y) 7 and analog-to-digital the forming (ADC) 8 with the first input of the processor (PR) 9. The second output of generator 1 is connected to a second input of the processor 9, the first output of the processor 9 is connected to a second input of the ADC 8 and the second output of the processor is the output device. The third input of the processor 9 receives a control signal signaling the fact of a level change and its direction of change.

The method of measuring the level of material in the tank is as follows.

Shaper 1 generates a sounding radio frequency signal with a linear stepwise periodic frequency modulation. This signal, after amplification in the microwave amplifier 2, enters through the directional coupler 3 and circulator 4 into the antenna 5 and is emitted in the direction of the controlled material. The proposed method of level measurement uses frequency modulation according to a triangular law. The level measurement occurs when using the MFR obtained in the interval 0 ÷ T _mod / 2, where T _mod is the period of frequency modulation. After the propagation time, the reflected signal is received by the antenna 5 and from the second output of the circulator 4 is fed to the first input of the mixer 6. The second input of the mixer 6 receives a part of the emitted signal from the second output of the directional coupler 3. The RF from the output of the mixer through the amplifier 7 is fed to the input of the ADC 8. The second input of the ADC 8 receives control pulses from the first output of the processor 9. Using the ADC, digital samples of the RMS u (k) k = 0, ..., N-1 are obtained. The RMS samples are digitally fed to the first input of the processor 9. A packet of synchronizing pulses corresponding to half the modulation period from the second output of the driver 1 is received at the second input of the processor 9. An auxiliary control signal is supplied to the third input of the processor 9, which allows one to determine the direction of level movement or the absence of movement . The result of measuring the level of material in the tank is fed to the second output of the processor 9.

, n=0, …, N-1 этого сигнала, умноженного на отсчеты весовой функции W(k), k=0, …, N-1 (например, весовой функции Блэкмана):A block diagram of the processor in the mode of measuring the level of material shown in figure 2. In block 10, the necessary variables are initialized. This is setting the sign of the first measurement Perv_Vkl = 1 and zeroing the sign of entry into the zone of increased measurement error Pr izn = 0, entering from the reprogrammable ROM the number K of zones with increased measurement error, their boundaries R _{1, k} , R _{2, k} , k = 1, ..., K, which were found in the training mode on an empty tank, and the values of the pouring speed v _{n, k} and the discharge rate of the material v _{c, k} , k = 1, ..., K in each zone with an increased measurement error, determined before starting measurements on technical specifications of tank equipment. In block 11, the digital samples u (k) k = 0, ..., N-1 RMS are input during one half-period of modulation. The moments of input of the RPS samples are set by the synchronizing signal supplied to the second input of the processor 9 from the shaper 1. In addition, the time is fixed at which the RPS readout has ended. Next, in block 11, a rough estimate of the frequency of the superfine frequency is performed;

, n = 0, ..., N-1 of this signal multiplied by the samples of the weight function W (k), k = 0, ..., N-1 (for example, the Blackman weight function):

n=0, …, N-1,

n = 0, ..., N-1,

и соответствующее значение дискретной частоты ω_nмакс=4πn_макс/T_мод.find the reference number n _max spectral component with maximum amplitude

and the corresponding value of the discrete frequency ω _nmax = 4πn _max / T _mod .

Then make an accurate estimate of the difference frequency. To do this, calculate the continuously-discrete Fourier transform of the RMS, multiplied by the samples of the weight function:

По полученному значению разностной частоты ω_макс вычисляют текущую i-ю оценку расстояния

до материала:vary the value of the frequency ω in the range from (ω _imax -4π / T _modes ) to (ω _imax + 4π / T _modes ) with a step Δω setting the error in the estimation of the frequency, and find the position ω _{max of} the spectrum modulus

Based on the obtained value of the differential frequency ω _max calculate the current i-th distance estimate

before material:

соответствующую максимуму спектра.where c is the propagation velocity of the electromagnetic wave inside the tank; ΔF is the frequency deviation at FM, and the phase is calculated

corresponding to the maximum of the spectrum.

In block 12, the sign of the first inclusion of Perv_Vkl is checked. If it is equal to 1, which corresponds to the first measurement, then this feature is zeroed in block 13, the result is fixed in block 14 and the transition to a new measurement cycle, i.e. to block 11.

с границами зон с повышенной погрешностью измерения R_1,k, R_2,k, k=1, …, К, и проверяют, находится ли уровень материала в одной из зон с повышенной погрешностью измерения вблизи соответствующего мешающего отражателя:If it is determined in block 12 that the measurement is not the first, then a transition is made to block 15, where the time interval t _ism calculated from the moment of the previous measurement is calculated. Next, in block 16, the distance estimate obtained in block 11 is compared

with the boundaries of zones with increased measurement error R _{1, k} , R _{2, k} , k = 1, ..., K, and check whether the material level is in one of the zones with increased measurement error near the corresponding interfering reflector:

R _{1, k} ≤R _i ≤R _{2, k} .

If the material level is within one of the zones with a high measurement error, the transition to block 17, where the sign Pr izn of the first measurement in the zone with a high measurement error, is checked. If Pr izn = 0, which corresponds to the first entrance to one of the zones with increased measurement error, then in block 18 the sign of the first measurement is inverted Pr izn = 1, the corresponding drain or fill rate is written to the reprogrammable ROM and go to block 19. If the measurement in the area of increased error is not the first, then such a recording is not performed, but proceeds immediately to block 19. In block 19, it is determined whether the material is drained (i.e., level reduction) by checking the value of the auxiliary control signal. If draining is performed, then in block 20 the result is predicted:

и

- соответственно текущее и предыдущее значение измеренного расстояния, и переход к блоку 14 для фиксации результата.Where

and

- respectively, the current and previous value of the measured distance, and the transition to block 14 to fix the result.

If no discharge is made, then in block 21 it is checked whether the filling is performed. If loading is done, then the result is predicted:

and proceeding to block 14 to record the result.

If loading is not performed, then from block 21 immediately proceeds to block 23 to fix the previous result.

Then, it returns to block 11 to enter a new array of RPS samples, etc. the level measurement procedure is repeated cyclically.

для k-й зоны с повышенной погрешностью измерения:If it is determined in block 16 that the level is not in the zone with an increased measurement error, or it has left it, then in block 24 the characteristic of the first measurement is reset to zero Pr izn = 0 and go to block 25, where it is checked whether the drain is performed. If draining is performed, then in block 26, an estimate of the drain rate

for the k-th zone with increased measurement error:

- накапливаемая статистика на i-м цикле измерений,

, М - количество накопленных значений оценки соответствующей скорости.Where

- accumulated statistics on the i-th measurement cycle,

, M is the number of accumulated values for evaluating the corresponding speed.

Next, go to block 14 to fix the measurement result, and then return to block 11 to repeat the measurement cycle.

для k-й зоны с повышенной погрешностью измерения:If no discharge is made, then in block 27 it is checked whether the filling is performed. If loading is done, then in block 28, the estimation of the loading speed

for the k-th zone with increased measurement error:

- накапливаемая статистика на i-м цикле измерений,

М - количество накопленных значений оценки соответствующей скорости. Далее производится переход к блоку 14 для фиксации результата измерения, и затем возврат к блоку 11 для повторения цикла измерения.Where

- accumulated statistics on the i-th measurement cycle,

M is the number of accumulated values for evaluating the corresponding speed. Next, go to block 14 to fix the measurement result, and then return to block 11 to repeat the measurement cycle.

If it is determined in block 27 that no loading is performed, then there is a transition to block 14 and then to block 11.

The learning process on an empty tank is carried out according to the block diagram of the program shown in figure 3.

, n=0, …, N -1 сигнала СРЧ с помощью БПФ. В блоке 31 в области первых четырех дискретных спектральных составляющих спектра СРЧ осуществляется поиск локального максимума S_ант, соответствующего отражению от кромки антенны и вычисление порогового значения:In block 29, the RMS is recorded on an empty tank; in block 30, the spectrum is calculated

, n = 0, ..., N -1 of the RF system using the FFT. In block 31, in the region of the first four discrete spectral components of the RHF spectrum, a local maximum S _ant corresponding to the reflection from the antenna edge is searched and a threshold value is calculated:

S _then = aS _ant

where a is the level of the threshold value with respect to the level of reflection from the edge of the antenna (for example, a = -20 dB).

, n=5, …, N-1, превысивших пороговое значение S_пор, определение их количества К, номеров отсчетов максимумов n_макс,k k=1, …, K и расчет дискретных частот ω_nмакс,k=4πn_макс,k/T_мод, k=1, …, K, соответствующих этим максимумам.Block 32 searches for spectrum maxima

, n = 5, ..., N-1, which exceeded the threshold value of S _then , determining their number K, maximum number counts n _{max, k} k = 1, ..., K and calculating discrete frequencies ω _{n max, k} = 4πn _{max, k} / T _modes , k = 1, ..., K, corresponding to these maxima.

In block 33, the discrete frequencies found are refined using the continuously discrete Fourier transform, similar to the processor’s operation in the level measurement mode, and the distances R _{mesh, k} , k = 1, ..., K, corresponding to interfering reflectors, are calculated.

In block 34, K pairs of boundaries R _{1, k} , R _{2, k} , k = 1, ..., K of zones with an increased measurement error are determined as the difference and the sum of the distances to each of the interfering reflectors and the triple value of the resolution element for the used signal with World Cup:

In block 35, the obtained values of the boundaries of the zones with increased measurement error are recorded in the reprogrammable ROM and exit the learning mode.

In the proposed method, when the level inside the zone changes with an increased measurement error near the interfering reflector, a constant prediction of the result takes into account the fact of the presence and direction of the level change and using the estimate of the rate of level change obtained immediately before entering this zone. This fact can significantly reduce the error in measuring the level of the material.

Modeling the level measurement process showed the high efficiency of the proposed method for measuring the level of material in the tank. So, figure 4 shows the dependence of the measurement error of the distance inside the zone with increased measurement error near the interfering reflector for two measurement methods. In figure 4, the arrow indicates the position of the interfering reflector. A thin solid line shows the error for a method based on the Fourier transform. The thick solid line shows the error for the proposed method. All graphs were obtained under the same measurement conditions for a carrier frequency of 10 GHz, a frequency deviation of 1 GHz, a noise level of 50 dB, and the formation of 1024 samples of a modeled RPS.

In figure 5, the arrow still indicates the position of the interfering reflector. A thin solid line shows the error for a method based on the Fourier transform. The thick solid line shows the error for the known method. At a certain distance from the interference, the known method really allows to significantly reduce the measurement error. However, there is a zone near the interfering reflector in which localization of the useful reflector by the maximum spectral component in the MUSIK method leads to a confusion of the useful and interfering reflectors due to incorrect transmission of the ratio of the intensities of the spectral components of these reflectors in the calculation of the spectrum. As a result, large sharp throws of the distance measurement error are observed, significantly exceeding the error provided by the measurement method based on the Fourier transform.

A comparison of FIGS. 4 and 5 shows that the proposed measurement method provides a significant reduction in the measurement error near the interfering reflector in comparison with known methods.

INFORMATION SOURCES

1. Theoretical foundations of radar. Ed. Shirmana Y.D. M .: Sov. Radio, 1970.

2. Marfin V.P., Kuznetsov F.V. Microwave level gauge. // Devices and control systems. 1979, No. 11. S.28-29.

3. RF patent No. 2234717, G01S 13/34, 03/04/2003.

4. US patent 5504430. MKI G01S 13/08.

5. Tikhonov V.I., Kharisov V.N. Statistical analysis and synthesis of radio engineering devices and systems. - M .: Radio and communications, 1991 .-- 608 p.

Claims (10)

1. The method of measuring the level of material in the tank, including the formation of a sounding RF signal with a linear stepwise periodic frequency modulation with known values of the carrier frequency and frequency deviation, the radiation of the generated signal in the direction of the probed material, reception, after the propagation time, of the reflected signal, mixing it with part radiated signal, extracting the differential frequency signal to obtain samples from the output of the analog-to-digital converter, calculating the spectrum of this with ignal multiplied by readings of the weight function (for example, the Blackman function) using the fast Fourier transform, a rough estimate of the distance to the material from a rough estimate of the frequency of the maximum spectral component, an accurate estimate of the distance to the material from an accurate estimate of the difference frequency, corresponding to the maximum spectral component found by calculating the continuous-discrete Fourier transform and varying the difference frequency within one frequency discreteness of the spectrum around a rough estimate of the frequency the maximum spectral component in both directions, characterized in that in the zone with increased measurement error near the interfering reflector, the level is predicted by the level value determined in the previous measurement cycle, the rate of change of the level measured outside the zone with increased measurement error and the measured time interval elapsed between two adjacent measurement cycles, and the level position is predicted taking into account the direction of its change, determined by the value of the auxiliary control signal. 2. The method according to claim 1, characterized in that in the absence of level displacement determined by the value of the auxiliary control signal, the result of the previous measurement is taken as the level value. 3. The method according to claim 1, characterized in that the determination of the rate of change of level is carried out outside the zone with increased measurement error based on the current and previous measured position of the level and the time interval between these two measurements, taking into account the direction of the level change. 4. The method according to claim 1, characterized in that the determination of the rate of change of level is made according to several previous measurement results using averaging. 5. The method according to claim 1, characterized in that during the first measurement, the determination of the rate of change of level is not performed. 6. The method according to claim 1, characterized in that the rate of change of the level of the material during filling and discharge in different zones with an increased measurement error is set in advance before measurements on the design features of the tank. 7. The method according to claim 1, characterized in that during the first measurement inside the zone with increased measurement error, the corresponding pouring or discharge rate is corrected based on the measurement result outside this zone. 8. The method according to claim 1, characterized in that the boundaries of the zones with increased measurement error are found using preliminary training on an empty tank. 9. The method according to claim 8, characterized in that the preliminary training is performed by recording the difference frequency signal on an empty tank, calculating the spectrum of the difference frequency signal using a fast Fourier transform, finding frequencies corresponding to all local spectrum maxima that exceed a predetermined threshold value, refinement the frequencies of local extrema found using the continuous-discrete Fourier transform, calculating the distances to the interfering reflectors from the found frequencies and calculating the near and far the boundaries of zones with increased error as, respectively, the difference or the sum of the calculated distances to the interfering reflectors and the triple value of the resolution element of the used signal with frequency modulation. 10. The method according to claim 9, characterized in that the threshold value of the spectrum during training is obtained as a given fraction, for example, 20 dB, from the local maximum of the spectrum of the difference frequency signal corresponding to reflection from the edge of the antenna.

Images