DE102022128063A1 - Transmission of spatially resolved level measurements - Google Patents

Abstract

The invention relates to data compression of three-dimensional filling material profiles or the compression of those location-related filling level values (Lx;y) that form the basis of the corresponding filling material profile and are recorded by a three-dimensional, location-resolving filling level measuring device (1). For this purpose, an evaluation unit of the filling level measuring device (1) creates a geometric model (MG) in which the location-related filling level values (Lx;y) are approximated by at least one geometric body (z1, k1, k2, B), for example on the basis of artificial intelligence. The advantage of this inventive idea is that geometric bodies (z1, k1, k2, B) can potentially be described with a significantly smaller amount of data than the corresponding filling level values (Lx;y) of the corresponding partial area on the filling material surface. This enables the level measuring device (1) to transmit the geometric model (MG) to external units, such as a process control center, via an internal communication unit, even if the corresponding transmission protocol only allows a very limited data rate, for example due to explosion protection regulations.

Description

The invention relates to a spatially resolving level measuring device and a method for the efficient data transmission of the corresponding level values.

In process automation technology, appropriate field devices are used to record relevant process parameters. In order to record the process parameters, suitable measuring principles are implemented in the respective field devices with which the corresponding process parameters, such as level, flow, pressure, temperature, pH value, redox potential or conductivity, can be recorded. A wide variety of such field device types are manufactured and sold by the “Endress + Hauser” group of companies.

Contactless measuring methods have been established for measuring the fill level of filling goods in containers because they are robust and low-maintenance. In the context of the invention, the term "container" also includes non-enclosed containers, such as basins, lakes or flowing water. Another advantage of contactless measuring methods is the ability to measure the fill level almost continuously. In the field of continuous fill level measurement, radar-based measuring methods are therefore mainly used (in the context of this patent application, the term "radar" refers to signals or electromagnetic waves with frequencies between 0.03 GHz and 300 GHz).

An established measuring principle is FMCW (“Frequency Modulated Continuous Wave”). The measuring principle of FMCW radar-based distance measuring methods is based on transmitting a continuous radar signal with a modulated frequency. A characteristic of FMCW is that the transmission frequency is periodically changed within a defined frequency band. Taking regulatory requirements into account, higher frequency bands in the range of a standardized center frequency are increasingly being used as development progresses: In addition to the 6 GHz band, the 26 GHz band or the 79 GHz band, frequencies of over 100 GHz have now been implemented. The advantage of high frequencies is that a larger absolute bandwidth (e.g. 4 GHz in the 100 GHz frequency band) can be used at higher frequencies. This in turn achieves a higher resolution and a higher accuracy of the level measurement.

The temporal change of the frequency within the frequency band is linear by default and has a sawtooth or triangular shape. A sinusoidal change can also be implemented in principle. The distance is determined in the FMCW method based on the current frequency difference between the currently received high-frequency signal after reflection on the measuring object and the radar signal currently emitted by the measuring device. The FMCW-based level measurement method is described, for example, in the published patent application EN 10 2013 108 490 A1 .

The FMCW method makes it possible to measure the distance or the fill level at least at certain points. The point at which the fill level is measured depends on the orientation of the transmitting/receiving antenna or the direction of its beam lobe (due to the generally reciprocal properties of antennas, the characteristics or beam angle of the beam lobe of the respective antenna are independent of whether it is transmitting or receiving; the term "angle" or "beam angle" in the context of this patent application refers to the angle at which the beam lobe has its maximum transmission intensity or reception sensitivity).

In the case of liquid filling materials with a homogeneous filling level, a point-based filling level measurement is sufficient. In these cases, the filling level measuring device is aligned so that the beam of the antenna is directed approximately vertically downwards towards the filling material and determines the distance to the filling material. However, in the case of solid filling materials such as gravel or grain, the filling level can be inhomogeneous due to bulk material cones, for example, so that the filling level value determined by the filling level measuring device is only partially meaningful. In such cases in particular, it is therefore desirable to be able to determine the distance or filling level with spatial resolution in the form of a two- or three-dimensional profile. In addition to the exact volume estimate, the visual 3D representation of imaging filling level measuring devices in particular offers great benefits for the automation of filling and dismantling processes. In addition, the visualization can be used to detect and avoid dangerous filling conditions, which can increase the reliability and safety of the corresponding process plants.

For spatially resolving level measurement, the beam of the radar-based level measuring device can be designed to be mechanically pivotable so that the filling material profile can be measured across the entire container cross-section. section or at least a partial area of the filling material surface can be detected. Due to the increased maintenance effort, such designs are only used in special applications, such as in mining.

ein.Radar-based distance measuring devices in which the beam can be electrically swiveled are also known from the state of the art: Among other things, the so-called "phased array" principle can be used, in which the measuring device comprises several antennas, with their radar signals being superimposed for evaluation purposes. The antennas are arranged in rows (beam swiveled along one axis) or in an array (beam swiveled around two axes). In order to emit or receive the high-frequency signal at a defined angle, the individual antennas are controlled according to their arrangement sequence with a phase shift that increases per antenna. The angle α of the beam is adjusted depending on the phase shift φ according to $$\mathrm{\alpha }~\text{arcsine}\left(\mathrm{\phi }\right)$$

a.

According to the current state of the art, the hardware required for this can be integrated so compactly that the antennas are housed as patch antennas together with the semiconductor component for signal generation/signal evaluation on a common circuit board or even as a jointly encapsulated radar IC (“integrated circuit”). A distance measuring device operating according to the phased array principle is described, among other things, in the German publication EN 100 36 131 A1 described.

In addition to the phased array principle, spatially resolving radar measuring devices can alternatively be designed on the basis of digital beam forming. In this case, each antenna in the antenna array has its own signal processing and its own digitization. The received radar signal is digitized in terms of both its amplitude and its phase position using an appropriate process. The summation takes place digitally after a virtual phase shift and amplitude scaling in a special computer, the so-called beamform processor. With digital beam forming, the radiation characteristics of the antenna can be shaped so that it has several independent main lobes for different directions. Both digital beam forming and the phased array principle can potentially achieve a high lateral and vertical spatial resolution of the level measurement.

Due to the high resolution, the amount of data recorded for location-based level values is correspondingly high. However, a large amount of data is particularly problematic for level measuring device types that are subject to increased explosion protection requirements. In these, the power transfer and thus the maximum data transfer rate is severely limited. Nevertheless, even under such operating conditions, it is often important to maintain as up-to-date an overview of the location-resolved level as possible. It is therefore only possible to a limited extent to minimize the rate at which the filling material profile is updated or newly recorded. If the amount of data per updated filling material profile is thinned out too much, this can in turn lead to the filling material profile grid no longer being fine enough.

It is therefore an object of the invention to provide a three-dimensionally resolving level measuring device which meets these conflicting design requirements as best as possible.

• - Eine Sende-/Empfangs-Einheit, die ausgelegt ist,
  • ◯ um Radar-Signale mit lateral veränderbarem Ortsbezug gen Füllgut auszusenden, und
  • ◯ um nach Reflektion an einem entsprechenden Ort der Füllgut-Oberfläche entsprechende-Empfangs-Signale zu empfangen, und
• - eine Auswerte-Einheit, die ausgelegt ist,
  • ◯ anhand der Empfangs-Signale ortsbezogene Füllstandswerte zu ermitteln, und
  • ◯ um das geometrische Modell zu erstellen, indem die ortsbezogenen Füllstandswerte durch einen oder mehrere geometrische Körper approximiert werden.

The invention solves this problem by means of a level measuring device for determining a geometric model of at least a partial area of the surface of a filling material. For this purpose, the level measuring device comprises at least the following components:

• - A transmitting/receiving unit designed
  • ◯ to transmit radar signals with laterally variable location reference to the filling material, and
  • ◯ to receive corresponding reception signals after reflection at a corresponding location on the product surface, and
• - an evaluation unit designed
  • ◯ to determine location-related fill level values based on the received signals, and
  • ◯ to create the geometric model by approximating the location-related level values using one or more geometric bodies.

The invention is therefore based on the idea of approximately representing at least the relevant sub-areas of the filling material surface, which are represented by the location-related filling level values, using geometric bodies. The advantage of this is that geometric bodies can potentially be described with a significantly smaller amount of data than the corresponding filling level values of the corresponding sub-area on the filling material surface. This enables the filling level measuring device to transmit the geometric model to external units, such as the process control center, via a communication unit, even if the corresponding transmission protocol only allows a very limited data rate. If designed accordingly, the communication unit can set the update rate with which it transmits the current model to the external unit, in particular inversely proportional to the approximation duration, so that the process control center always has the most current model possible.

In the context of the invention, the term "unit" is understood to mean in principle any separate arrangement or encapsulation of those electronic circuits that are intended for the specific purpose, e.g. for measuring signal processing or as an interface. Depending on the purpose, the respective unit can therefore comprise corresponding analog circuits for generating or processing analog signals. However, the unit can also comprise digital circuits such as FPGAs, microcontrollers or storage media in conjunction with corresponding programs. The program is designed to carry out the necessary process steps or to apply the necessary computing operations. In this context, different electronic circuits of the respective unit within the meaning of the invention can potentially also access a common physical memory or be operated using the same physical digital circuit. It is not relevant whether different electronic circuits within the unit are arranged on a common circuit board or on several interconnected circuit boards.

The approximation algorithm for the geometric model can be implemented in the evaluation unit, for example, using the least squares method (better known in English as “least square”) or a machine learning algorithm, in particular an artificial neural network. The target variable can be defined, for example, as the factor by which the amount of data is reduced. The target variable can be, for example, to select the number, shape and/or size of the geometric bodies in such a way that the amount of data in the model is compressed by a factor of 10 ² , 10 ³ or 10 ⁴ in relation to the location-related fill level values. The approximation algorithm can be left to decide which types of geometric bodies are used for the approximation. However, it is also conceivable to specify the type(s) of geometric body to the approximation algorithm. This could be advantageous, for example, if the shape of the filling material surface is already roughly known due to the physical properties of the filling material of the specific process application, such as the grain size. Irrespective of this, one or more cuboids, cones, cylinders, ogives and/or sphere sections can be used as geometric bodies to approximate the model. The geometric bodies are preferably linked to create the model using logical and/or arithmetic operations, in particular union, intersection and/or difference, if several geometric bodies are used for the approximation.

If the evaluation unit knows the geometry of the lateral inner cross-section of the specific container and if the geometric model is to be created for the entire filling material surface, the evaluation unit can optionally include the extruded inner cross-section shape to approximate the geometric model. This can improve the computational effort required to create the model or the accuracy of the model. This variant is not only applicable to containers whose inner cross-section is constant over the height. If the inner cross-section shape or area depends on the height, it is possible to include the geometry of the lateral inner cross-section in the model, provided, for example, that the dependency of the inner cross-section shape/area on the height is known.

Within the scope of the invention, it is not relevant which principle for spatially resolving radar measurement is implemented in the level measuring device. The transmitting/receiving unit can, for example, be designed to transmit the radar signals using the phased array principle or the digital beam forming principle with a laterally variable location reference and to receive reception signals accordingly. Corresponding to the principle implemented in the transmitting/receiving unit, the evaluation unit must also be designed to determine the location-related level values using the respective principle.

• - Aussenden von Radar-Signalen gen Füllgut mit jeweils lateral verändertem Ortsbezug,
• - Empfang der entsprechenden Radar-Signale nach Reflektion an den verschiedenen Orten auf der Füllgut-Oberfläche,
• - Ermittlung ortsbezogener Füllstandswerte anhand der Empfangs-Signale, und
• - Erstellung des geometrischen Modells, indem die ortsbezogenen Füllstandswerte durch zumindest einen geometrischen Körper approximiert werden.

Corresponding to the previously described level measuring device, the invention also includes the following method for creating a geometric model of at least a partial area of the surface of a filling material. The following method steps are provided for this purpose:

• - Transmission of radar signals towards the filling material with laterally changed location reference,
• - Reception of the corresponding radar signals after reflection at the various locations on the product surface,
• - Determination of location-related level values based on the received signals, and
• - Creation of the geometric model by approximating the location-related fill level values by at least one geometric body.

• 1: Ein ortsauflösendes Füllstandsmessgerät an einem Behälter,
• 2: eine erfindungsgemäße Modellierung der Füllgut-Oberfläche.

The invention is explained in more detail using the following figures. It shows:

• 1 : A spatially resolving level measuring device on a container,
• 2 : an inventive modelling of the filling material surface.

To understand the invention, 1 a container 3 with a filling material 2 is shown, the filling level L of which is to be determined. Depending on the type of filling material 2 and the area of application, the container 3 can be up to more than 100 m high.

In order to be able to determine the fill level L, a radar-based fill level measuring device 1 is mounted above the filling material 2 at a known installation height h above the brine of the container 3. The fill level measuring device 1 is attached to a corresponding opening of the container 3 in such a way that radar signals S _HF , R _HF can be transmitted vertically downwards into the container 3 towards the filling material 2 via an antenna arrangement or can be received after they are reflected on the filling material surface. Accordingly, the fill level measuring device 1 can be arranged essentially outside the container 3.

entsprechend proportional zum Abstand d zwischen dem Füllstandsmessgerät 1 und dem Füllgut 2. Bei „c“ handelt es sich in diesem Zusammenhang um die jeweils medienabhängige Radar-Ausbreitungsgeschwindigkeit. Zur Bestimmung der Signallaufzeit t kann im Füllstandsmessgerät 1 das FMCW- oder das Pulslaufzeit-Verfahren implementiert sein. Dementsprechend hat die Erzeugung der auszusendenden Radar-Signale S_HF sowie der Empfang der entsprechenden Empfangs-Signale R_HF innerhalb des Füllstandsmessgerätes 1 durch eine entsprechend ausgelegte Sende-/Empfangs-Einheit zu erfolgen. Im Falle des FMCW-Verfahrens kann die Sende-/Empfangs-Einheit beispielsweise auf Basis einer Phasenregelschleife („Phase Locked Loop“) ausgelegt werden. Im Falle des Pulslaufzeit-Verfahrens kann die Sende-/Empfangs-Einheit auf dem Prinzip der Puls-Unterabtastung basieren.After reflection of the emitted radar signals S _HF on the surface of the filling material, the level measuring device 1 receives the reflected radar signals R _HF . The resulting signal runtime t between emission and reception of the respective radar signal S _HF , R _HF is according to $$t=\frac{2\ast d}{c}$$

accordingly proportional to the distance d between the level measuring device 1 and the filling material 2. In this context, “c” is the media-dependent radar propagation speed. To determine the signal propagation time t, the FMCW or pulse propagation time method can be implemented in the level measuring device 1. Accordingly, the generation of the radar signals S _HF to be transmitted and the reception of the corresponding reception signals R _HF within the level measuring device 1 must take place by an appropriately designed transmitting/receiving unit. In the case of the FMCW method, the transmitting/receiving unit can be designed on the basis of a phase locked loop, for example. In the case of the pulse propagation time method, the transmitting/receiving unit can be based on the principle of pulse subsampling.

den Füllstand L zumindest punktuell bestimmen, sofern die Einbauhöhe h im Füllstandsmessgerät 1 hinterlegt wird.For example, after a corresponding calibration, the level measuring device 1 can assign the measured signal transit time t to the respective distance d. The level measuring device 1 can use this to $$d=H-L$$

determine the fill level L at least at certain points, provided that the installation height h is stored in the level measuring device 1.

As a rule, the level measuring device 1 is connected to a higher-level unit 4, such as a local process control system or a decentralized server system, via its own communication unit, in which, for example, “4-20 mA”, “PROFIBUS”, “HART”, or “Ethernet” is implemented as a communication protocol. The measured level values L can be transmitted via this, for example to control any inflows or outflows of the container 3. The communication unit can be used to However, other information about the general operating status of the level measuring device 1 can also be communicated.

In the 1 In the example shown, the surface of the filling material 2 is not planar. This can occur in particular with bulk-type filling materials 2, for example when material cones form when the container 3 is filled. In addition, when the filling material 2 is pumped out, funnels can form on the surface of the filling material. If the level measuring device 1 only determines the filling level L at one point on the surface of the filling material 2, this may lead to an incorrect interpretation of the filling level L. This can lead to an emptying process being stopped incorrectly if the level measuring device 1 has determined that the container 3 is empty, even though there is still filling material 2 at the edge of the inside of the container. In the opposite case, if the container 3 is full, a filling process may not be stopped even though a maximum filling level has already been exceeded at one point on the filling material surface, because this is not detected by the level measuring device 1.

For this reason, the 1 The level measuring device 1 shown measures the level L _x;y in relation to the horizontal plane x;y with spatial resolution. For this purpose, the phased array principle or the digital beam forming principle can be implemented in the transmitting/receiving unit or in the evaluation unit, for example, so that level values L _x;y are transmitted with a laterally changed location reference x;y and received signals R _HF are received accordingly. In order to create a filling material profile, it is advantageous to transmit or receive the individual radar signals S _HF in relation to the vertical with a different location reference x;y in such a way that a regular grid is formed. The evaluation unit can thus determine level values L _x;y for this grid based on the corresponding received signals R _HF . The grid ideally extends laterally over the entire interior of the container 3. Accordingly, the transmitting/receiving unit must be able to swivel the beam over a correspondingly wide solid angle range with regard to the phased array principle or the digital beam forming principle. It is advantageous if the level measuring device 1 is designed to record a grid or a solid angle range of at least 90° with a resolution of at least 2° per axis. This results in a correspondingly large data set of spatially resolved level values L _x;y , which is to be exported via the communication unit.

Three-dimensional product surface profiles based on the determined fill level values L _x;y can be described numerically in the form of networks or volume cells, for example. For a sufficiently accurate description, a corresponding number of support values and thus very large amounts of data are required. For example, the description of the container contents measured with a spatially resolving radar measuring device with 180 azimuth, 180 elevation and 500 distance cells as volume cells already results in 16 Mbit of data. Even with a data reduction of over 90%, the resulting data volume for narrowband radio networks and "HART" in "4-20 mA" operation cannot be transmitted in periods of a few minutes.

In particular, if the communication unit of the level measuring device 1 is based on the 4-20 mA transmission standard or its data transmission rate is otherwise limited due to explosion protection regulations, a delay-free transmission of the complete data set of the rasterized level values L _x;y is not possible depending on the size of the data set, although this may be necessary to control pumps or drains.

Therefore, according to the invention, the evaluation unit creates a geometric model M _G of preferably the entire filling material surface in the container 3, wherein the model M _G is based on one or more geometric bodies z1, k1, k2, B. In principle, it is not prescribed which type of geometric bodies z1, k1, k2, B is used to approximate the filling material surface. Simple geometric bodies such as cubes, cylinders or cones are conceivable. However, more complex bodies such as spherical segments or ogives are also conceivable. In a pragmatic manner, the geometric model M _G can initially be created by the evaluation unit in such a way that the base point of the model M _G is defined at the location of the level measuring device 1. A subsequent change to the base point by the user, so that the base point is defined at a location within the container, for example, can be achieved by appropriate coordinate translation.

The inventive creation of the geometric model M _G is visualized schematically in 2 : In the variant shown there, the model M _G is based on a base body B, which results from the cross-sectional geometry of the container interior extruded towards the filling material surface. In this case, it is necessary to calculate the cross-sectional geometry of the respective container 3 in the evaluation unit of the level measuring device 1. As in 2 As shown, the design variant there is a container 3 whose interior has a circular cross-sectional geometry over the entire height h.

To model the filling material surface, the base body B is 2 The embodiment shown is modified by three geometric bodies z, k1, k2. The modification takes place in the form of logical or arithmetic operations: A first cone k1 is subtracted from a cylinder z and a second cone k2 is added vertically offset by addition. The resulting partial body is cut with the extruded base body B, resulting in the geometric model M _G of the filling material surface. As shown in 2 As shown in the model M _G , the filling level L _x;y of the filling material 2 can be reproduced in a spatially resolved manner in that, for example, a bulk material cone and, at the same time, a depression funnel are formed in the container 3 by the flowing bulk material 2. It goes without saying that the 2 The variant shown is only a highly simplified example.

The advantage of this modeling of the filling material surface according to the invention is that, depending on the type, number and size of the geometric bodies z1, k1, k2 ,B on which the resulting model M _G is based, the model M _G can be represented in relation to the location-related fill level values L _x;y using a significantly reduced amount of data. This data compression is illustrated by the fact that, for example, a cone k1, k2 as a geometric body z1, k1, k2 ,B can only be represented mathematically by its height, its opening angle, its position and, if applicable, its spatial orientation. The amount of data required to mathematically describe these parameters is, depending on the corresponding area on the filling material surface, significantly smaller than the amount of data on spatially resolved fill level values L _x;y from the partial area of the filling material surface covered by the cone k1, k2.
In relation to the 2 The variant shown requires the transfer of the concrete model M _G according to # geom. solid Type parameter position 1 cylinder R, H x, y, h 2 cone R, H x, y, h 3 Elliptical cone R1, R2, H x, y, h 4 cylinder R, H x, y, h

2 bytes of data per parameter value and one byte for the object type. Less than 64 bytes are sufficient for the links.

The example makes it clear that the data reduction factor depends, among other things, on how closely the geometric model M _G is approximated to the underlying grid of fill level values L _x;y . In this case, either classic approximation algorithms, such as the least squares method, or machine or self-learning algorithms, such as an artificial neural network, can be implemented in the evaluation unit in order to approximate the geometric model M _G based on the fill level values L _x;y . In this context, it is conceivable to implement the approximation algorithm in the evaluation unit in such a way that the resulting geometric model M _G comprises a defined maximum amount of data, or so that the amount of data of the model M _G is compressed in relation to the underlying fill level values L _x;y by a minimum factor of, for example, 10 ² . This enables the communication unit to transmit the approximated model M _G of the filling material surface to the higher-level unit 4 even in the case of a limited data rate with a high update rate of e.g. 10 Hz.

The approximation time that the evaluation unit needs to approximate the current model M _G depends, among other things, on whether the model M _G needs to be significantly adjusted in relation to the previously generated model M _G. This in turn depends on whether the fill level L _x;y or the topology has changed significantly. If a machine learning algorithm is implemented in the evaluation unit for generating the geometric model M _G , the respective approximation time also depends on the current learning progress of the learning algorithm. Accordingly, the communication unit can be designed flexibly if necessary in that it independently sets the update rate with which the current model M _G is transmitted to the higher-level unit 4 depending on the approximation time.

If the approximation duration decreases, for example due to increasing learning progress, the communication unit can increase the update rate inversely proportional to the approximation duration, at least until the maximum possible data transfer rate is reached. In the other case, i.e. if the approximation duration increases despite any learning progress, this in turn indicates a major change in the filling material topology. In this scenario, the communication unit can, if designed accordingly, issue a warning signal to the process control center 4 if the approximation duration exceeds a defined maximum value, or if the approximation duration has increased by a minimum value in relation to a previously approximated model M _G. This can be used to warn of irregular conditions in the container 3, such as a flank slipping or a cornice breaking off.

List of reference symbols

1

Level measuring device

2

Filling material

3

container

4

Superior unit

11

Collecting lens

B

Internal cross-sectional shape of the container

d

distance

H

Height of the geometric body

H

Installation height

k1,2

cone

Lx;y

Level value

MG

Geometric model

R

radius

RHF, SHF

(Reflected) radar signal

x;y

Location coordinates

z

cylinder

QUOTES INCLUDED IN THE DESCRIPTION

This list of documents listed by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions.

Cited patent literature

• DE 102013108490 A1 [0005]
• DE 10036131 A1 [0010]

Claims (11)

Fill level measuring device for determining a geometric model (M _G ) of at least a section of the surface of a filling material (2), comprising: - a transmitting/receiving unit which is designed to ◯ emit radar signals (S _HF ) with a laterally variable location reference (x; y) towards the filling material (2), ◯ to receive received signals (R _HF ) corresponding to the filling material surface after reflection at a corresponding location (x; y), and - an evaluation unit which is designed to ◯ determine location-related fill level values (L _{x; y} ) based on the received signals (R _HF ), and ◯ to create the geometric model (M _G ) by approximating the location-related fill level values (L _{x; y} ) by at least one geometric body (z1, k1, k2 ,B). Level measuring device according to Claim 1 , comprising: - a communication unit by means of which the geometric model (M _G ) can be transmitted to an external unit (4). Level measuring device according to Claim 2 , wherein the communication unit is designed to set an update rate with which the respective current model M _G is transmitted to the external unit 4, in particular antiproportionally depending on an approximation duration. Level measuring device according to Claim 2 or 3 , wherein the communication unit is designed to output a warning signal if the approximation duration exceeds a defined maximum value and/or if the approximation duration increases by a minimum value in relation to a previously approximated model M _G. Level measuring device according to one of the preceding claims, wherein the transmitting/receiving unit is designed to transmit radar signals (S _HF ) according to the phased array principle or by means of the digital beam forming principle with a laterally variable location reference (x; y) and to receive reception signals (R _HF ) accordingly, and wherein the evaluation unit is designed to determine the location-related level values (L _{x; y} ) according to the corresponding principle. Method for creating a geometric model (M _G ) of at least a section of the surface of a filling material (2) by means of the level measuring device (1) according to one of the preceding claims, comprising the following method steps: - transmitting radar signals (S _HF ) towards the filling material (2), each with a laterally changed location reference (x; y), - receiving the corresponding radar signals (R _HF ) after reflection at the various locations (x; y) on the filling material surface, - determining location-related filling level values (L _{x; y} ) based on the received signals (R _HF ), and - creating the geometric model (M _G ) by approximating the location-related filling level values (L _{x; y} ) by at least one geometric body (z1, k1, k2, B). Procedure according to Claim 6 , wherein the number, shape and/or size of the geometric bodies (z1, k1, k2 ,B) are selected such that the data volume of the model (M _G ) is compressed by at least a factor of 10 ² in relation to the location-related fill level values (L _x;y ). Procedure according to Claim 6 or 7 , where one or more cuboids, cones (k1, k2), cylinders (z), ogives and/or sphere sections are used as geometric bodies to approximate the model (M _G ). Method according to one of the Claims 6 until 8th , wherein the filling material (2) is stored in a container (3) with a known internal cross-sectional shape (B), and wherein the evaluation unit is designed to include the extruded internal cross-sectional shape (B) for approximating the model (M _G ). Method according to one of the Claims 6 until 9 , wherein the geometric model (M _G ) is approximated by means of the least squares method or by means of a machine learning algorithm, in particular an artificial neural network. Procedure according to Claim 6 until 10 , wherein the geometric bodies (z1, k1, k2 ,B) for creating the model (M _G ) are formed according to logical and/or arithmetic operations, in particular union, intersection and/or difference, provided that several geometric bodies (z1, k1, k2 ,B) are used for approximation.

Images