Method for calibrating ABG filter to improve position signal of electric motor

Abstract

The invention relates to a method for calibrating an ABG filter for improving a position signal of an electric motor, the electric motor being a component of a machine, the method comprising the steps of:-creating a digital model (S10) of the electric motor, the model outputting values of a position x, a speed x'and an acceleration x '' of the electric motor, respectively; -specifying initial values of the factors alpha, beta and gamma (S12); creating a first state of the electric motor (S14); -starting the electric motor (S16) and determining a second state of the electric motor; -determining a modeled state (S18); -comparing (S20) the second state with the modeled state; -adjusting the factors alpha, beta and gamma (S22).

Description

Technical Field

[0001] The present invention relates to optimization of a position signal from an electric motor, in particular to a position signal of a resolver in the electric motor. Background Art

[0002] The rotor position of an electric motor can be determined in various ways, depending on the type of motor and the requirements of the application.

[0003] Resolvers are sensors that can detect the position of the rotor in electrical machines, including electric motors. Resolvers are particularly robust and are used in applications with harsh environments, such as aerospace.

[0004] An encoder is a sensor that directly measures the rotational position of a motor. Incremental encoders provide pulses proportional to the motor's rotational motion. Absolute encoders provide information about the motor's absolute position within a range of speeds. Encoder systems are commonly used in servo motors and other applications requiring precise position control.

[0005] Hall effect sensors can be used to determine the rotor position in a brushless DC motor (BLDC). These sensors detect changes in the motor's magnetic field to determine position. Most BLDC motors use three Hall effect sensors to determine position relative to the electrical phases.

[0006] Inertial sensors such as gyroscopes and accelerometers can be used to indirectly derive changes in motor position. By integrating the acceleration data, velocity can be estimated, and by integrating again, position can be estimated. However, this method can become inaccurate over time and requires a normalization procedure.

[0007] When determining the rotor position of an electric motor, various issues can arise that interfere with the measurement signal and generate noise. This can lead to inaccurate position indications from the sensor. For example, sensor errors can occur. Sensors can malfunction or wear out over time or due to weather conditions, leading to inaccurate positioning. Furthermore, some sensors can be sensitive to temperature changes, which can also cause positioning inaccuracies.

[0008] Furthermore, external interference can adversely affect the sensor's signal. For example, low frequencies due to pulse width modulation, nearby electric fields, or other noise sources can degrade the quality of the position signal. Summary of the Invention

[0009] The object of the present invention is therefore to provide a method with which the position signal can be optimized by post-processing.

[0010] This object is achieved by the subject matter of the independent claims.

[0011] According to a first aspect of the invention, this object is achieved by a method for calibrating an ABG filter to improve a position signal of an electric motor. The electric motor is a component of a machine that performs relatively complex processes.

[0012] The machine may be, for example, an excavator, a vehicle, a production plant or the like.

[0013] The method comprises the following steps:

[0014] - creating a digital model of the electric motor, the machine and / or a part of the machine,

[0015] wherein the model outputs the values of position x, velocity x' and acceleration x" of the motor, the machine and / or the part of the machine, respectively,

[0016] The value of position x is corrected by the following value, which is determined by the factor alpha and the first correction term,

[0017] The value of the speed x' is corrected to the following value, which is determined by the factor beta and the second correction term,

[0018] The value of acceleration x" is corrected to the following value, which is determined by the factor gamma and the third correction term:

[0019] - specify the initial values of factors alpha, beta and gamma;

[0020] - determining a first state of the motor, the machine and / or the part of the machine;

[0021] - starting the motor for a defined time t and determining a second state of the motor, the machine and / or a part of the machine;

[0022] - using the digital model of the electric motor, the machine and / or the part of the machine, determining the modeled state at the time t;

[0023] - comparing the second state to the modeled state and determining a loss function,

[0024] Among them, the loss function is a measure of the quality of the digital model;

[0025] - adjusting the factors alpha, beta, and gamma based on the comparison of the second state with the modeled state and the loss function; and

[0026] - Repeating the steps of comparing the second state with the modeled state and adjusting the factors alpha, beta and gamma until the loss function has reached an optimal value.

[0027] Signal post-processing algorithms are known in communications engineering. The difficulty lies in real-time processing and the given machine complexity.

[0028] The Alpha-Beta-Gamma (ABG) filter is a filtering technique used to estimate the position, velocity, and acceleration of a moving object. It is commonly used in navigation systems such as GPS to improve position tracking accuracy.

[0029] The ABG filter combines three separate filters—Alpha, Beta, and Gamma—to estimate the position, velocity, and acceleration of an object. Each filter plays a specific role in the estimation process.

[0030] The alpha filter is responsible for estimating the position of an object based on the available sensor measurements. This alpha filter takes into account the previous position estimate, the current sensor measurement, and a weighting factor alpha. The factor alpha determines the influence of the current measurement on the position estimate.

[0031] The Beta filter estimates the velocity of an object by taking into account the difference between the current position estimate and the previous position estimate. The Beta filter uses a similar weighting factor beta to determine the impact of the velocity estimate.

[0032] The gamma filter estimates the acceleration of an object by taking into account the difference between the current velocity estimate and the previous velocity estimate. The gamma filter uses a factor gamma in order to determine the influence of the acceleration estimate.

[0033] The acceleration target in an ABG filter is the desired acceleration of an object. By incorporating this acceleration target into the estimation process, the filter can adapt its estimates to the expected motion of the object. This helps reduce errors and improve the accuracy of position, velocity, and acceleration estimates.

[0034] The correction terms of the ABG filter can take various influences into account. For example, they can describe interference effects or include noise values. Furthermore, the correction terms for position x, velocity x', and acceleration x" can include different correction terms.

[0035] To calibrate the filter, at least a test run of the machine must be carried out. This can be done under laboratory conditions or in field tests. Furthermore, for this calibration, a model of the movement of the motor, the machine, or a part of the machine that is as accurate as possible is required. The numerical model is used to predict the movement, velocity, and acceleration, which are then compared with the actual position.

[0036] The first state of the electric motor, the machine, and / or a part of the machine can be, for example, a stopped or static position. If the drive component, in particular the monitored electric motor, is then started, the electric motor, the machine, and / or a part of the machine will move. For example, the electric motor can be used as a drive for a vehicle. Furthermore, the electric motor can be used, for example, as a drive for an excavator arm, as a drive for a pump in a hydraulic system, or as a drive for other components of the machine.

[0037] However, since the motor is not an isolated component, the inertia of the entire motion system acts on it. When the motor moves a portion of the machine, or even the entire machine, the machine will continue to move due to inertia. This in turn affects the motor, which must be taken into account in subsequent control.

[0038] Thus, in the second state, the position of the motor, the machine, and / or the part of the machine is again detected. Furthermore, the movement of the motor, the machine, or the part of the machine is simulated using the digital model. During the simulation, an ABG filter with factors alpha, beta, and gamma is used. The results of the simulation are then compared with the actually determined state of the motor, the machine, and / or the part of the machine.

[0039] To quantify this comparison, a loss function is determined, which can be selected in particular depending on the model.

[0040] In any case, the result can be used to adjust the factors alpha, beta and gamma. Since a single adjustment of these factors does not necessarily lead directly to the optimal adjustment of these factors, the adjustment process can be repeated.

[0041] To do this, the motor is restarted, causing the machine or part of it to move. Furthermore, the modeled state is determined and compared again with the new actual state. Based on this new comparison, new adjustments are made to the factors alpha, beta, and gamma.

[0042] This process can be performed, for example, during machine calibration. During machine use, the calibration of the factors of the ABG filter can be performed in the background.

[0043] When the loss function has reached an optimum and the factors alpha, beta and gamma are determined sufficiently well, the calibration process can be considered complete. The machine is ready for use, wherein the signals of the sensors for determining the position of the motor are determined as accurately as possible.

[0044] In one embodiment, the state of the motor, the machine and / or a part of the machine is determined using a resolver.

[0045] A resolver is a special type of sensor used to detect the rotor position of an electric motor. Essentially, a resolver operates through the interaction of electromagnetic fields. It consists of a stationary stator and a movable rotor. The stator is connected to an AC power source that generates a rotating magnetic field. The rotor, located within this field, is connected to the motor's rotor. The relative position between the stator's magnetic field and the rotor affects the resolver's output voltage.

[0046] By measuring this output voltage, the exact angular position of the rotor can be determined. Resolvers are commonly used in applications requiring precise position control and corresponding feedback, such as in servo systems for electric motors, particularly in industrial applications and automation technology. Resolvers provide a robust and reliable way to monitor the rotor's position in real time, even under harsh environmental conditions.

[0047] In one embodiment, the speed x' is determined as the time variation of the position x of the motor determined by the resolver.

[0048] Thereby, the direct physical relationship between position x and velocity x' is advantageously utilized.

[0049] In one embodiment, the acceleration x" is determined as the change in velocity x' over time.

[0050] Thereby, the direct physical relationship between the velocity x′ or position x and the acceleration x″ is advantageously utilized.

[0051] In one embodiment, the digital model includes at least three equations of motion, wherein the equations of motion describe position x, velocity x' and acceleration x".

[0052] The equations of motion are mathematical equations that are related to one another by physical relationships. These equations can be solved analytically, and thus the digital model can be created in a particularly simple manner.

[0053] In one embodiment, the digital model includes a random algorithm.

[0054] A stochastic algorithm is one based on random processes or probability. In contrast to deterministic algorithms, where the output is uniquely determined by the input data, stochastic algorithms account for uncertainty or random influences. These algorithms are often used in situations where there are uncertainties or unpredictable variables. This can include, for example, signal noise or environmental interference generated externally to the system.

[0055] In a first step using a random algorithm, the motor, the machine and / or the part of the machine is modeled. To this end, random variables or random processes are used to represent uncertainty or variability. This may include the use of random numbers, probability distributions or other random concepts.

[0056] Randomized algorithms make decisions based on probability. Rather than providing a definitive answer, these algorithms typically output the probability of a particular event occurring. This allows for a more flexible handling of uncertainty. An example of a tool used in randomized algorithms is Monte Carlo simulation. Here, random samples are drawn from a probability distribution to simulate possible scenarios and analyze their statistical properties.

[0057] Some stochastic algorithms can use adaptive learning to continuously adjust their models or decisions as new data becomes available. This enables them to adapt to changing conditions or environments.

[0058] Randomized algorithms provide an effective way to handle uncertainty in complex systems, but are more expensive than deterministic algorithms in terms of computational resources. The advantage of these randomized algorithms is that they can describe systems that cannot be described simply by equations of motion.

[0059] In one embodiment, the digital model includes a machine learning algorithm.

[0060] A machine learning algorithm is an algorithm that has been developed to automatically identify patterns and relationships in data and make predictions or decisions. The machine learning algorithm is created by training it on existing data and can then be applied to new, unknown data to generate predictions or classifications.

[0061] Machine learning algorithms can take different forms, such as linear models, decision trees, support vector machines, neural networks, etc. The machine learning algorithm is optimized by learning from the training data by identifying patterns and rules in order to make the best possible predictions or classifications for new data.

[0062] The effectiveness of machine learning algorithms depends on various factors, including the quality and quantity of training data, the choice of algorithm, model configuration, and evaluation metrics. Models are continuously refined and optimized to maximize accuracy and performance. To this end, factors such as alpha, beta, and gamma are used as model parameters.

[0063] In one embodiment, the machine learning algorithm comprises a neural network.

[0064] Neural networks are models that are particularly flexible and therefore suitable for a wide variety of applications. The architecture of a neural network comprises a plurality of nodes, neurons, or junctions arranged in layers.

[0065] Neural networks can be advantageously used for machines whose movements cannot be described accurately enough by random algorithms or in which random models provide results that vary too much.

[0066] Depending on the structure of the model used, i.e. the number of nodes and layers used, the computational effort of an implementation using a neural network can become relatively high. Thus, in principle, a trade-off is made between how much computing power is available on the machine and how accurately the movement of the motor, the machine or the part of the machine should be determined.

[0067] In another aspect, the invention relates to a computer program having a program code for performing the method as described above when the computer program is executed on a computer.

[0068] In another aspect, the invention relates to a computer-readable data carrier having a program code of a computer program in order to carry out the method as described above when the computer program is executed on a computer.

[0069] In another aspect, the invention relates to a system for controlling an electronic drive, wherein the system is designed to carry out the method as described above.

[0070] In another aspect, the invention relates to a machine for carrying out an industrial process, wherein the machine has a system for carrying out the method as described above.

[0071] An industrial process is any commercially applicable process that uses at least one machine. Thus, a machine can perform all possible industrial processes. Industrial processes may include, but are not limited to: transporting, cutting, drilling, milling, grinding, stamping, casting and injection molding, assembly, welding, painting, packaging, positioning, and more.

[0072] In another aspect, the invention relates to the use of an ABG filter for optimizing a resolver signal for controlling an electric motor, a machine or a part of a machine, wherein the ABG filter is calibrated according to the method as described above.

[0073] In summary, it should be emphasized that the present invention describes a method for calibrating an ABG filter to improve a position signal of an electric motor, a computer program with a program code, a computer-readable data carrier with a program code, a system for controlling an electronic drive, a machine with a corresponding system, and the use of an ABG filter for optimizing a resolver signal for controlling an electric motor.

[0074] The described embodiments and refinements can be combined with one another as desired.

[0075] Further possible embodiments, developments, and implementations of the present invention also include combinations of features of the present invention described above or below with reference to exemplary embodiments that are not explicitly mentioned. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] The accompanying drawings are intended to provide a further understanding of the embodiments of the present invention. These drawings illustrate the embodiments and, together with the description, serve to explain the principles and design of the present invention.

[0077] Further embodiments and several of the advantages mentioned can be derived with reference to these drawings. The elements represented in these drawings are not necessarily shown to scale with respect to one another.

[0078] in:

[0079] Figure 1 Schematically illustrates the process of the method according to the embodiment; and

[0080] Figure 2 The steps of comparing the modeled state with the measured state are schematically shown.

[0081] In the figures of the drawings, unless otherwise specified, the same reference numerals denote identical or functionally identical elements, components or assemblies. DETAILED DESCRIPTION

[0082] Figure 1 The flowchart of a method for calibrating an ABG filter to improve a position signal of an electric motor is schematically shown.

[0083] In a first step S10, the motor, the machine using the motor, or a part of the machine is described by a digital model. The digital model may be composed of motion equations, a stochastic model, or a machine learning model.

[0084] In any case, the position x, velocity x' and acceleration x" of the electric motor are modeled using the numerical model. To determine the position x, a measured value is used as an input variable, which is modified by a first correction term. The first correction term for the position x includes the factor alpha. To determine the velocity x", the measured value or the change over time of the position x is used, which is modified by a second correction term. The second correction term for the velocity x' includes the factor beta. To determine the acceleration x", the measured value or the change over time of the velocity x" is used, which is modified by a third correction term. The third correction term for the acceleration x" includes the factor gamma.

[0085] For alpha, beta and gamma, initial values must first be specified in step S12. These initial values are adjusted during the calibration process.

[0086] In step S14 , a first state of the motor, the machine, or the portion of the machine is determined. This first state can be determined using a sensor, particularly a resolver of the motor. More complex machines may have multiple motors, each of which can be configured with its own ABG filter. In particular, multiple ABG filters can be calibrated sequentially.

[0087] This first state serves as the starting point for the calibration. From this starting point, changes are determined. Within the scope of the present invention, a state refers in particular to a specific position of the machine, of the part of the machine, or of the attitude of the motor.

[0088] Furthermore, the position x can also be read as representing the angle phi of the motor, the speed x' as the angular speed phi', and the acceleration x" as the angular acceleration phi". This is particularly useful when the determined state does not involve a linear motion but a rotational motion.

[0089] If the state is determined, the motor is started for a fixed time t in the next step S16. By starting, the motor, the machine and / or the part of the machine enter another state. This state is also determined.

[0090] In step S18 , the modeled state is determined. To this end, the digital model of the motor is used and, starting from the first state, the state that the motor, the machine, and / or the part of the machine should be in according to the model is determined. For this purpose, an ABG filter is used.

[0091] In step S20, the modeled state is compared to the second state. The model and the measurement are thereby compared to each other. Based on the difference, it can be determined how accurately the simulation, including the correction using the ABG filter, matches the model.

[0092] Based on the difference, a loss function is determined, which reflects the quality of the ABG filter. In step S22, factors alpha, beta and gamma are adapted to the comparison result and the loss function.

[0093] Finally, steps S16 to S22 are repeated until the loss function reaches an optimal value or at least a defined threshold value in step S20. Optionally, step S16 can also be repeated to induce a different state of the motor, the machine, or the part of the machine. In this case, a new modeled state must also be determined, with which the new state can be compared. This embodiment has the advantage that the selection of the factors alpha, beta, and gamma is more adaptable to different states.

[0094] Figure 2 The interplay between modeled predictions and measurements of state is outlined.

[0095] An ABG filter is a recursive Bayesian filter used to obtain estimates of uncertain or noisy measurements.

[0096] The digital model created for the motor, the machine and / or the part of the machine contains state variables, input parameters, process noise and possible measurement noise. The model, due to its complexity, is usually represented by equations describing the system state over time.

[0097] Using the ABG filter, a prediction 10 of the current state of the system is determined based on the previous estimated state 12 and known input quantities. This step involves applying a system model to calculate the expected state and uncertainty in the prediction step.

[0098] The prediction is then compared to the actual measured values of position x, velocity x', and acceleration x". The ABG filter can take into account not only the prediction of the state but also the uncertainty of the measurement in order to calculate the best estimate of the current state.

[0099] The final step of the ABG filter is to update the state based on the measured information and the estimated state. This step improves the state estimate by integrating the uncertainty in the measurement. The cycle of prediction and measurement updates continues as new measurements become available. In this way, the ABG filter continuously optimizes the state estimate.

Claims

1. A method for calibrating an ABG filter to improve the position signal of a motor, in, The electric motor is a component of the machine, Wherein, the method comprises the following steps: - creating a digital model of the electric motor, the machine and / or a part of the machine (S10), wherein the model outputs values of position x, velocity x' and acceleration x" of the motor, the machine and / or a part of the machine, respectively, The value of the position x is corrected by the following value, which is determined by the factor alpha and the first correction term: The value of the speed x' is corrected to the following value, which is determined by the factor beta and the second correction term, The value of the acceleration x″ is corrected to the following value, which is determined by the factor gamma and the third correction term; - specifying initial values of factors alpha, beta and gamma (S12); - determining a first state of the electric motor, the machine and / or a part of the machine (S14); - starting the electric motor (S16) for a defined time t and determining a second state of the electric motor, the machine and / or a part of the machine; - using a digital model of the electric motor, the machine and / or the part of the machine, determining a modeled state at the time t (S18); - comparing said second state with said modeled state (S20) and determining a loss function, Wherein, the loss function is a measure of the quality of the digital model; - adjusting factors alpha, beta and gamma according to the comparison result of the second state and the modeled state and the loss function (S22); and - Repeating the steps of comparing the second state with the modeled state and adjusting the factors alpha, beta and gamma until the loss function has reached an optimal value.

2. The computer-implemented method of claim 1 , wherein: The state of the electric motor is determined using a resolver.

3. A computer-implemented method according to any one of the preceding claims, wherein: The speed x' is determined as the time variation of the position x of the motor determined by the resolver.

4. The computer-implemented method of claim 3, wherein: The acceleration x" is determined as the change in the velocity x' over time.

5. The computer-implemented method according to any one of the preceding claims, wherein: The digital model includes at least three motion equations, wherein the motion equations describe the position x, the velocity x' and the acceleration x".

6. A computer-implemented method according to any one of the preceding claims, wherein: The digital model includes a random algorithm.

7. A computer-implemented method according to any one of the preceding claims, wherein: The digital model includes a machine learning algorithm.

8. A computer-implemented method according to any one of the preceding claims, wherein: The machine learning algorithm includes a neural network. 9 . A computer program having a program code for carrying out the method according to claim 1 , when the computer program is executed on a computer. 10 . A computer-readable data carrier having a program code of a computer program in order to carry out the method according to claim 1 , when the computer program is executed on a computer.

11. A system for controlling an electronic drive, wherein: The system is designed to carry out the method according to any one of claims 1 to 8.

12. A machine for performing an industrial process, wherein: The machine has a system according to claim 11.

13. Application of ABG filters for optimizing resolver signals for controlling motors, machines or parts of machines, in, The ABG filter is calibrated according to the method of any one of claims 1 to 8.

Images